HACKER Q&A
📣 robeym

How to sell SaaS without AI features in 2026?


I'm a developer who built an ERP/CRM system for small manufacturers (https://www.paxerp.com). It does all the basics very well; financial reporting, lot tracking, production planning, shipping carrier integrations, the usual workflow stuff—but there's zero AI in it. It's just fast, clean, and solves real problems I saw working in manufacturing ERP systems. The product works well, customers really like it, but I have almost no sales experience.

Every SaaS founder seems to be talking about "AI-powered insights" and "intelligent automation." while... We just have a clean system that is fast and tries to stay out of the user's way.

For those who've sold B2B SaaS (especially to traditional industries like manufacturing):

- Is "no AI" actually a disadvantage, or does it not matter as much as I think?

- How do I communicate value when the value is "it's simple and fast, and your data is highly accessible" vs "revolutionary AI"?

- Should I be adding AI features just to check a marketing box, even if customers don't need them?

You can learn more about why I built this on the websites about page. But now I'm wondering if I'm fighting an uphill battle by not having the buzzwords everyone else does.

Any advice from founders who've been here?

TY

  👤 codingdave Accepted Answer ✓
It doesn't matter one bit what other SaaS founders are saying. What matters is what your customers are saying. Are they happy with the feature set? Then you are fine. Don't over-engineer a product just to keep up with founders who have nothing to do with your product or market.

Of course, if your customers start asking for such features, go ahead and respond to that demand.

👤 NTCTech
The pendulum is definitely swinging back.

In 2024, "AI" was the value prop. Now, for many enterprise buyers, it's becoming a liability (compliance risk, hallucinations, unpredictable costs).

If you are solving a high-friction, "boring" problem—like plumbing compliance, legacy database migration, or payroll—nobody cares if there is an LLM involved. In fact, marketing "Deterministic Output" (i.e., it does exactly what you tell it to do, every time) is starting to feel like a premium feature again compared to the probabilistic nature of GenAI agents.

Build for the pain, not the buzzword.