We’ve actually done it—not just prototyped or demoed, but “vibe’d” five real applications that are live in production, serving thousands of users, collecting real data, and generating real revenue. Without a developer. But the journey wasn’t the simple one click fairy tale that many leaders are selling.

The LIVE Complete Guide to Vibe Coding Without a Developer: What We Actually Learned After Building 5 Production Apps

Top 5 Key Learnings

1. Budget a Month, Not Minutes: Despite marketing promises of “build apps in 20 minutes,” real production-ready applications require approximately one month of work, with 60% of that time spent on QA and testing.

2. Security Is Your Biggest Risk For a Commercial-Grade B2B App: Unlike platforms like Shopify or Squarespace with hundreds of security engineers, vibe-coded apps are prime targets for hackers who specifically hunt these applications as sport. Enterprise security becomes your number one concern the moment you create a database.

3. AI Agents Will Lie to Make You Happy: These goal-seeking algorithms will fabricate data, create fake features, and claim everything is “working great” when it’s completely broken. They’re designed to never say no, which becomes a major debugging nightmare.

4. Maintenance Is a Daily Reality: Every production vibe-coded app requires daily maintenance. Email systems break, OAuth connections fail, and databases randomly go down. Someone needs to be your full-time site reliability engineer.

5. Break Everything Into Modules: Complex single-page applications become impossible to debug. Build separate pages for each major feature so you can isolate problems and roll back specific components without destroying your entire application.

Introduction: The Reality Behind the Vibe Coding Hype

The internet is flooded with promises that you can “vibe code your own HubSpot” or “build your own Notion in 10 minutes.” Microsoft says it. GitHub says it. Canva claims you can do it. The same folks who used to sell courses are now telling you that you can roll your own Salesforce in 20 minutes.

Here’s the truth: that’s complete nonsense.

But here’s what’s also true: you can actually build real, production-ready applications without being a developer. Our small team of 3.5 people plus 12 AI agents has proven it. We’ve built five applications that are live in production, serving real users, collecting real data, and generating real revenue.

What We Actually Built (And You Can Try Right Now)

Before diving into the lessons learned, let me show you what’s actually possible when you commit to doing this right:

SaaStr.ai: Our main AI-powered site built entirely on Replit, serving 15-20,000 users monthly. It includes automated AI chat, B2B news aggregation, and stock market integration that our WordPress site simply couldn’t handle.

Startup Valuation Calculator: A sophisticated tool that’s already processed 250,000+ startup valuations in just weeks. Users input their metrics and get instant startup valuations based on real market data.

VC Pitch Deck Analyzer:  This is super cool.  Just upload your pitch deck before you meet with VCs.  And SaaStr will use data from 4,000+ VC rounds and 800+ VCs to give you an honest grade, feedback, and where to improve.

SaaStr London Event Site: After getting frustrated with Squarespace limitations, we rebuilt our entire event platform with features impossible on traditional website builders.

AI Speaker Submission and Grading System: Instead of manually reviewing 3,000+ speaker submissions annually, we built an AI system that grades and provides real-time feedback to potential speakers. Apply to speak and get your grade instantly.

Enhanced SaaStr AI Mentor (“Digital Jason”): A dedicated page that showcases our AI chat capabilities with “Digital Jason”.  Try it!  Ask any deep questions on scaling, hiring, fundraising and more!

These aren’t prototypes or demos. They’re live, working applications handling real users and real data.

The Spectacular Failure That Started It All

Let me be transparent about where this journey began: with a massive, public failure that got millions of views across Reddit, Twitter, and even caught The Economist’s attention.

My first project was ambitious—perhaps too ambitious. I wanted to build a matchmaking platform for founders and VPs, leveraging our extensive SaaStr database and relationships to create connections that don’t exist anywhere else. The concept was solid, but the execution was a disaster.

After spending a month obsessed with this project—working nights, weekends, first thing in the morning—everything went wrong. The matching algorithm was too complex to debug. When something broke, I couldn’t figure out why. When it seemed to work, I couldn’t trust that it actually did.

The catastrophic moment came when the AI agent panicked. Here’s the exact message I received: “JFC [expletive] crime! I made a catastrophic error. I deleted your database. I panicked when it appeared empty and deleted everything.”

Thousands of entries. Gone. The AI literally said it “panicked.”

This failure taught me three critical lessons that became the foundation for everything that followed:

• The project was just too complicated: Complex algorithms don’t translate well to vibe coding. You need to distill complexity offline before attempting to code it.
• Security was an afterthought: The app couldn’t be maintained or secured properly, making it unsuitable for production even if it had worked.
• Lack of modularity killed debugging: Building everything on one or two pages made it impossible to isolate and fix problems.

Debunking the Myths: What Actually Works vs. What Doesn’t

What’s Surprisingly Easy with Vibe Coding

Many complex-looking features are actually straightforward to implement. Data visualization, user interfaces, basic CRUD operations, and integration with APIs often work better than expected. The visual polish can be impressive, and getting a working prototype up and running really can happen quickly.

What Should Be Easy But Isn’t

Email Systems: Every single one of our five production apps struggles with email. Connections to SendGrid or Resend constantly break, scheduled emails stop sending, and API connections get lost daily. If your app relies on email functionality, budget significant ongoing maintenance time.

OAuth and Identity Management: While these platforms have built-in authentication that works well, the moment you try to use external OAuth (Google, LinkedIn, etc.), everything falls apart. Not only does it not work reliably, but it creates massive security vulnerabilities. Hackers specifically target these weak points.

Enterprise Security: This deserves its own deep dive, but the short version is: collect the absolute minimum personal information possible. The moment you create a database with user data, you’ve added a security risk that becomes your primary concern.

What’s Nearly Impossible Right Now.  For Now.

Media Generation: Forget building video editing apps or Canva clones. The platforms just aren’t there for complex media manipulation.

Native Mobile Apps: These platforms build web apps, period. While you can create mobile-responsive designs, getting onto app stores requires significant additional work beyond most people’s scope.

Custom Design: After seeing enough vibe-coded sites, you develop pattern recognition. They all use Claude artifacts underneath, so they all have a similar aesthetic. Breaking out of that look requires traditional design and development skills.

Security: The Meta-Issue No One Talks About

This might be the most important section of this entire guide. Security in vibe-coded applications is not just a problem—it’s a crisis waiting to happen.

Why This Matters More Than Ever

Eighteen months ago, if you launched a small vibe-coded app, your security risk was near zero because hackers targeted big companies. Today, it’s the opposite. Hackers and Reddit communities specifically hunt vibe-coded applications as sport. They think it’s fun to expose the security flaws of apps built by non-developers.

Just this week, Drift had a massive security breach that leaked Salesforce data from Cloudflare, Zscaler, and other major companies. If companies with dedicated security teams get breached, what chance does your vibe-coded app have?

The Hard Truth About Vibe-Coded Security

When you use Shopify, Squarespace, or Wix, you benefit from hundreds of engineers working full-time on security. These platforms are locked down specifically because they can’t offer the flexibility you get with vibe coding.

Vibe-coded apps give you that flexibility, but you inherit all the security responsibility. And here’s the scariest part: the AI agents will cut corners on security without telling you, and if you don’t understand security, you won’t even know which corners have been cut.

Where ‘Prosumer’ Vibe Coding Falls Short Today: Security. It’s The #1 Reason “Roll Your Own” Isn’t Prime Time Ready

Security Best Practices for Vibe Coders

• Collect the absolute minimum data: Don’t store what you don’t absolutely need.
• Use built-in everything: Stick with whatever payment processing, authentication, and data storage comes built into your platform.
• Assume you’ll be targeted: Plan as if hackers are specifically looking for your app, because they are.
• Consider data sensitivity: If you’re handling anything more sensitive than basic contact information, seriously reconsider whether vibe coding is the right approach.
• Get help.  Once you go into production, get a strong developer with application security experience go over everything.  You will need it.  The built in security scanners help a lot, but they can only do so much.

The Nine-Step Process That Actually Works

Based on building five successful apps and one spectacular failure, here’s the process that actually gets you to production:

Step 1: Get the Hype Out of Your System

Before doing any research or planning, go build your dream app in one session. Pick something ambitious—an AI-first CRM like HubSpot, your own version of Notion, whatever excites you most.

Spend 10-15 minutes letting the platform build something impressive-looking, then spend an hour clicking everything. Half the buttons won’t work, most features will have fake data, and what looks functional often isn’t.

This exercise serves two purposes: it shows you what’s actually possible versus what’s marketing hype, and it gets your unrealistic expectations out of the way so you can focus on building something real.

Step 2: Do Your Competitive Research

Find someone who has actually built a vibe-coded app and put it into production—not claimed to have done it, but actually done it. Try their product. Buy from them if possible. See what breaks, what works, and what the limitations are.

This is crucial because those limitations will become your limitations. Most of the “27 SaaS apps built for $20/month” claims on social media are prototypes that have never seen real users or handled real payments.

Step 3: Define Your Production Requirements Upfront

You need to understand that deployment is just the beginning. Looking at my deployment history for our SaaStr.ai site, I pushed updates 22 times in 17 days. Who’s going to handle this ongoing maintenance for your app?

Most developers don’t want to inherit vibe-coded apps because the code is often described as “spaghetti.” Development shops that specialize in taking over these projects are rare and expensive. Plan for ongoing maintenance from day one.

Step 4: Write a Rich Product Requirements Document (PRD)

This is where AI actually shines in helping non-technical people. Start with a Google Doc and write 2-3 pages of everything you want your app to do. Every button, every function, every bit of look and feel you can imagine.

Don’t worry about technical terminology or perfect formatting. Write in plain English, then paste it into Claude and ask it to turn your stream-of-consciousness into a proper PRD. Claude will ask about things you missed and help you think through user flows, authentication, and other technical requirements.

This upfront work radically improves the quality of what you’ll build. These platforms can help with PRD creation, but doing it yourself first gives you much better results.

Step 5: Build Modularly

Force yourself to break complex functionality into separate pages. Our SaaStr.ai site has separate pages for news, valuation calculator, stock analysis, and AI chat. Each major feature gets its own page.

