SolutionsPrivate knowledge & retrieval

Give your team accurate answers, grounded in your own data.

We build permission-aware retrieval systems for internal knowledge, support, and research — with citations back to source and access control built in from day one.

Most internal search is poor for a simple reason: the answers live in many places, in many formats, behind different permissions. Enterprise RAG puts one retrieval layer over those sources and has a model answer only from what it finds, with citations to the source. Whether it works depends less on the model you pick than on retrieval quality, permissions, freshness and how you measure it.

That is where we spend our effort. A system that returns the wrong passage, shows a document to someone who should not see it, or cites a source that does not say what the answer claims will lose trust quickly. The sections below set out how we build against each of those failures. For a deeper technical walk-through, read our enterprise RAG architecture guide.

What this covers

Where we focus

Tailored to your systems and data, but these are the core building blocks.

Permission-aware retrieval

Answers respect the same access controls as your source systems — no accidental leaks.

Internal knowledge & enterprise search

One retrieval layer across docs, wikis, and databases instead of five separate search boxes.

Document Q&A with citations

Every answer traces back to its source, so it can actually be verified.

Access control & evaluation

Built-in evaluation loops to catch hallucinations before they reach a user.

The problem

Why enterprise RAG disappoints, and what we do about each cause

Most failed RAG projects fail in retrieval or permissions, not in the model. These are the usual causes.

What users seeUsual causeWhat we do
The answer ignores the right documentChunks lost their context, or search missed an exact term such as an ID or error codeAdd context to chunks, combine semantic and keyword search, and rerank the results
Someone sees content they should notPermissions were applied after retrieval, or not synced from the sourceEnforce access at query time using permissions carried with every chunk
The answer is out of dateThe index is not refreshed when sources changeContinuous sync with deletes and updates, and a visible last-updated date
A confident answer with a wrong or missing citationThe model answered beyond the evidenceAnswer only from retrieved passages, check citations, and say so when evidence is missing
Answers are slow or expensiveToo much text sent to the model on every queryRetrieve fewer, better passages, and cache what repeats

Security teams should note that OWASP's 2025 list of risks for LLM applications includes sensitive information disclosure and, new that year, vector and embedding weaknesses. Weak access control at the retrieval layer is a common route to both.

Retrieval pipeline

How we build retrieval that finds the right passage

Each stage has one job, and each is tested on your own documents rather than assumed to work.

StageWhat we doWhy it matters
Ingest and parseRead documents, tables, PDFs and pages, keeping structure and source detailsPoor parsing, such as broken tables, cannot be fixed later
Chunk with contextSplit text sensibly and prepend a short explanation of where each chunk comes fromA chunk that says the company grew 3% is useless without knowing which company and quarter
Hybrid searchCombine semantic search with keyword search, and merge the resultsSemantic search can miss exact identifiers that keyword search finds
RerankScore a wider set of candidates and pass only the best to the modelBetter answers with less text, so lower cost and latency
Generate with citationsAnswer only from the passages, and link each claim to its sourceLets people verify, and builds trust
Decline when unsureSay the sources do not contain the answer instead of guessingA refusal costs less than a confident wrong answer

Anthropic's published tests of this approach found that adding context to chunks cut failed top-20 retrievals by 49 percent, and by 67 percent with reranking added, in their test set. Results vary by data, which is why we measure on yours. They also point out that a knowledge base under roughly 200,000 tokens, about 500 pages, can simply go in the prompt with caching and may not need RAG at all, a point we will raise with you if it applies.

Permission-aware retrieval

Making sure answers respect who is asking

An internal assistant must not become a way around your access controls. The right approach depends on how your sources manage permissions.

ApproachHow it worksTrade-off
Permission metadata filtered at query timeEach chunk carries the users or groups allowed to see it, and the search only considers allowed chunksNeeds permissions synced from the source and kept current
Separate indexes per team or tenantEach group has its own indexStrong isolation, but more to run and duplicate content
Check access against the source at answer timeConfirm the user can open each cited document before showing itMost accurate, at the cost of extra calls and latency
Filtering results after generationRemove disallowed content from the answer afterwardsWe avoid this on its own, because restricted text has already reached the model

We usually combine a query-time filter with a source check on the final citations, and test it with accounts that should and should not see specific documents.

Evaluation

How we know the answers are good before users do

We build a test set of real questions with known good sources and answers, and score every change against it.

MeasureWhat it tells youHow we check it
Retrieval failure rateHow often the right passage is missing from the top resultsCompare retrieved passages with the known source for each test question
GroundednessWhether every claim in the answer is supported by the cited textReview samples, with automated checks to flag unsupported claims
Answer correctnessWhether the final answer is rightScore against reference answers written by your experts
Citation accuracyWhether each citation points to the right placeCheck that the cited passage supports the claim
Correct refusalsWhether it declines when the answer is not in the sourcesInclude unanswerable questions in the test set
Latency and costResponse time and spend per queryTrack p50 and p95 latency and tokens per answer

The same test set runs after every change to chunking, models or prompts, so quality does not drift unseen. Our guide to evaluating AI agents for production covers the approach in more detail.

Technology choices

Choosing where the vectors live

We pick the storage layer to fit your scale and your existing stack, not the other way round.

OptionFits whenWatch for
PostgreSQL with pgvectorYou already run Postgres, the corpus is moderate, and you want permissions and metadata in plain SQLVery large corpora and heavy query loads need careful tuning
A managed vector database such as PineconeYou want managed scale and low operational loadAnother system to secure, and permissions must be synced into it
WeaviateYou want an open-source vector database with built-in hybrid searchYou take on hosting unless you use a managed offering
A search engine with vector supportYou already run a search platform and want one system for keyword and semantic searchTuning relevance across both takes effort

For the trade-offs in depth, read Production RAG with PostgreSQL and pgvector and RAG knowledge base tools: build vs buy. For the delivery side, see our RAG knowledge systems service.

Proof, not just a pitch

Real numbers from named engagements.

Every figure links to the case study or source behind it.

FAQ

Common questions about this work.

Straight answers, grounded in what we've actually shipped.

What is enterprise RAG?

Retrieval-augmented generation connects a language model to your own documents and data. The system finds the relevant passages for each question and has the model answer from them, with citations, instead of relying on what it learned in training.

Do we need RAG, or can we put our documents in the prompt?

If your knowledge base is small, roughly under 200,000 tokens or about 500 pages, putting it in the prompt with caching can work well and is simpler. RAG earns its place when the content is larger, changes often or needs permissions applied per user.

Should we use RAG or fine-tune a model?

For knowledge that changes and needs citations, RAG is usually the better fit. Fine-tuning suits changing a model's style or behaviour more than teaching it facts. We explain the trade-offs in our guide to RAG versus fine-tuning.

How do you stop the system showing people documents they should not see?

We carry each document's permissions with its chunks and enforce them at query time, then check the final citations against the source. We test with accounts that should and should not have access.

How do you reduce hallucinations?

By answering only from retrieved passages, requiring a citation for each claim, checking those citations, and declining to answer when the sources do not contain the information. We measure groundedness on a test set.

How do you keep the knowledge up to date?

We sync from your source systems on a schedule or on change events, handle updates and deletions, and show when content was last refreshed.

Can it run in our own cloud environment?

In most cases yes. We can deploy inside your AWS or other cloud account so data stays within your boundary. Data handling and compliance requirements are confirmed during discovery.

Ready when you are

Have a workflow worth automating?

Tell me what's slow, manual, or fragile today, and I'll help you figure out the smartest first build.