July 22, 2026

407 | The SaaSpocalypse: Why Startups Fail in 2026 | Brian Herr

The SaaS world is in the middle of a brutal reckoning. Products that looked innovative 18 months ago are quietly becoming redundant — not because markets disappeared, but because the floor rose. In this episode, Jeff Mains sits down with Brian Herr, a 30-year technology and SaaS veteran, to dissect what it actually takes to build a software business that survives — and wins — in the age of AI.

Brian brings sharp investor-grade thinking to the conversation, drawing on his work with startups, venture studios, and PE-backed companies. They cover the death of thin-wrapper SaaS, why blocking AI agents is a catastrophic mistake, how security and compliance have become unexpected competitive moats, and the critical distinction between a product that helps and one that solves. If you build software or provide services, this episode is non-negotiable.

Key Takeaways

4:08 — The value expectation from SaaS platforms is shifting fast. Thin wrappers around someone else's AI model have no future — customers will ask why they're paying when they can do it themselves.

4:53 — Companies that survive will be the ones that solve real problems, curate the right data, and give meaningful feedback — not just deliver a slick interface.

6:46 — Investor rubrics have changed. A key new question before committing capital: “Can this be replicated as a Claude skill or agent in six months?” If yes, it's not fundable.

7:39 — Where physical world meets digital data is a major investment magnet. These companies have stronger moats, are more AI-resistant, and occupy underserved territory.

13:53 — Natural language interfaces are no longer a differentiator — they're an expectation. And Brian's crystal ball: local on-device AI will push this even further into everyday life.

14:29 — Natural language is democratizing technology for older users. If you don't have a conversational interface, the market will pass you by.

20:23 — Agents are no longer just for technologists. CFOs and revenue officers are using them. Blocking agents is a strategic blunder — competitors are advertising agent compatibility while you're building walls.

21:00 — The smart play: figure out what people are doing with agents hitting your platform and monetize it. Blocking just pushes them to your API — or to a competitor.

21:20 — Every SaaS company needs a quarterly gut-check: What is my value? What do I do well? How do I evolve? A business plan from one year ago doesn't fit today's market.

33:04 — Security, compliance, and certifiability are the new defensible moat. You literally cannot vibe-code your way into SOC 2, HIPAA, or AI trust scores. That's the value story.

33:59 — The AIUC-1 framework is making AI applications insurable for the first time. MITRE has joined the consortium. If your SaaS uses AI, this becomes part of your trust story.

39:50 — The single most important product question: Does it help, or does it solve? Helpful gets cut from budgets. Essential doesn't.

43:09 — Going niche gives you orders-of-magnitude higher odds of success. Trying to do what everyone else is doing? Your chance of success drops to 13% or less.

47:37 — Brand trust and human relationships are more important than ever. People do business with people. When you become indifferent to your customers, you become a vendor. Vendors don't survive.

Tweetable Quotes

“If someone opened a fresh ChatGPT window right now and got roughly the same result your product delivers — would your customers notice the difference, or would they even care?” — Jeff Mains

“The thin wrappers aren't going to make it very long. What's going to survive is companies that still solve real problems, curate the right data, and give the right feedback.” — Brian Herr

“One of our investment rubrics now: Can this be turned into a Claude skill or agent in six months? If so, it doesn't make sense for us to invest.” — Brian Herr

“Natural language interfaces are now an expectation, not a differentiator. If you think you'll eventually get around to it, the market will pass you.” — Brian Herr

“Helpful solutions get cut from the budget first. Solutions that solve don't. Stop asking whether you can bolt on AI and start asking whether customers actually need YOUR data and process to make it work at all.” — Jeff Mains

“Agents are becoming for everyone — especially as the interface evolves. Blocking them is evolve or die.” — Brian Herr

“Figure out what people are doing with agents and monetize it. People will pay for it. By being a blocker, you're just pushing them to find another way.” — Brian Herr

“People do business with people. When you become indifferent to your customers, you stop being a partner and become a vendor. Vendors have a hard time surviving.” — Brian Herr

“Success is a journey, not an endpoint. The founders who make it understand you're going to be a little wrong — as long as you course correct in the right direction.” — Brian Herr

SaaS Leadership Lessons

1. Moat = Data + IP + Experience, Not Interface A beautiful UI sitting on top of a commodity model is not a business — it's a countdown clock. Your defensible moat is proprietary data, domain expertise, and institutional knowledge that competitors cannot prompt their way into.

2. Run a Quarterly Value Audit Especially in the $5M–$15M revenue range, ask yourself every quarter: Does my business plan still match the market? What do I do well, and how am I evolving? Founders who don't course-correct veer further off target every quarter until they no longer recognize where the target moved.

3. Embrace Agents as a Revenue Channel, Not a Threat Your API traffic spikes are signals, not attacks. When agents are hitting your platform, that's demand you haven't monetized yet. Build for agent access, charge for it, and let your competitors play defense while you build offense.

4. Compliance and Trust Are Your Unfair Advantage In a world where anyone can vibe-code a competitor over a weekend, the thing they cannot replicate is your certifications, your compliance posture, your years of regulated-market experience, and your insurance-grade AI trust scores (AIUC-1). Make this part of your sales story.

5. Help vs. Solve Is the Only Product Question That Matters Helpful products live in discretionary budgets — they're the first cut when times get hard. Products that solve real, urgent problems command non-negotiable budget lines. Every feature you build, every market you target: ask which one it is.

6. Stay a Partner, Never Become a Vendor When customers feel like a transaction to you, you become a commodity to them. In the $5M–$15M range, clients know your team personally — that trust is a competitive advantage. Build systems to maintain it as you scale, or risk waking up one day to find out you've been quietly moved to the vendor pile.