Company knowledge, searchable — and still confidential.
Internal search fails twice: it can't find the document, or it shows one it shouldn't. AACSearch enforces document-level access in every query and answer, and connects people, projects and decisions across sources.
Queries plain keyword search gets wrong
политика отпусков 2026policyкто отвечает за проект Xpeople + projectдоговор с поставщиком Y условия оплатыrestricted documentпочему отказались от Zdecision historyWhere search loses customers today
Permission leaks
One search result or AI answer from a document the employee can't open is a security incident.
Scattered sources
Wiki, drive, tickets and chats each have their own search.
Questions that span documents
“Who decided X and why?” needs facts from several documents. Keyword search returns them one by one, if at all.
How AACSearch solves it, layer by layer
Each card names the capability behind it. Items marked “in development” are designed and scheduled, not live yet.
Finds the right item from the first letters — typos, synonyms, filters, meaning.
Document-level ACL
Users and groups on every document; filtering happens before hits, counts and facets are computed.
Widens the basket: similar, frequently together, personal.
Answers questions a filter can't, citing your own catalogue and docs.
Permission-aware answers
The assistant retrieves only documents the user may read, and cites them.
Learns how items, queries and topics connect — and leads the visitor to the goal.
Knowledge graph
Will extract people, projects, systems and decisions and link them across documents.
Global questions
Will answer broad questions (“what are our main supplier risks?”) by summarising communities of related documents — still within the user's permissions.
One visitor, from first query to goal
What the visitor does on the left, what AACSearch does on the right.
- 1
Search
An employee searches for a supplier contract.
Shows it only if their groups have access; totals don't reveal hidden documents.
Document-level access - 2
Ask
Asks about the payment terms.
Answers from the contract with a citation.
AI answers with citations - 3
Connect
Asks why the previous supplier was replaced.
Connects meeting notes, tickets and the contract into one cited answer.
GraphRAG· in development
Your first two weeks
A typical rollout for this industry. Most teams are live on day one and spend the rest of the time measuring and tuning.
Day 1
Connect the first source with user and group ids on each document.
Day 2
Test as two employees with different access.
Week 1
Add further sources and the assistant.
Week 2
Run leak tests with your security team.
What to measure
The numbers to judge a rollout by. Compare them with an A/B split or a before/after window on the same traffic.
- Time to find
- Minutes from question to document.
- Answer citation rate
- Share of answers with at least one cited source.
- ACL test coverage
- Cross-user leak tests must pass on every release.
Connects to
- REST API
- JSON / CSV export
- Sitemap crawl
- SDK (TypeScript, Python)
Capabilities used
Coming next
Questions
Search for Internal knowledge, live this week
Connect a feed or a site, embed one script, and measure the difference on your own traffic.