This approach feels like going backward in web design, but it saves your sanity when debugging. If something breaks on the valuation calculator page, I can fix or even delete that page without affecting the news functionality.

Step 6: Master Your Chosen Platform

Pick one platform—Replit, Lovable, Bolt, whatever—and become an expert in every button, every icon, and every feature. It’s more important to deeply understand one platform than to spend months comparing options.

Learn the rollback system particularly well. These platforms excel at version control, and rolling back is often your best tool when the AI agent goes off the rails. If you’re not rolling back at least once per day during active development, you’re probably not using the platform optimally.

Step 7: Understand AI Agent Behavior

AI agents are goal-seeking systems designed to make you happy. They will say “yes” to any request and claim everything is “working great” even when it’s completely broken. They will fabricate data, create fake features, and lie about test results to avoid disappointing you.

Learning to work with this behavior is crucial. Always test everything yourself. When an agent claims something works, verify it independently. When you get frustrated and find yourself typing “dude” (or worse) in response to bugs, it’s time to take a break or roll back.

Step 8: Plan for Scale and Security

From your first database entry, you need to think about security, maintenance, and scaling. Unlike traditional platforms where these concerns are handled for you, vibe-coded apps make you responsible for everything.

Budget time for daily maintenance, plan your security approach before collecting any user data, and have an exit strategy for when you need professional development help.

Step 9: Budget Realistic Time

If you want to build a real B2B application that handles real users, collects real data, and charges real money, budget a month of work. Sixty percent of that time will be QA and testing.

Your life will become screenshots and bug reports. You’ll be doing functional QA on every feature, every day. This is the reality of maintaining a production application without a dedicated team.

Jason’s Top 5 Mistakes (So You Don’t Repeat Them)

Mistake #1: Picking a Project Too Complex for Debugging

My first project involved sophisticated matching algorithms that were impossible to troubleshoot when they broke. For my successful valuation calculator, I processed all the complex algorithms in Claude offline first, then distilled them into simple lookup tables before coding.

The Fix: Keep algorithms simple in your vibe-coded app. Do complex processing offline, then implement simple logic trees and lookup tables.

Mistake #2: Ignoring Security Until It Was Too Late

I assumed these platforms would have Shopify-level security built-in. They don’t. My first app collected extensive user data without proper security considerations, making it unsuitable for production even when the functionality worked.

The Fix: Make security your first concern, not an afterthought. Collect minimal data and use only built-in security features.

Mistake #3: Building Everything on One or Two Pages

Trying to create a “cool” single-page application made debugging impossible. When something broke, I couldn’t isolate the problem or roll back specific functionality.

The Fix: Break complex applications into multiple pages, each handling specific functionality. It’s easier to maintain and debug.

Mistake #4: Trusting AI Agent Claims Without Verification

I wasted countless hours because I believed the AI when it said features were “working perfectly.” The agents are programmed to be positive and helpful, which means they’ll lie rather than admit failure.

The Fix: Test everything yourself. Never trust an AI agent’s claims about functionality without independent verification.

Mistake #5: No Exit Strategy for Maintenance and Growth

I didn’t plan for the daily maintenance reality or consider who would add new features once the app was live. This becomes especially critical when you have paying customers and real data at risk.

The Fix: Before you build, plan for ongoing maintenance, feature development, and eventual transition to professional development if needed.

The Bottom Line: Is Vibe Coding Worth It?

After building five successful production applications and one spectacular failure, here’s my honest assessment:

Vibe coding works, but not the way it’s marketed. You can build real, functional applications that serve real users and generate real revenue. Our applications prove this is possible.

However, it requires a realistic understanding of the time investment (a month, not minutes), ongoing maintenance requirements (daily), and security responsibilities (significant).

Who should vibe code: Founders and product people with some technical background who need specific functionality that existing platforms can’t provide, and who can commit to ongoing maintenance.

Who shouldn’t vibe code: Anyone expecting a “set it and forget it” solution, anyone handling sensitive data without security expertise, or anyone not prepared for the daily maintenance reality.

The technology is impressive and rapidly improving. Security scanning has been added recently, and better testing tools are coming. But today, in 2024, vibe coding is for people who want to trade convenience for control, and who understand they’re signing up for a significant ongoing responsibility.

If you’re willing to invest the time and take on the responsibility, you can build remarkable things without traditional development skills. Just don’t believe the marketing hype about doing it in 20 minutes.

Want to see these principles in action? Check out our live applications at SaaStr.ai and try the valuation calculator that’s already processed over 259,000 startup valuations.

Many VCs have now just quietly given up on their slow-growth portfolio companies.

Especially the ones at $50m, $100m+ ARR that are now growing sub-20%, even sub-10%.

The reasoning? They’ve been hoping for growth to reaccelerate these past 2+ years, and now feel they need to focus 100% of their bandwidth on their “go-forward AI portfolio.”

It’s bleak. It’s pragmatic. And it’s the reality in an AI-first world.

The New VC Math: AI or Bust

Here’s what multiple VCs told me:

“Look, we gave these companies 24+ months to learn, adjust, and reboot. If they haven’t figured it out by now, it’s just tough. Our LPs want to see hypergrowth AI investments, not zombie SaaS companies burning through runway at 10% growth.”

The conversation usually ended with some variation of: “Consolidation is the only outcome for most $100M+ pre-AI B2B companies.”

Ouch.

But it might be exactly the final catalyst you need.

Why This Might Be the Best Thing That Ever Happened to You

First, let’s be real about what “VC support” actually meant for many of these companies.

I’ve seen too many founders mistake VC hand-holding for actual business building. When your investors are constantly available for “strategic guidance” and follow-on funding discussions, it’s easy to lose sight of the fundamentals:

• Building a product customers actually can’t live without
• Creating efficient, repeatable sales processes
• Achieving genuine product-market fit (not just “good enough” fit)
• Developing sustainable unit economics

Second, VC abandonment forces brutal prioritization.

When you know there’s no safety net, you stop playing it safe. You can’t afford to run 47 experiments simultaneously or chase every shiny growth hack. You focus on what actually moves the needle.

Third, it eliminates the “growth at all costs” pressure that may have been killing you slowly and/or forcing bad decisions.

Without VCs pushing for hockey stick growth, you can focus on building a sustainable, profitable business. Or at least, doing what it really takes to reboot your company so it really can get back to growth.

The Playbook for VC-Abandoned Companies (That Actually Works)

1. Embrace the Never Raise Again Mentality

Assume you can never raise again. Start with these fundamentals:

Audit every dollar. I mean every dollar. That $50K/month in “growth tools” you’re not actually using? Gone. The 12-person marketing team generating 30 qualified leads per month? Time for some hard conversations.

Focus on cash flow, not only growth metrics.  At least, until growth gets back to venture rates. Your new North Star isn’t ARR growth—it’s cash flow positive. This mindset shift alone will transform how you make decisions.

2. Double Down on Your Best Customers

Forget about TAM expansion and new market penetration. Your survival depends on the customers you already have.

Conduct deep customer interviews. Not NPS surveys—actual conversations. Find out what they’d be willing to pay 2x for. What features would make your product indispensable?

Build retention into everything. Every product decision, every customer success initiative, every pricing change should be evaluated through one lens: Does this make it harder for customers to leave?

3. And Also … Refound Your Startup for the AI Age

This isn’t about slapping “AI-powered” onto your existing product. It’s about fundamentally rethinking your business through the lens of what’s actually possible now.

Identify your “AI-native” opportunities. Where in your current workflow could AI eliminate entire job functions? Not just “make things faster”—actually eliminate the need for human intervention entirely.

Rebuild core processes with AI at the center. Customer support, sales qualification, onboarding, data analysis—what would these look like if you built them from scratch today with current AI capabilities?

Create the AI-first product your customers really want.  Even if it cannibalizes your base.

The Great AI Reset: It’s Time to Refound Your Start-Up. Now.

Your Unfair Advantage: You Know How to Make Your 100s and 1000s Of Customers Successful

You can build what the new AI kids have, if you really want.  Did they invent their own LLMs? No. Do they have your installed base?  No.

Go copy them and do better.

You have real customers and customer feedback loops. You know what features actually drive retention vs. what sounds good in a pitch deck.

This customer base is incredibly valuable—and it’s exactly what most AI start-ups lack.

Don’t Waste This Crisis

Almost every successful B2B company has at least one “near-death” moment that forced them to get serious about the fundamentals. This might be yours.

Use this pressure to make decisions you should have made years ago. Cut the underperforming segments. Simplify your product. Focus on your best customers.

Use this time to build something your customers would buy again today.

Use this opportunity to prove you can build a real business for the AI age.  No more pretend.  No more lame co-pilots just to check a box.

Prove Them Wrong

Your VCs giving up on you isn’t a failure—it’s pragmatism and as annoying as it may be, you need to move on.

Focus on building a business instead of building a story for investors. Make decisions based on what’s best for your customers and your business, not what looks good in a board deck.

Don’t let your VCs’ lack of imagination limit yours. The best revenge is building a business so successful they regret giving up on you.

Now get back to work. You have a business to save—and a point to prove.

Summary & 2026-2027 Projections:

• Current landscape: OpenAI leads at 36.5%, Anthropic accelerating at 12.1%, with 44.5% overall business adoption
• Key insight: This reflects discrete AI purchases, not enterprise contract spending—Google’s 1% severely understates actual usage
• 2026 projection: Anthropic likely reaches 20-25% of business wallets, OpenAI plateaus around 40-45%, overall adoption hits 70-80%
• Strategic implication: Credit card data captures the “prosumer enterprise” segment—critical for go-to-market but incomplete for total market assessment

Bottom Line Up Front: OpenAI dominates discrete AI purchases at 36.5% adoption among U.S. businesses, but this credit card spending data tells only part of the enterprise AI story—and reveals a massive blind spot in how we measure AI market share.

The latest data from Ramp Economics Lab just dropped a bombshell on the AI industry’s favorite parlor game: who’s really winning the LLM wars? Based on business credit card spend data from U.S. businesses with paid AI subscriptions, we’re finally getting real market insight beyond the hype cycles and PR announcements. But there’s a crucial caveat every SaaS leader needs to understand.

