Ask your documents questions in plain language
AI document search lets people type a question and get the passage that answers it, with the document it came from, instead of guessing the exact keywords.
What AI document search is
Ordinary search matches words. AI search matches meaning, so "how do I return an item" can find the page titled "Refunds and exchanges". The answer always shows the passage and document it came from, so a person can check it.
We build it on the material you choose, keep it up to date as your content changes, and agree what it may and may not search before we start.
What you can search
- Policies and manuals
- Staff find the right rule or procedure by asking a question.
- Your website and help centre
- Visitors describe what they need and land on the right page.
- Reports and records
- Find the right record by asking, with the source shown.
- Product and catalogue data
- Match a question to the right item or specification.
How a document search project runs
We begin with one set of documents and prove it works before adding more.
1
Choose the documents
We agree which files and pages the search covers, who may see what, and what a good answer looks like.
2
Index and prototype
We prepare your content so the search can find passages by meaning, then show you results for your real questions.
3
Check the answers
We test with questions your team really asks, tune how results are ranked, and make sure every answer links to its source.
4
Launch and keep it fresh
We connect it to your sources so new and changed documents appear in search, and we review what people look for.
How we keep search safe and accurate
- Source shown every time
- Every answer shows the passage and document it came from, so people can check it.
- Access rules respected
- People only see results from documents they are allowed to open, matching your existing permissions.
- Careful with data
- We agree where documents are stored and processed before we build, and API keys stay on the server.
- Limits and logs
- Usage limits and logs are part of the build, and the logs show what people search for and cannot find.
Questions about AI document search
Can it search our private files?
Yes, if you choose to include them. We agree what is included and what stays out, and who may see what, before we build.
Does it replace our current search?
It can sit beside it or replace it. We start with a small set of documents and compare the results with your current search.
How do I know an answer is right?
Each answer shows the passage and the document it came from, so a person can check it.
Where does our data go?
We agree the data flow with you first. Keys stay on the server, and we send models only what the task needs. Tell us about any data you cannot share.
Want your team to find things faster?
Tell us what people search for and cannot find, and we will tell you what is realistic.