Why I'm Not Scared of AI Agents Eating My SaaS

Everyone keeps telling me AI agents are going to cannibalize SaaS. After 15 years in digital, I've heard this song before — and I'm not running.

The hype is real (but incomplete)

The Hacker News thread "AI agents are starting to eat SaaS" blew up a few months back and it's still being referenced in every founder Slack I'm in. The thesis is clean: if AI can orchestrate workflows on demand, why pay monthly for dedicated software? Large enterprises are already cloning tools internally instead of buying them. Indie SaaS founders are rattled.

And honestly? They should be paying attention. But panicking? That's a different story.

The threat is real — for a specific category of product. I'm talking about the dashboard wrappers. The tools that are basically a nicer UI on top of data you could get elsewhere. The SaaS that became a business because building it was hard, not because the workflow was genuinely unique. AI agents will commoditize those. Fast.

What I actually think is happening

I'm building four SaaS products right now — ListingVid (AI video for real estate agents), EST8 (real estate CRM), OhMyLead (lead gen for indie hackers), and AIAnswer.to. None of them are dashboards. All of them own a specific daily workflow for a specific type of person.

Here's the thing about a real estate agent in Milan or Madrid: they're not setting up AI agent workflows. They're listing 6 properties a week, drowning in WhatsApps, and they want a button that makes a polished video in 30 seconds. The friction of "configure your own agent pipeline" is infinite for them.

That's the gap AI agents can't close on their own: the last mile. Getting a non-technical user to adopt a working system, trust it, and make it part of their daily routine. That's distribution, not technology.

The indie hackers who will lose are the ones building generic AI wrappers with no workflow moat. The ones who will win are building opinionated, sticky tools for specific people with specific daily problems.

3 things worth taking away

• Lazy SaaS dies first — if your product is a dashboard with no real workflow lock-in, the agents thesis is a direct threat. Either go deeper or pivot.
• Daily use = survival — the stickiest products are ones used every single day, where your data accumulates and switching hurts. Build toward that.
• Agents are a tailwind, not a headwind — a 2-person team can now build what used to take 20. The barrier to building good software just dropped. Use that.

The move

Stop treating "AI will eat SaaS" as an existential threat and start treating it as a product strategy question. What would an AI agent actually struggle to replace about your tool? If the answer is "nothing," that's the problem — not the agents.

Build something people open every morning. Make the workflow theirs. The agents can wait.

---

Building in public and thinking out loud at @lmoncany. What's your take — is the agents-eat-SaaS thesis actually playing out in your market?