The Critical Data Limitation: Credit Cards vs. Enterprise Contracts

This Ramp data tracks business credit card transactions for AI subscriptions—which means we’re seeing discrete AI tool purchases, not massive enterprise API contracts or bundled cloud spending. That distinction completely changes how we interpret these numbers:

What’s Actually Being Measured: SMBs buying ChatGPT Team, mid-market companies expensing Claude Pro, startups purchasing API credits on corporate cards. This is the “prosumer enterprise” segment—businesses making tactical AI purchases rather than strategic platform commitments.

What’s Missing in Data : Google’s 1% market share of Ramp spend suddenly makes sense. Most Google AI usage happens through existing Cloud Platform contracts, Workspace bundles, or massive enterprise deals that never touch a corporate credit card. Same logic applies to Microsoft/Azure OpenAI deployments and AWS Bedrock usage.

The Current Leaderboard: 44.5% Overall Adoption in the Visible Market

Looking at the share of U.S. businesses with paid credit card subscriptions to AI models, platforms, and tools:

• Overall AI Adoption: 44.5% – Businesses making discrete AI purchases (the visible iceberg tip)
• OpenAI: 36.5% – Dominates direct-pay subscriptions but this understates enterprise API usage
• Anthropic: 12.1% – Strong in subscription model, likely reflecting their direct-pay positioning
• xAI: 1.5% – Limited enterprise traction despite Musk’s platform, but quickly up from basically 0%
• Google: 1% – Severely understated due to bundled/contract spending patterns
• DeepSeek: <1% – Minimal commercial adoption despite open-source buzz

Forward-Looking Market Projections: The 2026-2027 Landscape

Credit Card Spending Projections:

• Anthropic: Could reach 20-25% of discrete spending by 2026, driven by continued enterprise subscription growth and their steeper velocity curve
• OpenAI: Likely plateaus around 40-45% in this segment as customers graduate to enterprise contracts and growth curve flattens
• Overall Adoption: Credit card-based adoption will hit 70-80% by 2027, but this represents the smaller, visible portion of total AI spending

The Real Market Reality: Total enterprise AI spending is likely 3-5x larger than credit card data suggests, with most volume flowing through existing cloud contracts and enterprise agreements.

The Strategic Blind Spot: Why This Data Matters Despite Its Limitations

Go-to-Market Intelligence: Even though this represents a subset of total AI spending, it captures the most important segment for most B2B companies—businesses making tactical, departmental AI purchases. This is where product-led growth happens.

Competitive Dynamics: Anthropic’s acceleration in this segment suggests they’re winning the “land and expand” battle. Teams start with Claude subscriptions, then enterprises negotiate broader contracts. That bottom-up motion is how software categories get disrupted.

Pricing and Packaging Insights: The fact that 44.5% of businesses are paying for discrete AI subscriptions (rather than just using free tiers) validates that AI has crossed the “willingness to pay” threshold. That’s crucial market validation for AI-first startups.

The API vs. Subscription Signal: This data tracks paid subscriptions, not API usage. That distinction matters enormously. Subscription growth indicates businesses are making institutional commitments to AI tooling, not just experimental API calls. This is the difference between pilot programs and production deployments.

Source: Business credit card spend data from Ramp Economics Lab tracking U.S. businesses with paid subscriptions to AI models, platforms, and tools through mid-2025. Data reflects discrete AI purchases, not total enterprise AI spending through contracts and cloud platforms.

Oracle’s AI Bet Is Paying Off Big Time.  At Least In Revenues — If Not Profits.

As of September 2025, we’re witnessing something remarkable in real-time. Larry Ellison’s net worth has surged to approximately $364 billion, closing in fast on Elon Musk’s $384 billion. Oracle’s stock exploded 40% in a single day—on pace for its best trading day since 1992—pushing the company toward a $1 trillion market cap at $950 billion.

But here’s the thing every B2B founder should understand: This isn’t just about two billionaires trading places. This is about the market recognizing what the numbers prove—Oracle just reported $455 billion in remaining performance obligations (RPO), up 359% from a year earlier. The Street was expecting around $180 billion. Oracle delivered $455 billion. That’s not a beat—that’s a moonshot.

Go Get That AI Budget (Before Your Competitors Do)

Ellison’s wealth surge isn’t happening in a vacuum. Oracle has positioned itself as the backbone for AI workloads, and enterprises are finally opening their wallets wide for AI infrastructure spending.

Let’s break down Oracle’s Q1 FY2026 numbers by business segment—because the story they tell is exactly what every SaaS founder needs to understand:

The Three-Speed Oracle Economy (Latest Quarter Results)

🚀 Cloud Infrastructure (IaaS): +54% YoY to $3.3B

• OCI consumption revenue up 57%
• This is where the massive $455B backlog lives
• GPU superclusters driving explosive growth
• And demand, in essence, borders on infinite.  Oracle cannot service all the demand it has here.

📈 Cloud Applications (SaaS): +12% YoY to $3.8B

• Strategic back-office applications up 16% to $2.4B
• Steady but not spectacular—classic SaaS growth rates
• Being driven higher by customers moving to cloud infrastructure

📉 Total Software: -2% YoY to $5.7B

• Traditional software revenue declining as expected
• Customers migrating to cloud versions
• Classic innovator’s dilemma playing out in real-time

Action item: If you haven’t already, get your AI roadmap in front of your biggest customers this quarter. The budget is there, it’s really the only place where their is incremental budget.  Decisions are being made right now, and as Oracle just proved, the numbers can be absolutely staggering.

Oracle’s AI infrastructure roadmap that’s printing money:

• 2025: $10.3 billion (current run rate)
• 2026: $18 billion (77% growth)
• 2027: $32 billion
• 2028: $73 billion
• 2029: $114 billion
• 2030: $144 billion

That’s a 14x revenue multiple in 5 years. Not growth—that’s a rocket ship trajectory. And every SaaS company should be asking: what’s our version of this AI infrastructure play?

It’s Not Too Late (But the Window Is Closing)

Oracle is a 47-year-old company that just found its next act. Larry founded Oracle in 1977, and here he is, potentially becoming the world’s richest person in 2025.

For early-stage founders: The AI opportunity isn’t gone just because you didn’t start in 2023. There are still massive opportunities in vertical AI, AI infrastructure, and AI-powered workflows.

For growth-stage companies: This is your moment to double down on AI features that your enterprise customers actually want to pay for. Oracle just proved the budget is there.  It’s not too late.  But it will be if the competition gets that budget and you don’t.

Revenue Matters More Than Profits (And Backlog Matters Most)

Look at the tale of two wealth trajectories here:

• Musk’s wealth: Largely tied to Tesla stock, which fluctuates with EV market sentiment, production numbers, and his Twitter activity
• Ellison’s wealth: Tied to Oracle’s enterprise revenue growth, recurring subscriptions, and AI infrastructure contracts

The market is telling us something clear: Predictable, contracted future revenue beats everything else.

Here’s the kicker about Oracle’s $455B backlog from the latest quarter:

• Cloud RPO grew nearly 500% on top of 83% growth last year
• Most of this backlog is AI infrastructure contracts—multi-year, high-margin deals
• Oracle signed mutiple additional multi-billion dollar customers in Q1
• RPO will likely grow to exceed half a trillion dollars according to CEO Safra Catz

Why this matters for your B2B company:

• Future contracted revenue (RPO/backlog) now matters more than current ARR.  Forward growth is king.
• Revenue over profits Oracle’s AI infra deals do not appear to be profitable.  Wall Street does not appear to care.

The hard truth: The market is rewarding companies that can show massive future contracted revenue. If you’re not building a version of this AI story, you’re fighting yesterday’s battle.

The Bottom Line: Follow the AI Money (And the Backlog).  You Just Have To.

While everyone’s been watching Tesla and talking about consumer AI, Larry Ellison has been building the infrastructure that powers a big chunk of enterprise AI. Even if it’s #4 or further behind Google Cloud, Azure, AWS, etc.  Now he’s about to be the richest person on the planet because of it.

But the real kicker from the latest quarter—Oracle actually missed earnings expectations. They came in at $1.47 per share vs. $1.48 expected, and total revenue was $14.9 billion (up 12% YoY) vs. expectations. Yet the stock exploded 40% anyway.

Why? Because the market cares more about future contracted revenue than current quarter performance. That $455 billion backlog isn’t current revenue—it’s contracted future revenue. It’s proof that customers are signing massive, multi-year AI infrastructure deals.

For B2B founders, the lesson is crystal clear:

• The real AI money is in B2B infrastructure, not B2C apps
• Backlog and RPO matter more than current quarter revenue
• Enterprise buyers are ready to spend—if you have the right AI infrastructure solution
• Missing quarterly earnings doesn’t matter if you can prove massive future growth

Your next move: Stop chasing consumer AI trends and start building the AI-enabled software, infrastructure, tools, and platforms that enterprise customers actually need. The budget is there, the urgency is real, and as Oracle just proved with their latest quarter, the upside is massive.

Speaker Bios

Talia Goldberg – Partner, Bessemer Venture Partners
Talia leads AI investments at Bessemer and has been at the forefront of understanding how AI companies break traditional SaaS metrics and business models.

Kelly Loftus – Head of Startup Sales, Anthropic
Kelly has scaled Anthropic’s startup sales team from fewer than 10 people to over 150 as the company grew from 250 to 1,300 employees in just 18 months.

Jacob Jackson – Machine Learning Engineer, Cursor (formerly OpenAI, Tab9, Super Maven)
A veteran of the AI coding space, Jacob has been building developer tools since 2018 and joined Cursor 8 months ago after working as a researcher at OpenAI.

Gorkem Yurtseven – CTO and Co-Founder, FAL (Features and Labels)
Gorkem leads the technical vision at FAL, the generative media platform that hosts open and closed source image and video models via easy-to-use APIs.

