Use case · Updated
Keep documentation up to date automatically
Every docs chatbot hits the same wall: it is only as good as the pages behind it. When the docs are missing or out of date, the bot either guesses or gives up. usedocs treats each of those moments as a fix to draft, not a dead end.
TL;DR
Short answer
To keep documentation up to date, the work has to start when the product changes, not when a customer complains. usedocs reads merged pull requests and proposes edits to the articles they affect, turns questions the docs could not answer into drafted articles, and answers from the docs with a link to the source. You approve everything in one review queue.
Who this is for
SaaS teams with product docs or a help center who want customers to get answers without opening a ticket. Typical owners are founders, support leads, or a product manager who maintains docs part time. Best fit when the product changes often and the docs struggle to keep up.
Problems this solves
- The chatbot repeats outdated steps because the docs never caught up with the last release.
- Answers sound confident without showing where they came from.
- When the docs have no answer, the bot invents one or loops on a polite apology.
- Nobody knows which docs to write next, so the same questions keep failing.
- Teams pay for a chatbot and a docs platform that never talk to each other.
How usedocs works for this job
Bring your docs Import from your docs URL, Zendesk Guide, Intercom, GitBook, Notion, or an OpenAPI file, or write in the usedocs editor. Articles arrive as drafts with redirects.
Answer with sources The assistant answers on your help center and in a widget, linking the article behind every answer, and says when the docs do not cover a question.
Turn misses into drafts Unanswered questions, "not helpful" votes, and imported tickets become ranked gaps, each with a drafted article.
Follow the code Connect GitHub and merged pull requests propose edits to the articles they change, so answers match the product.
What you get
Source on every answer
Customers can open the article behind an answer, and your team can fix the right page when something is off.
Honest refusals
Weak matches get a clear "the docs do not cover this" and become gaps instead of guesses.
Ranked gap list
Missed questions grouped by theme and ranked by how often customers ask.
Edits from merged PRs
Proposed article edits with the code change as evidence.
AI ready docs
A Markdown copy of every page, llms.txt, and a Copy page menu for ChatGPT, Claude, and Cursor.
Test before you embed
Ask the assistant the ten questions your inbox sees most. Open every source link and check the article says what the answer says. Then ask something the docs do not cover and confirm it refuses. If an answer is wrong because the article is wrong, fix the article, not the bot; that fix helps every reader, not only chat.
Run the weekly loop
Once a week, open the review queue. Approve the drafted articles for the top gaps, then clear the edits proposed from merged pull requests. Ten to thirty minutes a week keeps a small team's docs complete and current, and the chatbot gets better without any prompt tuning.
Measure the right thing
Chat volume says little. Track the share of answers with a source, the number of gaps closed, and whether repeat tickets on those themes fall the next month. Those three numbers show whether the docs, and so the chatbot, are getting better.
Write for the question, not the feature
Drafted articles start from the customer's own wording, so keep it. A title like "Why is my invoice in draft?" matches what people search and ask the assistant, while "Invoice states" does not. Keep each article to one job, link related articles instead of repeating them, and put the answer in the first two sentences. The assistant quotes those first lines most often, and readers skim them first.
Why teams pick usedocs
- Answers link their sources
- Misses become drafted articles
- Merged PRs become proposed edits
- One plan at $99 a month with a 7 day trial
- Markdown copies and llms.txt for AI tools
Not for you if
Skip this if you want a general chatbot for marketing pages, a voice bot, or an internal knowledge assistant over thousands of private files. usedocs is built for customer facing product docs.
FAQ
How do you keep documentation up to date?
An assistant that answers questions from your product docs, ideally with a link to the source and an honest answer when the docs are silent.
How does usedocs avoid made up answers?
It answers only from your articles, links the source, and refuses when the match is weak. Each refusal becomes a gap to fix.
Do I have to move my docs to usedocs?
No. You can keep your docs site and use the widget or the Answer API, or let usedocs host the help center with a custom domain.
How does the chatbot stay current?
Merged pull requests propose edits to the affected articles, and live articles are checked against your code on a schedule.
Can parts of the docs stay private?
Yes. Make the help center private, or limit collections to signed-in readers with magic links, a shared password, or a token from your app.
Can the docs be translated?
Yes. Growth includes four languages beyond your primary one, translated automatically on every publish, sent to review first, or by hand.
Is there a free trial?
Yes. A 7 day Growth trial with no card, then one plan at $99 a month.