Your SaaS Product Was Built for Humans. AI Computer-Use Is About to Change That.

Six people. $75 million. A $500 million valuation. And the product? Teaching AI to use software the same way you do — by looking at the screen and clicking around.

Standard Intelligence just raised that round. And while most coverage focused on the funding milestone, the more interesting question is what it implies for everyone else building software products.

Because here’s the thing: every product decision you’ve ever made rested on a silent assumption. That the person using your software is a person.

Computer-use AI breaks that assumption at the root.

This isn’t about chatbots embedded in your product or API integrations that let AI access your data. Computer-use AI means a model can open your app, look at the UI, navigate menus, fill in forms, and complete workflows — without any native API, without custom connectors, without you doing anything. It interacts with your product the same way a human intern would on day one.

The implications for SaaS are weirder than most people are talking about.

Per-seat pricing assumes a seat is a person. Your pricing model is almost certainly built around counting humans. One login, one seat, one bill. But what happens when a company deploys 50 AI agents to use your product simultaneously? Do you charge per agent? Per task? Per outcome? The whole model starts to look like it was designed for a world that’s already changing underneath it.

Onboarding flows assume a learner. You spent real time on your first-run experience. Tooltips, progressive disclosure, empty states with helpful copy, guided checklists. All of it was designed to reduce cognitive load for a human who needs to learn. An AI agent doesn’t need the tour. It doesn’t get confused by a cluttered UI. In fact, an AI might be better served by a dense, information-rich interface than the simplified one you built for humans. The “intuitive” UI you were so proud of is now a design constraint.

Workflows optimized for human hands may be the wrong shape. Think about anything in your product that requires clicking through multiple confirmation dialogs, scrolling through long lists, or toggling between views to complete a task. You probably kept some of that friction intentionally — to make sure humans were paying attention, to reduce errors, to create audit trails. Computer-use AI will complete those steps faithfully. But the question is whether you designed those steps for humans and are now just hoping AI can navigate them, versus designing workflows that work well for both.

The non-obvious risk isn’t that AI will replace your users. It’s that the stuff you built as competitive advantages — the carefully designed UI, the guided flows, the deeply human experience — might be less defensible than you thought.

On the flip side, there’s a version of this where computer-use AI dramatically expands your addressable market. Workflows that were too complex or time-consuming for a human to do manually become suddenly automatable without any custom integration work. The long tail of tasks people avoided in your product because they were tedious? An AI agent will just… do them.

The more useful frame isn’t “will AI use my product” — it will, if it’s useful. The better question is: “have I accidentally built my product in a way that assumes the user is human, and does that assumption create drag for AI-native workflows?”

There’s going to be a generation of products that get rebuilt — or simply replaced — not because they were bad, but because they were designed for a user who no longer exists as the primary actor in the workflow.

The companies that think about this now, before it’s obvious, will be the ones designing intentionally for both. Not abandoning the human experience, but not assuming it’s the only experience that matters.

Are you designing your product for humans, or is it time to start thinking about AI as the user?