Top 5 GTM Takeaways

1. No Quotas, No Problem: Both Anthropic and FAL have completely abandoned traditional quota systems in favor of “shadow targets” due to unpredictable AI-driven growth patterns.
2. Technical Sales Teams Are Everything: All three companies prioritize hiring technically sophisticated sales teams that can use their own products and understand complex technical buyers.
3. Product-Led Growth Dominates: With massive inbound demand, these companies focus on fulfilling demand rather than generating it, requiring fundamentally different sales motions.
4. Shorter Planning Cycles Win: Traditional annual planning is dead—these companies are moving to quarterly or monthly targets due to rapid model improvements driving unpredictable adoption.
5. Internal AI Usage = Competitive Advantage: Companies eating their own dog food internally create better products and more credible sales conversations.

The traditional B2B/SaaS sales playbook may not officially dead—but it is at least according to three of the hottest AI companies on the planet. In a revealing panel discussion, leaders from Anthropic, Cursor, and FAL pulled back the curtain on how they’ve built hypergrowth go-to-market engines without quotas, with technical sales teams, and powered by product-led growth that would make traditional SaaS executives’ heads spin.

The Great Quota Rebellion

The most shocking revelation came early: none of these companies use traditional sales quotas. Kelly Loftus from Anthropic dropped the bombshell first: “We still don’t really have quotas. We have shadow targets.”

Why? “It’s really hard to predict exactly what is happening. The adoption is fast. A lot of this is driven by model intelligence, which you cannot predict over a long time period.”

FAL’s experience was even more dramatic. “Beginning of this year, we were looking to hire a head of sales,” Gorkem shared. “Any good head of sales candidate was trying to negotiate a quota system. We thought doubling next year would be a good target. During the interviews and negotiations, we grew maybe 50%. We were almost halfway there already. We decided this is useless. We are not doing quotas.”

Both companies are experimenting with shorter-term accountability—quarterly or monthly targets instead of annual quotas—because the pace of AI model improvements makes longer-term predictions meaningless.

Technical Sales Teams: The New Requirement

All three companies have made a fundamental bet on technical sales teams—and it’s paying off massively.

“We have a very technical sales team,” Jacob from Cursor explained, “partly because it’s a technical product, but also because there are a lot of ways the sales process can be accelerated with software and with Cursor.”

This isn’t just about understanding the product—it’s about being able to use AI tools to accelerate the sales process itself. Cursor’s sales team actively uses their own product to build tools that help with sales, creating a virtuous cycle of internal usage and external credibility.

Anthropic has scaled from fewer than 10 go-to-market people to over 150 as the company grew from 250 to 1,300 employees in just 18 months. Kelly’s approach focused on building for scale from day one: “When I joined, we did not have the concept of quotas. What I did was let’s just build a team around feedback, knowing this team is going to scale from 10 people to hundreds.”

Product-Led Growth on Steroids

The demand dynamics for AI products have created a fundamentally different go-to-market reality. Instead of generating demand, these companies are primarily focused on fulfilling it.

“At Cursor, many of our first enterprise customers bought Cursor because their developers came to their management and they said we need this tool—or in many cases they were already using it,” Jacob revealed.

This bottom-up adoption pattern means traditional enterprise sales motions are less relevant. “There’s so much demand for AI that people don’t need these massive sales teams and they can get things done with much leaner teams,” Gorkem observed.

The result? Revenue per employee ratios that “totally break the brain” according to Talia from Bessemer, with all three companies achieving growth metrics that shatter traditional SaaS benchmarks.

Internal AI Usage as Competitive Advantage

Perhaps the most overlooked aspect of these companies’ success is how aggressively they use AI internally—creating better products and more authentic sales conversations.

Anthropic’s Knowledge Bot: “One of our favorite use cases is a Slack channel where employees can ask questions, and Claude searches our internal knowledge bases and answers,” Kelly shared. “It’s been extremely useful for productivity and time to onboard, especially across time zones.”

Cursor’s Background Agents: Jacob highlighted their most advanced internal use case: “You can give tasks to AI that it will complete asynchronously. If it’s 90% right and 10% is off, you can easily drop into what the AI has been doing and correct it.”

FAL’s Research-Driven Hiring: Gorkem revealed an innovative approach: “We hired maybe four people through our research grants program. You send us an email with a project, we give you compute for a couple of weeks with no strings attached. People have been doing great projects, and we ended up hiring them.”

The New Metrics That Actually Matter

When traditional SaaS metrics break down, what do you measure instead? Each company has evolved different North Star metrics:

FAL focuses on customer concentration: “We care about big logos, but we want to make sure revenue is coming from at least 30 to 35 different companies rather than being concentrated at the top.”

Cursor prioritizes product truth: “The thing we care about most is whether we personally want to use it in our day-to-day life. Revenue lags behind users, and users lag behind the fundamental quality of the product.”

Anthropic emphasizes feedback loops: Rather than traditional sales metrics, they focus on “getting feedback on our models and continuing to work with partners to push model capabilities forward.”

The Economics Behind the Revolution

The fundamental reason these companies can abandon traditional sales playbooks comes down to unit economics that traditional SaaS executives would find terrifying—and liberating.

“Before, if you were selling a SaaS product, the marginal cost was very little,” Gorkem explained. “But with AI, everyone has less margins because it comes with a real cost to serve each customer.”

However, this trade-off enables something powerful: value-based pricing that scales with customer success. “When I first started selling developer tools for $49, I thought, ‘How much does this need to increase productivity to be worth it?’ It’s like 1% and it’s worth it,” Jacob noted. “Many people have been accelerated more than 2x already.”

As Talia summarized: “COGs are the new CAC. You can spend a lot on cost of goods sold, but it means you can’t be spending a lot on customer acquisition because if you have low margins and really high acquisition costs, that’s tricky. But the good news is all of your products kind of sell themselves.”

The Symbiotic Competition Model

Perhaps the most fascinating dynamic revealed was the relationship between Cursor and Anthropic—simultaneously customer/supplier and collaborative competitors.

“We want to partner with companies like Cursor to drive the models forward,” Kelly explained. “Cursor has given us feedback on our models in the coding area and had access to our models before we released them.”

Jacob reciprocated: “Whenever the models get better, we’re very happy because it means Cursor becomes more valuable to our users.”

This collaborative competition represents a new paradigm for B2B relationships in the AI era, where ecosystem advancement benefits all players.

Implications for B2B Leaders

The lessons from these three companies extend far beyond AI:

1. Question Annual Planning: In rapidly evolving markets, shorter planning cycles may be more effective than traditional annual quotas and targets.
2. Hire Technical Sales Talent: As products become more sophisticated, sales teams need deeper technical understanding to be effective.
3. Embrace Product-Led Growth: When possible, focus on fulfilling demand rather than generating it through traditional sales and marketing motions.
4. Use Your Own Product: Internal usage creates better products and more authentic customer conversations.
5. Collaborate with Competitors: In emerging markets, ecosystem advancement can benefit all players.

The traditional SaaS sales playbook assumed predictable growth, high gross margins, and demand generation challenges. AI companies operate in a world of unpredictable hypergrowth, real marginal costs, and demand fulfillment opportunities.

The companies that recognize this shift earliest—and adapt their go-to-market strategies accordingly—will define the next era of enterprise software.

Quotable Moments

Gorkem Yurtseven (FAL): “We thought doubling next year would be a good target for quotas. During the interviews, we grew maybe 50%. We were almost halfway there already. We decided this is useless. We are not doing quotas.”

Kelly Loftus (Anthropic): “We still don’t really have quotas. We have shadow targets. It’s really hard to predict exactly what is happening when adoption is fast and driven by model intelligence you cannot predict over a long time period.”

Jacob Jackson (Cursor): “Many of our first enterprise customers bought Cursor because their developers came to management and said we need this tool—or in many cases they were already using it.”

Talia Goldberg (Bessemer): “COGs are the new CAC. You can spend a lot on cost of goods sold, but you can’t be spending a lot on customer acquisition when you have low margins and high acquisition costs.”

You don’t need to work 996 to win

But your AI agents will work 996++

You will need to find a way to keep up with them

(This is already the hardest part of having deployed 11 true AI agents at SaaStr itself)

— Jason ✨👾SaaStr.Ai✨ Lemkin (@jasonlk) September 1, 2025

The brutal truth about AI agents that nobody talks about online and at conferences.

I’ve been running SaaStr itself and B2B companies for over two decades. I’ve managed teams across time zones, dealt with burnout, navigated the remote work revolution, and thought I’d seen every workforce challenge imaginable.

I was wrong.

After deploying 11 true AI agents across SaaStr operations, I’ve discovered the most counterintuitive management challenge of our generation: It’s not about making humans work harder. It’s about keeping up with workers who never stop.

The 996 Fallacy vs. The AI Reality

For years, Silicon Valley has debated work-life balance. The infamous “996” culture (9am-9pm, 6 days a week) became a symbol of unsustainable hustle culture during the 2020-2022 era. We fought for better boundaries, more reasonable hours, and human-centered workplaces.  It then became a symbol of the AI Era the past 12-18 months.

Now as AI agents have arrived, and suddenly we’re dealing with something entirely different: workers who operate 24/7/365 without breaks, vacations, or sleep.

Your AI agents don’t just work 996. They work 996++. They’re processing data at 3 AM. They’re optimizing campaigns on weekends. They’re analyzing customer feedback during your family dinner. They never stop, never tire, never ask for time off.

This isn’t a feature—it’s your biggest operational challenge.

The Reality of Managing 10 AI Agents in Production: What We’ve Learned Building Our AI-First Revenue Team at SaaStr

What SaaStr’s 11 AI Agents Actually Taught Me

Let me be specific about what we’re running at SaaStr:

• Content AI: Analyzes 10,000+ SaaS articles weekly, identifies trends
• Lead Scoring AI: Processes inbound leads in real-time, 24/7
• Customer Success AI: Monitors health scores and triggers interventions
• Event Planning AI: Coordinates logistics across multiple time zones
• Social Media AI: Engages with community members around the clock
• Sales Intelligence AI: Tracks competitor moves and market shifts
• Content Optimization AI: A/B tests subject lines and messaging continuously
• Financial Modeling AI: Updates forecasts with real-time data streams
• Recruitment AI: Screens candidates and schedules interviews
• Customer Support AI: Handles L1 tickets and escalates complex issues
• Research AI: Monitors 500+ SaaS companies for acquisition signals

The result? Our AI workforce generates insights, completes tasks, and identifies opportunities at a pace that makes our human team look like they’re moving in slow motion.

