# The best interface might be no interface

By Tommy Magee

Type: Writing
Canonical: https://tommymagee.com/library/less-interface

If AI is supposed to remove work, why are we creating another place for humans to work?

Expanded website edition of Tommy’s public LinkedIn post. Added examples are illustrative. The source links to the original.

SaaS founders, a request: build products my team and I don’t have to log into.

I actively solicit demos and get pitched new tools. Lately, I keep coming back to the same question. Do we really need another interface? A product can solve a narrow problem and still give the organization more work overall.

Another tool means another workflow to learn. Someone has to train the team, manage access, keep the information current, and support it when something changes. Eventually, someone has to replace it. Those costs rarely fit into the demo.

## Start with what needs to happen

When I see a promising solution, I increasingly ask whether it can do the work for us as a service, or make its capabilities available to our agents. I’m interested in the outcome. The screen is one possible way to get there.

Take a hypothetical project update. If the information already exists across the systems a team uses, asking everyone to enter it into one more dashboard adds a coordination job. A more useful product might assemble the current picture, show its sources, and flag the few things that need a person’s attention.

That still requires an interface somewhere. People need to check the answer, understand what changed, and correct mistakes. But there’s a difference between giving someone control and making them operate every step.

## There’s no maintenance-free option

Internally built agents have maintenance costs too. So does vibe coding. Someone needs to own the connections, permissions, quality, and behavior when the system fails. Building it yourself doesn’t make that responsibility disappear.

The comparison I care about is the total work involved. What does this remove? What does it add? Who ends up looking after it? A smaller interface is useful only if the complexity hasn’t just moved somewhere harder to see.

For teams buying software, I’d put those questions into the evaluation early. For founders, I think there’s an opportunity to make the product useful wherever the customer already works, including through an agent.

If AI is supposed to remove work, a new place for people to manage it is a pretty incomplete proposition. Show me the work that goes away. That’s where it gets interesting.

Original LinkedIn post: September 2026 (2026-09-06)

Source: https://www.linkedin.com/feed/update/urn:li:activity:7502513748235096064/

Topics: AI & work, Applied AI