Guide · 8 min read

Building an internal
AI knowledge base

Your organisation already knows the answer to most questions your team asks. It is just spread across a drive, a wiki, six mailboxes and one person who is on holiday. Here is how to make it answerable.

00Why this matters

The knowledge exists. Nobody can find it in time.

Ask a new starter what slows them down and it is rarely the difficulty of the work. It is not knowing where the answer lives, and not wanting to interrupt the person who knows. Meanwhile the same handful of colleagues answer the same questions every week.

Traditional search does not solve this, because the answer is usually not in one document. It is a policy in one place, an exception noted in an email, and a spreadsheet nobody linked. You need something that reads across all of it and assembles an answer.

That is achievable now, and it is also the category where badly-built systems do the most damage, because a confident wrong answer about your own policy is worse than no answer.

GUIDEThe playbook

How to build one your team actually trusts.

01Scope

Start with questions, not with documents

The instinct is to index everything. Resist it. Begin from the questions your team actually asks, and only bring in the content needed to answer them.

  • Collect the real questions first. Two weeks of the questions arriving in your team chat is better source material than any content audit. It also gives you a test set to evaluate against.
  • Pick one domain to begin with. Internal policy, or product knowledge, or onboarding. One domain with current content beats everything at once with mixed quality.
  • Name a content owner. Someone has to be responsible for whether the source material is right. Without that, the project becomes a very sophisticated way to repeat outdated documents.
  • Decide what is out of scope. Explicitly. Personnel files, legal advice, anything under specific confidentiality. Saying so up front avoids an uncomfortable discovery later.
02Retrieval

Why retrieval quality is the whole game

The model gets the attention, but the answer quality is decided almost entirely by whether the right passages were found in the first place.

  • Chunk documents thoughtfully. How content is split determines what can be retrieved. Splitting mid-clause or mid-table is a common cause of answers that are almost right.
  • Combine meaning-based and keyword search. Semantic search handles paraphrasing; keyword search catches product codes, names and identifiers where exact matching matters. Most good systems do both.
  • Keep dates and versions attached. Retrieval should prefer the current policy over the 2023 draft. That requires knowing which is which, which requires metadata nobody enjoys adding.
  • Evaluate retrieval separately. Before judging answers, check whether the right passages were retrieved at all. Most disappointing answers are retrieval failures wearing a model's clothing.
03Trust

Citations, refusals and confidence

Trust is not built by the system being right most of the time. It is built by the system being checkable, and by it admitting when it does not know.

Cite every claim to a source

Each part of the answer should link to the document and section it came from. This is non-negotiable. Without it your team is asked to take an AI's word for internal policy.

Make 'I don't know' a good outcome

When nothing relevant is retrieved, the system should say so and point at who to ask. Systems that always produce an answer teach people to distrust all of them.

Answer only from retrieved content

The model should not fall back on general knowledge for a question about your business. Constrain it, and test that the constraint holds.

Give feedback a single click

A thumbs-down with the question attached tells you exactly which content is missing or wrong. This becomes your improvement backlog.

04Permissions

Access control that mirrors reality

This is the part quickly-built internal tools get wrong, and it is the one that turns a helpful project into an incident.

  • Enforce permissions at retrieval. Filter what can be retrieved for that user, before the model sees anything. Filtering the answer afterwards means sensitive content was already used to produce it.
  • Reuse your existing groups. Do not invent a second permission model. It will drift out of step with the real one, and nobody will notice until it matters.
  • Log who asked what. Not to police your team, but so you can answer the question 'could this person have seen that' with evidence rather than reasoning.
  • Handle removals promptly. When a document is deleted or a person leaves, the index has to follow. A stale index is a quiet access-control problem.
05Adoption

Getting people to actually use it

A knowledge base nobody opens is the most common outcome, and it usually has nothing to do with answer quality.

  • Put it where people already work. Inside the chat tool or the intranet they open anyway. A separate site requires a habit, and habits are harder to build than software.
  • Launch with the top questions already working. Test against the real questions you collected. First impressions are decisive, and one bad early answer costs a dozen good ones.
  • Have the experts endorse it. If the people who currently answer the questions publicly point to it, adoption follows. If they quietly ignore it, so will everyone else.
  • Treat content as ongoing work. Review the questions it could not answer every month and fix the source material. That review is the difference between a launch and a system.
CTATalk to Brains

This is what Second Brain does.

We built a product for exactly this, and we adapt it to how each organisation is set up. Tell us what your team keeps asking and where the answers live, and we will tell you whether this is worth building for you. See our products or get in touch.

Guide FAQ

Common questions about internal AI knowledge bases.

What is an internal AI knowledge base?

A system your team can ask questions of in plain language, which answers from your own documents rather than the open internet, and shows which document each part of the answer came from. Technically it usually means retrieval-augmented generation: find the relevant passages first, then have the model answer using only those.

How is this different from search?

Search returns documents and leaves the reading to you. A knowledge base returns an answer assembled from several documents, with citations so you can verify it. The difference matters most when the answer lives in three places, which is the normal case.

Will it make things up?

It can, and preventing that is the main design work. A well-built system answers only from retrieved passages, cites them, and says it does not know when nothing relevant was found. If you cannot click through to the source of a claim, do not deploy it.

How do permissions work?

They have to mirror the permissions you already have, enforced when documents are retrieved rather than filtered afterwards. If a person cannot open a file today, no answer they receive should be built from it. This is the most common serious flaw in quickly-built internal tools.

How much content do we need?

Enough that finding things is genuinely a problem, which is usually hundreds of documents rather than tens. Below that, better organisation beats an AI layer. The more important factor is that the content is current, because a knowledge base repeats outdated policy with total confidence.

What is the most common reason these projects fail?

Stale or contradictory content. The system is only as good as what it retrieves, so if three versions of a policy exist, it will quote one of them convincingly. Content ownership is not a side task; it is the project.