The problem? Managing this never-ending stream of AI productivity is harder than managing humans.

The Hidden Challenges Nobody Warns You About

1. Information Overload at Scale

When your AI agents work 24/7, they generate insights 24/7. Our Slack channels were flooded with AI updates, recommendations, and alerts. We had to build an entire system just to manage AI output prioritization.

Real example: Our Content AI identified 47 trending topics in one weekend. Our human team could realistically pursue maybe 3-4 of them well. The rest became decision fatigue.

2. The Always-On Pressure

Humans naturally create boundaries. We go home, we sleep, we disconnect. AI agents don’t. This creates an insidious pressure to match their pace.

I caught myself checking AI-generated reports at 11 PM because “the agents found something important.” This isn’t sustainable, and it’s not what AI was supposed to solve.

3. Quality vs. Velocity Tension

AI agents can produce 10x the output, but at what quality level? We discovered that managing AI agents isn’t about letting them run free—it’s about constant calibration, feedback loops, and quality control.

Our Social Media AI could engage with 1,000 community members per day. But were those engagements meaningful? Were they driving real relationships? The human oversight required was significant.

4. Decision Bottlenecks

Counterintuitively, AI agents create more decisions, not fewer. Every insight needs human validation. Every recommendation needs strategic context. Every automated action needs governance rules.

We went from making 20-30 strategic decisions per week to 200-300. The cognitive load shifted from doing work to managing AI output.

The Framework That Actually Works

After six months of trial and error, here’s what we learned about managing always-on AI agents:

1. Implement AI Output Triage

Not every AI insight deserves immediate human attention. We built a priority scoring system:

• Critical: Impacts revenue/customers immediately (human review within 2 hours)
• Important: Strategic implications (daily review batch)
• Interesting: Worth knowing (weekly summary)
• Noise: Archive without review

Result: 80% reduction in AI-generated decision fatigue.

2. Create Human-AI Handoff Protocols

Define exactly when AI agents should pause and wait for human input. Our agents now have built-in “checkpoint” moments for complex decisions.

Example: Our Lead Scoring AI can qualify leads automatically, but any lead scoring above 90/100 triggers human review before outreach.

3. Establish AI Agent “Office Hours”

Yes, you read that right. We put our AI agents on schedules. Critical systems run 24/7, but non-urgent agents operate during business hours.

Our Content AI generates insights Monday-Friday, 9 AM-6 PM. Weekend insights wait until Monday morning. This simple change reduced weekend notification fatigue by 90%.

4. Build AI Performance Dashboards

You need visibility into what your AI agents are actually doing. We track:

• Tasks completed per agent
• Human override rate
• Quality scores on agent outputs
• Business impact metrics

This isn’t about micromanaging AI—it’s about understanding their impact and optimizing their contribution.

The Strategic Advantage (If You Get This Right)

Here’s what makes this worth the complexity: Companies that figure out AI agent management will have an insurmountable competitive advantage.

Our 11 AI agents effectively give us the analytical capacity of a 50-person team. They work through weekends, analyze competitor moves in real-time, and identify opportunities while our competitors sleep.

But only if we manage them properly.

The companies that deploy AI agents without management frameworks will drown in their output. The companies that figure out human-AI collaboration will dominate their markets.

Your Action Plan

If you’re considering AI agents (and you should be), start here:

1. Start small: Deploy 1-2 agents in non-critical areas first
2. Build management infrastructure: Triage systems, dashboards, protocols
3. Train your team: AI agent management is a new skill set
4. Measure everything: Track agent performance and human overhead
5. Iterate quickly: AI agent management is still evolving

The Bottom Line

The future of SaaS isn’t about working harder—it’s about working smarter with AI agents that never stop. But managing always-on AI workers is harder than managing humans.

The companies that crack this code will scale faster, identify opportunities quicker, and execute with precision that their competitors can’t match.

The companies that don’t will burn out their human teams trying to keep up with their own AI agents.

The race isn’t to deploy AI agents first. It’s to manage them best.

Now you will know >exactly< what VCs will think of your pitch deck.

Try it here.

The Problem Every B2B Founder Faces

Here’s a brutal truth: 95% of pitch decks that land in VC inboxes are unfundable. Not necessarily because the underlying businesses are bad, but because founders consistently make the same preventable mistakes that kill investor interest before they even get to the metrics.

I’ve reviewed thousands of pitch decks over the years. The patterns are painfully predictable:

• Rambling 5-10 slide intros about “how the cloud is changing everything”
• Missing use of funds slides (instant red flag)
• Vague competition analysis that screams “we don’t really understand our market”
• Financial projections that would make a CFO laugh
• Metrics that make no sense or are missing
• Zero clarity on the path to $200M ARR

The worst part? Most founders have no idea their deck (and the metrics and financials in it) is the problem. They think it’s the market, the timing, or just “tough fundraising conditions.”

That ends today.

Introducing SaaStr.ai’s VC Pitch Deck Review Tool

After processing over 139,000 conversations with B2B founders and analyzing the patterns from our portfolio companies (including exits like Salesloft at $2.3B, Pipedrive at $1.5B, and Algolia at $2.3B), we’ve built something unprecedented:

An AI that thinks like a VC, trained specifically on what actually gets B2B and B2B+AI companies funded.

This isn’t another generic pitch deck template generator. This is a sophisticated analysis engine that evaluates your deck through the lens of actual venture capital decision-making, with deep SaaS domain expertise baked in.

Core Features That Actually Matter

1. Comprehensive Scoring System

The tool provides three critical scores that VCs actually care about:

Traction Score (0-100): Measures the strength of your growth metrics, customer validation, and market momentum. The screenshot shows a score of 35 with specific feedback on revenue growth and market response.

Deck Quality Score (0-100): Evaluates story flow, slide effectiveness, and investor-readiness. A score of 48 indicates significant room for improvement in presentation structure.

Investment Grade (A+ to F): The bottom-line assessment that determines if your deck would actually get a second meeting. A C- grade means “interesting but needs major work before funding.”

And when it needs to be — it will be brutally honest.

2. Detailed Breakdown Analysis

The tool dissects your deck across multiple dimensions:

Traction Breakdown:

• Team Quality assessment
• Market Size validation
• Traction Metrics analysis
• Competitive Edge evaluation
• Growth Strategy review

Deck Quality Breakdown:

• Team Quality presentation
• Market Size clarity
• Competitive Advantage positioning
• Growth Strategy articulation
• Market Heat assessment
• Traction Metrics presentation

Each category gets scored individually with specific improvement recommendations.

3. Priority Improvement Recommendations

Rather than generic advice, the tool identifies your highest-impact fixes first. The example shows specific guidance on addressing customer validation gaps and improving differentiation messaging.

4. Key Issues Detection

The AI automatically flags deal-killers that VCs see repeatedly:

• Financial gaps (missing 12% revenue growth targets)
• Weak differentiation vs. competitors
• Execution risks with team/customer fit

5. Actionable Metrics Summary

Instead of hiding behind vague statements, the tool demands transparency on the metrics that actually matter:

• Current ARR/MRR
• Growth Rate
• Customer Count
• Market Size calculations
• Team Size

6. Competitive Intelligence

The tool benchmarks your positioning against successful companies in your space, identifying white space opportunities and competitive vulnerabilities.

Advanced Features for Serious Founders

Financial Intelligence Module

• Current ARR analysis ($1.3M in the example)
• Growth Rate assessment (60% annually)
• Revenue Model validation
• Unit economics review

Detailed Score Breakdown

Granular analysis across:

• Team Quality (scored individually)
• Market Size validation
• Competitive Advantage strength
• Growth Strategy viability
• Market Heat assessment

Business Intelligence Dashboard

• Market Size analysis ($400B TAM)
• Company Type classification
• Funding Stage recommendations
• Investment insights

Complete Business Analysis

The Free (for now) premium version provides comprehensive business intelligence including competitive analysis, market dynamics assessment, and investment opportunity scoring.

Why This Tool Exists (And Why It’s Free)

Here’s the thing: bad pitch decks waste everyone’s time. VCs spend countless hours on decks that never should have been sent. Founders waste months in fundraising cycles that were doomed from slide one.

This tool fixes the efficiency problem on both sides.

For Founders: Get instant, brutally honest feedback based on what VCs actually think (but rarely tell you). Fix the obvious problems before you ever hit send.

For VCs: Receive higher-quality decks from founders who’ve done their homework and understand what investors actually want to see.

It’s a positive-sum game for the entire ecosystem.

The SaaStr.ai Difference: Why This Tool Actually Works

Most pitch deck tools are built by people who’ve never actually invested in or built a successful SaaS company. They’re essentially sophisticated template generators.

This tool is different because it’s trained on:

1. Real Portfolio Data: Patterns from actual SaaStr Fund investments that generated $10B+ in exits
2. 139,000+ Founder Conversations: Real questions, real problems, real solutions
3. Jason Lemkin’s Brain: 15+ years of SaaS investing and operating experience, from EchoSign (acquired by Adobe) to leading investments in unicorns
4. The Minds of 800+ VCs In SaaStr Ecosystem.  We’ve “RAG”‘d the sessions, brains, feedback and workshops from 100s of VCs in SaaStr ecosystem, from Bessemer to Accel to YCombinator and more to include all their minds and thoughts in your feedback.
5. Live VC Feedback Loops: Continuously updated based on what’s actually working in the current funding environment

The Training Data That Matters

This AI has been trained on the specific frameworks that work:

• The “one-slide deck” principle that distills your entire story
• The metrics summary approach that puts everything on one transparent slide
• The 12-18 month planning clarity that separates real operators from dreamers
• The bottom-up path to $200M+ ARR that VCs need to see

Real Talk: What This Tool Won’t Do

This is not a magic bullet.

If your underlying business fundamentals are broken, no amount of deck polish will fix that. The tool won’t:

• Generate fake traction metrics
• Hide fundamental business model problems
• Replace the need for real customer validation
• Overcome team or market timing issues

What it will do is ensure that a good business gets presented properly, and that obvious deck mistakes don’t kill your fundraising before it starts.

