跳到正文
DEV Community· Samuel Kimama·· 2 小時前AI 評分55

AI 對話介面把待辦清單當成普通文字，令用戶難以追蹤進度

Your AI hands users a plan, then lets them lose it

AI 導讀

作者指出，AI 對話介面把模型列出的待辦清單當作普通文字，令用戶難以回看項目、辨認後續訊息所指內容及追蹤完成狀態。文章提出由前端固定清單、由用戶手動勾選，到以穩定 ID 自動同步項目狀態的分階段做法。自動勾選應提供明顯的撤銷操作；手機上，面板則可收起為顯示未完成項目數量的提示。

正文

Cover image for Your AI hands users a plan, then lets them lose it

Samuel Kimama

You ask an LLM to review your setup and get back five numbered issues. Good answer. You start on #1, and a few messages later the list is a screen and a half above you.

To get to #3, you scroll up, reread it, scroll back down and type "for #3…", hoping the model knows which list you mean. Later replies bring numbered lists of their own, so "#3" can mean two different things. And nothing tracks which items you actually finished.

A chat where the reply listing five fixes is scrolling off the top, and the user types

A later reply with its own numbered list. The user says

Why it stays broken

To the interface, a list of things to work through is just formatting. It renders and scrolls exactly like five paragraphs of explanation. Some tools now show the model's own multi-step work as a checklist, but when the list is work for the user, it's still just text.

The principle: when a reply hands the user a set of things to work through, the interface should hold onto it as work, not as text.

A fix

The same chat with a floating

How you'd build it

Smallest version: a pin control on any rendered ordered list that moves it into a sticky side panel. Users check items off themselves. It's front-end only, with no model changes, and it already fixes the scrolling problem.

Solid version: the model marks a list as actionable and gives each item a stable ID. On later turns it reports which item it's addressing and whether it's resolved, so the panel can highlight and check items automatically. Clicking an item attaches its ID to your next message, so "this one" is never ambiguous.

Ambitious version: lists become part of the conversation's state. Later messages can add or split items, finishing the list closes its span, and scrolling back through history shows each list as it stood at that point.

Four frames: a list with a pin control, the list floating as a panel, items checked off, and the panel closing after the last check

Trade-offs to watch

  • A wrong checkmark is worse than none, so auto-checks need an obvious undo.
  • Not every numbered list is work. Ranked options or a recipe shouldn't follow you around.
  • On a phone, the panel collapses to a chip like "2 of 5 open."

The full post has a checklist for auditing your own product.

If you're building an AI feature that hands users steps to follow, how are you handling this?

來源：DEV Community · dev.to

#教學/實踐