In 2023, an HR tech startup lost 800 monthly subscribers in three months — even as they kept shipping new features. A competitor had started offering the same functionality for free through a GPT-based plugin. The CEO's postmortem was blunt: "We were selling software. What our customers wanted was speed of hire."

The weight of that story doesn't come simply from "competition got fierce." It touches a more fundamental question for founders right now: what exactly are you supposed to be selling when software features are commoditizing faster than ever? The SaaS subscription model — pay monthly, ship updates, reduce churn — has been the standard playbook for startups over the past decade. Investors knew it, founders were trained on it. Now, with AI and large language models shaking the foundations of that equation, voices are growing louder that the pitch deck needs a rewrite.

Features Commoditize Faster Than You Can Ship Them

Ivan Nikkhoo, venture investor and co-founder of Navigate, laid out the challenge clearly in a recent piece for Crunchbase News. The argument is straightforward: AI has accelerated the commoditization of software features, making competitive advantage based on features alone increasingly difficult to sustain.

Several trends converged on the SaaS industry over the past two to three years. Revenue multiples for many SaaS companies fell to less than half their 2021 peaks. Time and again, a GPT-4-powered startup replicated in a team of two what an established SaaS company had built with ten. And customers started asking — more often, more sharply — "What actually changes in our business when we use this tool?"

What Nikkhoo warns against is the most common response founders reach for under this pressure. When software competition intensifies, the instinct is to bolt on consulting or managed services. That generates revenue in the short run — but it compresses margins and kills scalability. The moment a software company starts operating like an agency, investors pull out a different measuring stick: service multiples, not software multiples. The deeper problem is that this approach still doesn't answer the core question: why should customers keep paying?

Before You Accept the "SaaS Is Dead" Diagnosis

There are, of course, credible counterarguments. Enterprise SaaS players like Salesforce, ServiceNow, and Workday have actually sustained growth since the AI wave hit. They've integrated AI as a feature enhancer rather than treating it as an external threat — broadening their customer bases and growing ARR steadily. The evidence against a sweeping "SaaS is over" declaration is substantial.

Skepticism about outcome-based pricing is also warranted. How do you measure customer outcomes? How do you distinguish between poor results caused by the software versus poor results caused by how the customer operates? What do you show investors when revenue predictability gets murky? Outcome-based models are logically compelling, but actually implementing them requires contract design, measurement infrastructure, and organizational readiness — all at once. Some startups have rushed this transition and ended up with tangled cash flows. This isn't the right prescription for every founder, and that's worth acknowledging.

That said, Nikkhoo's core argument carries real weight. His claim isn't that SaaS is disappearing — it's that competing purely on features has hit a wall. Blurring that distinction leads to the wrong diagnosis and the wrong fix.

The Only Way to Prevent Churn Is to Embed in the Workflow

At the center of Nikkhoo's survival strategy is what he calls "workflow ownership." The idea isn't just to provide a useful tool — it's to get your software so deeply embedded in the daily processes customers must run through every day that switching becomes a different category of decision entirely.

The practical reason this matters: cloning a single feature and rebuilding a team's entire daily work process are completely different orders of decision-making. The first is something any well-funded competitor can do quickly. The second is something customers don't undertake lightly. Software embedded in the workflow sees lower churn, and it naturally supports usage-based or outcome-linked pricing — the more customers use it, the more revenue grows alongside them.

Finance thinking helps clarify the picture. When customers evaluate software, what they're actually calculating isn't the feature count. It's: "What return does paying $500 a month for this tool generate for our business?" A growing cohort of customers approaches this like a capital allocation decision — what measurable improvement does this cash outlay produce? A founder who opens the conversation in those terms has a completely different dialogue with a customer than one who leads with a feature list — even if they're selling the same product.

How you position AI also divides along this line. "Our product now includes AI" and "customers who use our product have cut average processing time by 40%" aren't just sentences of different lengths. The first adds another checkbox to the feature race. The second signals that you have the capability to measure and prove outcomes in your customers' actual work. That gap is, I'd argue, the most important fork in SaaS pitches right now.

Own the Workflow — Don't Just Sell the Software

For solo founders or small teams building software-based businesses, this shift arrives from a slightly different angle.

If you've heard directly why customers churn, start there. "The price is too high" is, in many cases, just another way of saying "the results aren't clear enough." Behind that complaint there's almost always a workflow that's still unsolved — and finding it is the clue to your product direction.

You also need to understand where your software actually sits in your customer's workday. If it's something they can take or leave, the cancellation decision will come fast the moment a comparable tool appears. Becoming the thing they have to open every morning requires working through, together with your customers, what happens when it isn't there. The bigger that disruption, the stronger your defensibility.

When you add AI features, ask the same question first: can you state in concrete numbers how much faster this AI feature handles which specific task for customers? If you can, that's the reason they keep paying. If you can't, the AI feature may generate a burst of curiosity, but it won't sustain retention.

The "deep moat" Nikkhoo emphasizes actually maps to a realistic path for small teams. Competing head-to-head with large players means losing on scale. Going narrow and deep — into a specific industry, a specific workflow, a specific customer segment — is how you build real defensibility. Go wide and you need scale; go narrow and deep and the depth itself becomes the asset.

SaaS isn't going away. But selling it by reciting a feature list is no longer a viable strategy. The founders who embed themselves in customers' workflows and can show the results in numbers are the ones who make it to the next stage. The moment the first page of your pitch deck is filled with proof of outcomes, the conversation you have with every customer changes entirely.