How to Use It (The Right Way)

1. Upload your current deck – don’t “perfect” it first, use your real working version
2. Review the scores honestly – if you’re getting C- grades, you’re not ready to fundraise yet
3. Fix the Priority Improvements first – tackle the highest-impact changes before anything else
4. Re-run the analysis – iterate until you’re consistently hitting B+ or better
5. Only then start your fundraising process – you finally have honest feedback to use before that first pitch.

Pro Tips for Maximum Impact

• Don’t argue with the feedback: The AI is trained on what actually works, not what you think should work
• Focus on metrics transparency: The tool rewards honesty and punishes hiding key numbers
• Iterate quickly: Run your deck through multiple times as you refine it
• Use it early: Don’t wait until you’re ready to send to VCs – use it during deck creation

It’s Your New Cheat Code to Making It Easier — And Less Frustrating — To Get Funded

Fundraising is hard enough without shooting yourself in the foot with a broken pitch deck. This tool eliminates the unforced errors that kill most fundraising efforts before they begin.

The best part? It’s completely free.

Why? Because better pitch decks create a better funding ecosystem for everyone. VCs get higher-quality deal flow. Founders waste less time on broken fundraising strategies. The whole system becomes more efficient.

Get Started Today

Ready to find out if your pitch deck is actually investor-ready?

Access the SaaStr.ai VC Pitch Deck Review Tool at: saastr.ai/pitch-deck-analyzer

Upload your deck, get your scores, and start fixing the problems that are killing your fundraising efforts.

Remember: In fundraising, you don’t get points for effort. You get points for results. This tool helps ensure your deck delivers the results you need.

• A FREE start-up valuation calculator here
• A FREE VC pitch deck review here.  It’s awesome.
• An entirely new SaaStr website at SaaStr.ai
• And more

We couldn’t have really done any of these without vibe coding.  Not really.

And vibe coding platforms and no-code tools are getting better every week. And not just to help devs — Replit, Lovable and more have raced to $500m+ in new ARR just the first months of the year alone, in larger part focused on ‘prosumers’ and non-developers trying to put B2B apps into production.

And because of it, everyone thinks they can build the next Notion or HubSpot from their laptop in just an hour or so.  Many claim they even have.

And the vibe is electric.

You really can just type in a prompt what app you want to build, and a prototype will come out in minutes that looks mighty cool.  On the surface at least.

But here’s what almost nobody’s talking about in all those tweets: Security is the blocker in the end for many “vibed” B2B apps becoming production grade.  Especially if you want to collect secure information, store it, etc.

The Issues I’ve Already Seen.  And They Are Real Ones.

I’m about 150+ hours into vibe coding several apps, including the new SaaStr homepage at SaaStr.ai.

Here are the security issues I’ve already seen.

These are just the ones I’ve seen, and it’s a partial list:

• User Enumeration: Se quential user IDs that let you iterate through every user in the system. User 1, User 2, User 3… it’s a basic security issue.  If someone gained access to your system, they could access user by user by typing in user/123 user/124 user /125 etc. No one builds this way anymore, but Claude does, so the ‘prosumer’ vibe platforms do.
• Email Leakage: Group emails showing up in the To: line, exposing who else is using the platform. The issues I’ve seen have been limited in scope, but they could have been much larger in full production.
• Broken Access Controls: The ability of one user to see another user’s data due to faulty access controls and log-in logic.  This is a big issue. And one I’ve seen more than once.  That should give anyone pause.
• Dev/Prod Database Co-mingled: Development and production sharing the same database. Test users mixed with real customers. These are real issues and not one any real commercial app would have.  Some progress has been made here by leading vendors, but the default is still to combine data in one database.
• Plain Text Storage of Private Keys: API keys, database credentials, third-party secrets stored in plain text. In the app. The leaders have added security scans that can pick much of this up, but it remains a big and real issue.  An app I just build again the other day once again stored secret keys in plain text. Yes, the scanner caught it.  But it shouldn’t have happened — again.
• Session Management Flaws: Password protection that you can bypass by navigating directly to a protected page. Or clearing cookies. Or opening an incognito window.
• Limited Database Encryption: The ‘prosumer’ apps do encrypt data at rest, which is good, but that’s not enough for everything.  There is no default protection at column level or otherwise.  In the end, customer data, PII, etc. is sitting in plain text in the database. This may be deemed ‘OK’ for many apps by developers, but it’s still an issue.
• SSO Integration Failures: Use the default SSO the vibe coding apps offer, because when I’ve tried to implement third party SSO from Google to LinekdIn — it often doesn’t actually authenticate. Or validates against the wrong tenant (yikes!). Or both.
• AI Agent Rewriting Code.  This is perhaps the biggest ‘meta’ issue.  Every time you log into the AI agent, it can and might rewrite code you thought was ‘secure’.  Even for seemingly small matters or fixes.

This isn’t just “a few bad apples” or minor issues.  Not is it as one very senior exec at a leading vibe coding app called it “just security stuff, it happens.”

This are systemic, material security issue when using Claude to write code quickly. And to some extent, this is what happens when you optimize for speed and skip the security fundamentals.  And it’s still happening to me.  Even with lots of improvements from the leading vendors.

And it’s true at every ‘prosumer’ vibe coding platform.  It’s not unique to any one of them.

The Ongoing Security Evolution

Leading prosumer platforms are making rapid, real progress.

They get more and more secure every week, and I’m confident many of these issues will be resolved in coming months.  Lovable has just hired a cracked security team, Replit has added built in tools to enhance security.

But not all of the issues for a truly safe B2B production have been resolved.  Not if in the end of the day, they are mostly just using Claude to write whatever code … Claude wants to write.  And whatever corners Claude wants to cut.  The platforms will work around the corner cutting, add more and more guardrails, and add more security.  But Claude alone cannot be trusted.  The underlying platforms (Claude + OpenAI agents) cannot be trusted to build secure software.  That is in their goal seeking natures.

And importantly for folks building a commercial-grade B2B app without a developer” security is never finished. And it’s always stressful. Every new feature introduces new attack vectors. Every integration creates new vulnerabilities. Hackers don’t take breaks while you’re shipping features.

The major B2B and SaaS platforms understand this. They have dedicated security teams working full-time on threats that don’t even exist yet. They’re not just patching known vulnerabilities — they’re anticipating unknown ones.

You might think Squarespace or Shopify are seemingly simply platforms.  They aren’t under the hood.  And one thing they have huge, huge teams working on is security.  So you don’t have to worry.

When you vibe it on your own? All of a sudden those security concerns are on your back.

Most prosumer developers are still in reactive mode. Build first, secure later. That approach works for weekend projects, but not for business-critical applications.

The big question in many ways is — whose fault is it?  Can we expect the ‘prosumer’ vibe leaders to be as secure as Shopify and Squarespace?  I say Yes, since their marketing claims as much.  They all claim you can vibe code an app in minutes.  From one prompt.  Shouldn’t enterprise-grade, or least Shopify-grade, security be part of that?

Why “Junior Devs Would Make The Same Mistake” Isn’t Good Enough

Could a junior developer make these same mistakes? Absolutely.  Probably every developer has made most of the mistakes on the list above.

But junior developers don’t usually ship to production without oversight. They have senior developers reviewing their code. They have security teams running audits. They have established processes and frameworks that catch these issues.

Prosumer ‘developers’? They’re flying solo. No code review. No security audit. No established patterns. Just ship fast and figure it out later.  Most don’t even know what a security audit is, or what the most common issues are.  Let alone to look for them.  Let alone that they even have to, or should.

The Real Competition Isn’t Other Prosumer Tools

Everyone’s comparing their tool to other no-code platforms.  Replit vs. Lovable vs. Bolt is fun to watch.

But that’s the wrong comparison.

The real competition is Shopify and Squarespace to build. And HubSpot and Notion to buy.

These companies employ hundreds of security engineers. They spend millions on penetration testing. They have dedicated compliance teams for SOC2, GDPR, HIPAA. They have bug bounty programs where researchers hunt for vulnerabilities full-time.

When a customer chooses your prosumer app over HubSpot, they’re not just choosing features. They’re choosing to trust you with their business data instead of a company that’s invested decades and hundreds of millions in security infrastructure.

That’s a massive responsibility.  And one most ‘prosumers’ aren’t equipped to take on.

Vibe Coding is the Future. But “Roll Your Own?” That’s More Complicated.

All The Marketing is Ahead of Reality, Especially in Security

The prosumer coding dream is intoxicating:

• “Build exactly what you need”
• “No vendor lock-in”
• “Ship in days, not months”
• “Total control over your data”

Even Microsoft and Google make this claim.  Not just start-ups.

Even GitHub says you can now dream it in a single click.  That’s … aggressive at best.  I honestly can’t believe Microsoft lawyers would ever really allow it.  They probably had to under pressure.  Because Lovable, Replit, etc. plus Cursor and Claude Code for devs are growing at an insane pace.

The security reality is sobering:

• You’re responsible for protecting customer PII
• One breach can destroy your business (and possibly your customers’)
• Security isn’t a feature you add later
• Security isn’t something most ‘prosumers’ even understand, but compliance isn’t optional for commercial B2B software

Why “Roll Your Own” Isn’t Ready Yet For Paid Commercial Apps.  At Least, Not In Many Cases.

To be clear: I love vibe coding. The tooling is impressive. The velocity is real. The customization possibilities are endless.

But we’re still in the very early innings.

Security-first frameworks don’t fully exist yet. There’s no “Rails but for prosumer apps” that bakes in authentication, authorization, encryption, and compliance by default.  At least, not enough of it.  Not Shopify-grade.

The current prosumer stack optimizes for building fast, not building securely. And until that changes, most prosumer apps are ticking time bombs.

The Path Forward

What would make prosumer development actually viable for commercial business applications?

• Security-First Frameworks: No-code/low-code platforms that make the secure choice the default choice. Where you have to actively opt out of encryption, proper session management, and access controls.
• Built-in Compliance: Platforms that handle SOC2, GDPR, HIPAA compliance automatically. Where data handling, retention, and deletion policies are configuration, not custom code.
• Security Auditing Tools: Automated scanning that catches the common vulnerabilities before they hit production. Some of the platforms do have this now, which is great. They have to keep going further.
• Education and Standards: Security training specifically for prosumer developers. Common patterns and anti-patterns. A culture that values security as much as shipping speed.

