Skip to content

Search across everything

Find what you've already written. Across every format.

Instant full-text search across articles, FAQs, uploaded documents, and slideshow training content. The article exists, the answer's in there, search just has to do its job.

The article exists. The slide says it. The PDF spells it out.

The information your team needs is somewhere in your knowledge base. The problem is, it's spread across formats. Some of it lives in articles. Some in PDFs your ops team uploaded. Some on slide 14 of last quarter's training deck. Most knowledge bases search articles and stop there. The PDFs and slides become an island.

So your team searches, gets nothing useful, and either gives up or asks a colleague who happens to remember which deck it was in.

Search needs to reach into every format you've already published.

Full-text search. Every format. Every word.

KnowledgeScout indexes the full text of everything you publish. Article bodies, FAQ entries, collapsible sections, uploaded documents, and slideshow content. A search runs across all of it at once and returns a ranked list of matching results.

Results are permission-scoped to the asker and tenant-isolated by default. A Reader sees only what their role and team-scoped articles permit. Customers on the public help centre see only public articles. There is no shared index across tenants and no risk of leakage.

For questions where the asker doesn't know the right terminology, search hands off to the AI Chatbot, which uses an agentic loop to keep trying with rewritten queries. Two paths, one source of truth underneath.

Three surfaces, all searched together

Most knowledge bases search articles and skip the rest. We index everything you publish.

Articles

Article titles, metadata, body text, FAQ entries, and collapsible sections. The whole article gets indexed, not just the title.

Documents

Uploaded PDFs, Word docs, and spreadsheets are indexed at upload. The body text is searchable, not just the file name.

Slideshow training content

Text on every slide of every presentation is indexed and searchable. Training material stops being an island and joins the rest of your knowledge base.

Two ways to find an answer

Search is fast and deterministic. AI Chat is smarter when the asker doesn't know the right words.

Search

Type a phrase. Get a ranked list.

Full-text search across articles, documents, and slides. Instant results, no AI credits, runs on every plan. Best for known queries: a policy by name, a procedure you wrote last month, a slide you remember from training.

AI Chat (with agentic loop on B+E)

Ask in plain English. Watch the AI work.

For questions where the asker doesn't know the right terminology, AI Chat takes over. The agentic loop keeps trying — rewriting the query, retrying the search — with each attempt visible to the user. Best for fuzzy questions: "what do we do if a customer asks for a refund three months later?"

More on AI Chat →

What makes the search useful

Full-text, not just titles

A search for "refund window" hits articles, FAQs, and PDF policies that contain the phrase, not just ones with it in the title.

Document content indexed

PDFs, Word docs, spreadsheets become searchable as soon as they're uploaded. No manual tagging or transcription needed.

Slide-level training search

Slideshow training content is indexed slide by slide. The exact slide where something was mentioned is findable in seconds.

Permission-scoped

Results respect the asker's role and team-scoped article access. The right person sees the right content, no leaks.

Tenant-isolated

Each customer gets a private database with its own index. No shared search across tenants. Content is never used to train AI models.

Search analytics

See what your team is searching, what returned no results, and which queries handed off to AI Chat. The questions that don't have answers become your content roadmap.

Who this is for

Teams with content scattered across formats

If your knowledge sits in articles, PDFs, and training decks, you need a search that reaches into all three. Most don't.

Contact centres mid-call

Search runs instantly with no AI credit cost. Agents can run lookups during a live call without worrying about latency or burning credits.

Customers on the help centre

Public help centre search uses the same engine, scoped to public content. Customers find their own answers in seconds, fewer tickets reach your team.

AI agents calling via MCP

Agents use the same search through the MCP server. Same scope, same permissions, same content surfaces.

Why KnowledgeScout's search

1.

Searches every format you publish

Articles, FAQs, sections, PDFs, Word docs, spreadsheets, slideshows. Most KBs index articles and call it a day. Yours doesn't have to.

2.

Works with or without AI

Search alone is a complete way to use the knowledge base. No AI required, no AI credits consumed, no AI features to disable. AI Chat is on every plan; the agentic loop is on Business and Enterprise as an opt-in toggle for harder questions. You don't have to use either.

3.

Permission-scoped and tenant-isolated

No shared search index. No content leaving your tenant. No risk of training data exposure. Each customer's search is its own world.

4.

Analytics that surface the gaps

Searches that returned nothing, queries that escalated to AI Chat, popular content. Your search analytics become a content roadmap, not just a usage chart.

Common questions

What does search cover?

Full-text search runs across every content type: article titles, metadata, body text, FAQs, collapsible sections, uploaded documents (PDFs, Word docs, spreadsheets), and slideshow training content. If it's published in your knowledge base, it's findable.

Does search work inside uploaded documents?

Yes. PDFs, Word documents, and spreadsheets are indexed at upload. A search hit will return the document name, and clicking through opens the file with the matching term highlighted (where the format supports it).

Does it search inside slideshow training content?

Yes. Text on every slide of a presentation is indexed and searchable. Most knowledge bases skip this and leave training content as an island. We don't, because training material is often where the most up-to-date wording of a process lives.

What if my team doesn't know the right keywords?

That's where the AI Chatbot picks up. The chatbot uses an agentic loop — when an initial query returns weak results, it rewrites the query (handling synonyms, alternate phrasing, related concepts) and tries again, with the rewritten searches visible to the user. Search is the fast deterministic path. AI Chat is the path for harder questions.

Is search permission-scoped?

Yes. Search results respect the asker's role and team-based article access. A Reader on the Sales team will only see articles scoped to Sales (or to everyone) in their results. The same query run by an Admin will return more. Customers using a public help centre only see articles marked public.

Is search isolated per tenant?

Yes. Each customer gets a private database with its own search index. There is no shared index across tenants, and your content is never used to train AI models. Search runs entirely within your tenant.

The article exists. Make it findable.

Full-text search across articles, documents, and slides on every plan. AI Chat on every plan; agentic loop on Business and Enterprise for the harder questions.