# How do you design an AI agent approval screen people actually read?

Show 4 things on 1 screen: the action, the exact amount, the reason, and the rule that stopped it. Then 1 tap to approve or decline. Send it to the 1 person who owns the decision, on the device they carry. An approval that needs digging gets clicked through unread.

By Surinder Thakur, AI Product Designer and Builder. Published 17 Nov 2026. Topic: [Trusting AI with money](https://www.surinder.design/work?type=trusting-ai-with-money).

## Why do people click approve without reading?

Because the request arrives too often, or hides the detail behind another click. People learn that approving is safe and stop looking.

Fewer requests fix the first problem. A screen with everything on it fixes the second.

## What does the approver need to see?

The action, the amount, the reason and the rule that stopped it. A new supplier, or an order over the limit.

In the Ordering agent design, the workshop owner gets the exact amount on their own phone. 1 tap approves or declines.

## What happens while they decide?

Nothing is lost. The agent holds the work, like reserved stock, and moves on to other cases.

If nobody answers in time, it follows a rule you set: remind, pass to a backup, or cancel.

## How do you know the approvals work?

Watch 2 numbers: how long a decision takes, and how often people decline. A decline rate near zero can mean nobody reads.

Some declines are healthy. They show the rule is catching what it should.

## In 1 picture

An approval people read shows the action, amount, reason and the rule that stopped it.

Approval request. The agent fills action, amount, reason and rule. Action, amount and reason check out. Rule is flagged.

A payee it has never paid: the owner decides in 1 tap.

Everything needed to decide, on 1 screen.

## Next questions

- [What should an AI agent be allowed to do without asking?](https://www.surinder.design/work/what-should-an-ai-agent-do-without-asking): Anything a person can check and undo in time. Reading, sorting and drafting need no approval, because nothing leaves the building.
- [How do you stop an AI agent making a mistake with money?](https://www.surinder.design/work/ai-agent-mistake-with-money): You limit what it can do alone. The agent pays only inside the rules you set: approved payees, a limit, a monthly budget.
- [How do you show people what an AI agent is doing while it works?](https://www.surinder.design/work/how-to-show-what-an-ai-agent-is-doing): Show 4 things. The plan before it starts, so people can stop it early.
- [What should an AI agent for a finance team actually do?](https://www.surinder.design/work/what-should-an-ai-agent-for-finance-do): The matching, chasing and checking that fills month end. It matches invoices to orders, flags duplicates, chases late documents and drafts the close list.

## Questions that lead here

- [What is agentic UX, and why does it matter?](https://www.surinder.design/work/what-is-agentic-ux): Agentic UX is design for software that acts on a person's behalf.

Page: https://www.surinder.design/work/how-to-design-an-ai-agent-approval-screen