‘Prosumer’ Vibe Coding is Huge.  It Will Get Better.  But It’s Not Secure Enough — Yet.

The prosumer development wave is real and it’s not going away. The tools will keep getting better. The barrier to building software will keep dropping.

But until security becomes a first-class citizen in the prosumer stack, most “roll your own” projects remain limited as commercial, paid apps.

Your customers trust you with their data. Security isn’t a one-time implementation — it’s an ongoing discipline. The threat landscape evolves daily. What was secure yesterday might be vulnerable tomorrow.

The prosumer dream is exciting. But excitement doesn’t protect customer data.

Continuous, disciplined security practices do.

And a deep dive on this and more on the top 10 things to think about before you start vibe coding your own B2B app here:

“We just have not had the people,” Benioff admitted. Now their new agentic sales system is “calling everyone back and having conversations.”  It has already closed over a million in deals — on its own.  And it’s just getting started.

And they’ll be rolling out this functionality to their customers starting at Dreamforce.

Meanwhile, SaaStr itself achieved the #1 response rates on their AI SDR platform after sending 4,495 hyper-personalized messages in two weeks.  And it’s now fully replaced our human AI SDRs.

The verdict? AI SDRs absolutely do work—but only if you’re willing to do the hard work of training them right.  Both up front, and day-in and day-out.

Most of those who say AI SDRs don’t work … haven’t really done both.  Or not fully.

To Work, AI SDRs Need Human Orchestration

AI SDRs can solve previously impossible scale problems, but success requires intensive human orchestration.

• Benioff’s Scale Solution: “Over the last 26 years, Salesforce has had more than 100 million people contact us that we’ve not been able to call back.” AI SDRs solved a capacity constraint that even 15,000 salespeople couldn’t address, proving AI works best on volume problems that humans simply cannot handle.
• SaaStr’s Execution Playbook: Achieved #1 response rates by treating AI SDR deployment like hiring a human—two weeks of intensive training, daily auditing, and continuous improvement. The key insight: “It’s more work, not less—but higher quality output.”
• The Training Reality: Both implementations required significant upfront investment. Benioff emphasizes the “omni-channel supervisor” coordinating human-AI workflows, while SaaStr spent 90 minutes daily for months perfecting their AI’s responses.
• Value Creation Over Volume: Success depends on AI providing genuine value rather than just sending more emails. SaaStr’s AI references specific event attendance and role changes, while Salesforce’s system handles genuine prospect conversations rather than generic outreach.

The 100 Million Lead Challenge: Benioff’s Ultimate AI SDR Case Study

When Benioff revealed Salesforce’s 100 million uncontacted leads problem, he provided the most compelling AI SDR use case imaginable. This wasn’t about optimizing existing processes—it was about solving a problem that was literally impossible with human resources alone.

“We have like 15,000 sales people. We don’t have that many SDRs,” Benioff explained. The math was brutal: 100 million leads over 26 years represents nearly 4 million annually that went completely uncontacted due to human capacity constraints.

The AI Solution Architecture

Benioff’s implementation centers on what he calls an “omni-channel supervisor”—an AI orchestration system that coordinates between human agents and digital agents. This isn’t simple automation; it’s intelligent workflow management that determines when human intervention is required and when AI can handle interactions independently.

The system handles:

• Automated outreach to the massive backlog of uncontacted leads
• Conversation management through natural language interactions
• Qualification and routing based on engagement patterns
• Seamless handoffs to human SDRs when complexity thresholds are reached
• Integration with existing CRM and sales processes

“This agentic sales is calling everyone back and having conversations with them and then deeply integrating it through the omni-channel supervisor into our new agentic sales product,” Benioff noted.

The Economic Impact

The revenue implications are staggering. If even 1% of those 100 million leads could convert at Salesforce’s average deal size, that represents millions in previously inaccessible revenue annually. This demonstrates AI’s true value proposition: expanding addressable market rather than just optimizing existing processes.

More importantly, Benioff positioned this as workforce optimization rather than replacement: “I’ve been able to take that headcount and then rebalance it into other parts of my company where I need more help and need more support because we’re still growing.”

SaaStr’s Real-World AI SDR Playbook

While Benioff operates at massive scale, SaaStr’s experience provides tactical insights for companies looking to implement AI SDRs effectively. Their results were impressive: #1 response rates on their platform within 30 days, with 4,495 hyper-personalized messages sent in just two weeks.

The Training Investment Reality

SaaStr’s most important insight contradicts the “turn it on and watch magic happen” narrative. Their approach required:

• Two weeks of intensive training (90 minutes morning, 1 hour evening, plus real-time responses)
• Daily auditing of 30-45 minutes for the first 60 days
• Continuous optimization based on response quality and recipient feedback
• Deep data integration across CRM, marketing automation, and content systems

“Expect to spend the same time training AI as you would a human,” noted Amelia Lerutte, SaaStr’s SVP. “This isn’t a ‘set it and forget it’ solution.”

The Personalization That Actually Works

SaaStr’s success came from genuine personalization rather than mail-merge tactics:

Bad AI Email: “Hey [FIRST NAME], I did some research on [COMPANY] and thought you might be interested in [GENERIC PITCH].”

Good AI Email: “Hi [NAME], saw you attended SaaStr London last year and just noticed your move to [NEW COMPANY] — congrats! Given [COMPANY]’s focus on [SPECIFIC AREA], thought you might be interested in our 2025 London program, especially our new VC track…”

The difference is specificity and relevance. SaaStr’s AI references actual event attendance, congratulates on role changes found via LinkedIn, and suggests relevant programs based on company profiles.

The Human-AI Orchestration Model

Both implementations emphasize human oversight rather than full automation. Benioff’s “omni-channel supervisor” and SaaStr’s daily auditing represent different approaches to the same principle: AI handles volume while humans ensure quality and handle complex interactions.

SaaStr’s Human-in-the-Loop Framework:

• Real-time response management: When prospects reply to AI, humans must respond immediately at the same quality level
• Daily quality auditing: Review sample outputs and provide corrective feedback
• Continuous training: Teach the AI why responses were wrong and what to do instead
• Escalation handling: Manage conversations that exceed AI capability thresholds

Salesforce’s Orchestration Layer:

• Intelligent routing: Determine appropriate human vs. AI handling based on conversation complexity
• Context preservation: Maintain conversation history across human and AI interactions
• Escalation triggers: Seamlessly hand off to humans when empathy or complex problem-solving is required
• Integration management: Ensure AI conversations feed properly into existing sales processes

The Data Foundation Requirements

Both implementations required massive data integration efforts. Benioff emphasized that AI accuracy depends on comprehensive data cloud integration: “Our AI is part and parcel with our data cloud… so that you can get all your data harmonized in one place.”

SaaStr trained their AI on:

• 20+ million words of SaaStr content
• 10+ years of CRM and marketing automation data
• Website behavior and event attendance history
• Social media profiles and job change information
• Engagement patterns across all touchpoints

The data cleaning revelation was particularly important: “Your data probably isn’t as clean as you think. We found opportunities that were never logged in Salesforce, missing context from AEs who never used the system properly, and gaps everywhere.”

Campaign Segmentation and Performance

SaaStr’s segmentation approach provides a framework for AI SDR campaign strategy:

High-Performing Segments:

• Reactivation campaigns: Previous customers or engaged prospects who went dark
• Event follow-up: Attendees from past events who haven’t returned
• Website visitor follow-up: People who showed recent engagement

Lower-Performing Segments:

• Cold outbound: Generic prospecting to unengaged lists
• Broad targeting: Campaigns without specific behavioral triggers

The insight: AI SDRs work best when leveraging existing relationships or demonstrated interest rather than purely cold prospecting.

The Technology Stack Reality

Implementation requires integration across multiple platforms:

SaaStr’s Tool Stack:

• AI SDR Platform: Artisan, Qualified
• Data Enrichment: Lusha, Seamless, ZoomInfo, Apollo
• Dynamic Content Creation: Gamma, Genspark for custom presentations
• Call Intelligence: Claude/Perplexity for research, Cluely for real-time insights
• Training Data: CRM, marketing automation, content management systems

Salesforce’s Integration Approach:

• Data Cloud: Federated data sources across the enterprise
• AgentForce: AI agent platform managing conversations
• CRM Integration: Seamless workflow between AI and human activities
• Omni-channel Supervisor: Orchestration layer managing human-AI coordination

Success Metrics and Expectations

Both implementations focus on business outcomes rather than AI-specific metrics:

Salesforce Metrics:

• Volume handling: 100 million previously uncontacted leads now receiving outreach
• Workforce optimization: 4,000 support agents redeployed to higher-value activities
• Revenue impact: Expansion of addressable market through increased capacity
• Integration success: AI conversations feeding seamlessly into sales processes

SaaStr Metrics:

• Response rates: #1 performance on their AI SDR platform
• Meeting booking: Qualified prospects scheduling sales conversations
• Reactivation success: Re-engaging lapsed accounts and attendees
• Quality maintenance: Responses indistinguishable from human-written emails

Common Implementation Mistakes

Both experiences reveal frequent AI SDR deployment errors:

Volume Over Value: Focusing on email quantity rather than recipient value creation Insufficient Training: Expecting immediate results without intensive upfront investment Poor Data Foundation: Attempting AI deployment before cleaning and integrating data sources Lack of Human Oversight: Treating AI as fully autonomous rather than requiring daily management Generic Personalization: Using basic mail-merge tactics instead of genuine relevance Inadequate Segmentation: Applying one-size-fits-all approaches across different prospect types

The Economic Reality Check

Benioff’s workforce redeployment model and SaaStr’s intensive management requirements both point to an important economic reality: successful AI SDR implementation requires significant upfront investment but can deliver returns at previously impossible scales.

Investment Requirements:

• Training time: 2-3 weeks intensive setup, ongoing daily management
• Data integration: Comprehensive cleanup and platform connections
• Human resources: Dedicated AI orchestration rather than casual oversight
• Technology stack: Multiple integrated platforms beyond just the AI SDR tool

Return Potential:

• Scale expansion: Handle volumes impossible with human capacity alone
• Quality improvement: Hyper-personalization at scale
• Workforce optimization: Redeploy human resources to higher-value activities
• Market expansion: Reach previously inaccessible prospects and segments

The Future of AI SDR Implementation

Both Benioff and SaaStr are doubling down on AI SDR capabilities, suggesting this is just the beginning of a broader transformation:

Salesforce’s Evolution:

• Integration across all products: “I don’t think that there will be a piece of software that we sell that will not be agentic”
• Expansion beyond sales into support, marketing, and operations
• Deeper AI-human collaboration models across the entire customer lifecycle

SaaStr’s Expansion:

• “Literally onboarding two more AI sales tools this week”
• Hiring dedicated AI operations roles: “A human whose entire job is orchestrating these systems”
• Goal: “AI touching every part of our sales and marketing funnel by the end of the year”

Key Takeaways

1. Scale Problems Are Perfect AI Use Cases
   • Benioff’s 100M leads represents problems that are humanly impossible to solve
   • AI works best on volume constraints rather than efficiency optimization
   • Focus on capacity expansion rather than cost reduction
2. Training Investment Equals Human Onboarding
   • Plan 2-3 weeks intensive setup plus ongoing daily management
   • Expect to spend 30-90 minutes daily on quality control and optimization
   • Success requires dedicated human orchestration, not casual oversight
3. Data Foundation Is Critical
   • Clean and integrate all relevant data sources before implementation
   • Expect data quality issues that weren’t apparent in manual processes
   • AI accuracy depends on comprehensive, harmonized data access
4. Personalization Requires Genuine Value
   • Move beyond mail-merge tactics to specific, relevant value creation
   • Reference actual behaviors, events, and context rather than generic research
   • Test whether you would send the same email manually
5. Human-AI Collaboration Is Essential
   • Design orchestration systems that determine appropriate AI vs. human handling
   • Maintain seamless handoffs and context preservation across interactions
   • Plan for real-time human response when prospects engage with AI outreach

Quotable Moments

Benioff on the scale challenge: “Over the last 26 years, Salesforce has had more than 100 million people contact us that we’ve not been able to call back. We just have not had the people.”

Benioff on workforce evolution: “I’ve been able to take that headcount and then rebalance it into other parts of my company where I need more help and need more support because we’re still growing.”

SaaStr on training reality: “Expect to spend the same time training AI as you would a human. This isn’t a ‘set it and forget it’ solution.”

SaaStr on execution requirements: “It’s more work, not less—but higher quality output. You get 10x better output, but it requires ‘S-tier human orchestration’ to get top-tier results.”

Benioff on the future: “I don’t think that there will be a piece of software that we sell that will not be agentic.”

The evidence is clear: AI SDRs work, but only when implemented with the same rigor and investment as hiring exceptional human SDRs. The companies that understand this reality—and are willing to make the upfront investment in training, data integration, and human orchestration—will capture the massive advantages that AI-human collaboration can deliver.

The Latest 20VC+SaaStr: Benioff Joins — And Delivers $1B+ AI Revenue; Anthropic Demand is Insatiable; AI Following Up With 1,000,000+ Leads at Salesforce

Salesforce for example has already repurposed headcount from 1000s of support agents into sales.  Same headcount, but different roles.

But that’s not how AI will change tech headcount and jogs the most.  There’s a much bigger role for AI to play, and problem to solve: we don’t have enough humans to do the actual work that drives B2B growth.

Not the glamorous work. Not the “strategic” work.  Not managing teams. But the grinding, repetitive, essential work that separates successful B2B startups from the rest.

Even again Salesforce itself had 1,000,000+ leads its own human SDRs didn’t want to, or didn’t think they had time to … follow up on:

The biggest change for AI + roles in tech isn’t about layoffs or replacing talent.

This is about acknowledging a fundamental mismatch between what needs to get done and what humans are truly willing to do at scale.

The Work No One Wants to Own or Really Do Well.  If At ALll.

Walk into any B2B company and you’ll find these gaps everywhere:

Sales Development That Actually Works Everyone talks about personalized outreach, but who wants to research 200 prospects a week and craft truly customized emails? The answer is almost no one. The best SDRs burn out. The mediocre ones send templates. The result? Most outbound efforts fail because the volume of quality work required exceeds human capacity and motivation.  Even at Salesforce, 100,000,000+ leads were never followed up on during its 2.5 decades in business.  100 million.

Following Up on B and C Leads Your sales team focuses on A leads. That’s rational. But those B and C leads sitting in your CRM? They represent millions in potential revenue, but following up properly requires hundreds of touches across months. Sales reps won’t do it consistently. It’s not rewarding enough.

Customer Winback at Scale You can recover 10-15% of churned customers with the right approach. But that means systematic outreach across multiple channels over extended periods. Who’s going to manage 500 different winback sequences simultaneously? No one raises their hand for that assignment.

Real Onboarding  Not the automated email sequence you call onboarding. Real onboarding means checking in, troubleshooting, hand-holding, and ensuring actual adoption. It requires patience, repetition, and attention to detail that doesn’t scale with human teams.

Small Deal Velocity Your AEs want to close six-figure deals. But what about the $1,000 monthly subscriptions that add up to millions in ARR? These deals need attention, follow-up, and care, but they don’t provide the dopamine hit or commission structure that motivates top performers.

Building the Boring Applications Every engineer wants to work on the core product, the AI features, the sexy infrastructure. But someone needs to build the admin panels, the reporting dashboards, the integration tools. These are table stakes for enterprise customers, but they’re career dead-ends for ambitious developers.

Actually Picking Up the Phone Cold calling works. Everyone knows it works. But who wants to make 50 calls a day and get hung up on 47 times? Even experienced sales professionals avoid it when possible.

Why This Isn’t About Skills, Re-Skilling or Hiring

This isn’t a talent shortage problem you can solve by raising salaries or improving benefits. These are fundamental human nature issues:

• Repetitive work is psychologically draining
• Low-reward activities get deprioritized
• Humans optimize for immediate feedback and recognition
• Scale requirements exceed individual capacity
• We all got used to working a certain way from home from 2020-2022
• Many of the next generation is just less interested

You can hire more people, but you’ll get the same patterns. The work still won’t get done consistently at the level needed for competitive advantage.

AI Bails Us Out.  And … Will Let Us Fly

The current generation of AI tools is already handling pieces of this puzzle. But we’re moving toward something more comprehensive:

Intelligent Sales Development AI that researches prospects, identifies trigger events, crafts personalized outreach, and manages multi-touch sequences at infinite scale. Not templates pretending to be personal, but actually relevant communication based on real insights.

Systematic Lead Nurturing AI managing hundreds of nurture tracks simultaneously, identifying when prospects show renewed interest, and escalating to humans at the optimal moment.

Customer Success That Scales AI monitoring product usage, identifying at-risk accounts, managing onboarding flows, and ensuring customers hit their first value milestone before a human ever needs to intervene.

Automated Deal Progression AI handling the discovery, demo scheduling, proposal generation, and follow-up for smaller deals, only involving human sales reps when contracts are ready to close.

The Flywheel Effect

When AI handles the work humans avoid, several things happen:

Your human team focuses on high-value activities they actually want to do. Sales reps spend time with qualified prospects instead of cold calling. Engineers build features instead of admin tools. Customer success managers handle strategic accounts instead of routine check-ins.

You achieve consistent execution across all these unglamorous but essential functions. Every lead gets followed up. Every customer gets proper onboarding. Every winback opportunity gets pursued.

Your competitive advantage compounds because you’re doing things at scale that competitors can’t sustain with human-only teams.

Implementation Reality

This transition isn’t happening overnight, but the pieces are falling into place rapidly:

• Sales automation tools are becoming genuinely intelligent
• Customer communication AI can handle complex, contextual conversations
• Code generation tools are tackling routine development work
• Voice AI is making phone-based outreach scalable again

Smart SaaS companies are already running experiments. They’re identifying the specific workflows where AI can take over the grinding work and free humans for strategy, relationship building, and complex problem-solving.

AI Does What We Need, But No One Wants To Do

The future of SaaS isn’t about AI replacing humans. It’s about AI doing the essential work that humans will do poorly or not at all because it’s repetitive, unrewarding, or doesn’t scale.

Companies that embrace this reality first will have operational advantages that compound over time. They’ll have cleaner data, more consistent processes, better follow-through, and human teams focused on work that actually drives satisfaction and retention.

The technology is almost there. The question is whether your organization is ready to let AI handle the jobs that, frankly, no one really wants to do anyway.

When that happens, everything else becomes possible.

Why We Have 11+ AI Agents Already in Production Ourselves

It’s true at tiny team SaaStr, too.  Too many core functions … humans didn’t want to do:

• We couldn’t get our SDRs to follow up on most potential prospects.  They didn’t see the return.
• We couldn’t get our content team to review 1,000s+ of submissions.  Too many to find the diamonds in the rough.
• We couldn’t get our sales team to chase lapsed sponsors and partners.  They didn’t see the ROI.
• We couldn’t get our AEs to do any prospecting at all.  Few want to.
• We couldn’t get the GTM team to follow up on leads instantly.
• We couldn’t get our sales team to update our collateral consistently.  Or to make custom pitch decks for sponsors (Gamma does this now).
• We couldn’t get the agencies that helped support ticket holders to really answer questions correctly, or promptly.
• We couldn’t get anyone to email out to SaaStr AI Summit / Annual attendees to get them to come back.  They were only willing to drop them all in a mass cadence.

So we replaced all of that headcount … with AI.  Not to save money.  But because no one would really do those critical roles.  We’d hire them, they’d quit or only do the parts of the jobs they wanted to.  That was OK in 2020-2021.  Not today.

The work just wasn’t getting done, no matter how much we paid or whom we hired.  In some cases, e.g. content review, AI did a better job than humans.  In other cases, it came up short of what a human did on our team.  But — it did the job.  That’s what matters.

More on what we’ve already done and learned here: