Skip to main content
Documents and knowledge

Documents and knowledge

Help new staff find answers in your documents on day one

A searchable knowledge base turns your files into a resource that answers questions and cites where the answer came from.

· 5 min read · Document AI · Onboarding · Knowledge management · Process automation

Every time you hire someone, you spend hours answering the same questions. Where do we keep client contracts? What's our policy on scope creep? How did we handle this situation last year? These questions are not stupid. They're necessary. But they're also a drain on your senior people, and they slow down the new hire.

A searchable knowledge base solves this by letting new staff ask questions of your documents the way they'd ask a colleague, and get answers that point to the exact page where the answer lives.

Here's what this looks like in practice.

The problem onboarding creates

A property manager in Surrey with a 12-person team spends roughly 4 hours per new hire answering questions about strata bylaws, precedent decisions, and how the firm handles common disputes. A dental clinic in Kelowna has a similar pattern: new hygienists and receptionists ask where to find consent forms, how to handle insurance queries, and what the cancellation policy is. A 15-person accounting firm in Burnaby has the same issue with new junior staff asking about client file structure, tax deadlines, and past interpretations of ambiguous rules.

In each case, the answers exist. They're in old emails, past contracts, policy documents, and the heads of the people who've been there longest. But they're not organized in a way a new person can find them.

The cost is real. A new hire who spends their first week asking questions is not yet productive. A senior person who spends 3 to 4 hours onboarding is not doing billable work. And the new hire often gets inconsistent answers, depending on who they ask.

What a searchable knowledge base does

Instead of asking a person, a new employee types a question into a search tool. The tool reads through your documents—contracts, policies, past emails, decision logs, whatever you feed it—and returns answers that cite the source.

A new property manager asks: "What's our process for handling noise complaints?" The tool finds the relevant clause in the bylaws, the procedure document from three years ago, and an email thread about a specific case, and shows all three with page numbers.

A new dental receptionist asks: "Do we bill insurance directly or ask the patient to claim?" The tool pulls the relevant policy and the last three invoices that show how it was done.

A junior accountant asks: "How did we handle this client's home office deduction last year?" The tool finds the working file and the tax memo.

The answers are not perfect. The tool sometimes returns things that are close but not quite right, or misses a relevant document. That's why a person still needs to review the answer before acting on it. But the tool gets the new hire 80% of the way there in seconds, not hours.

What it takes to set up

You don't need to rewrite anything. You use the documents you already have.

The real work is deciding what goes in. A property manager might include bylaws, past council meeting minutes, policy documents, and a log of major decisions. An accounting firm might include tax guides, past client memos, and procedure documents. A plumbing company in Richmond might include warranty policies, past customer agreements, and equipment manuals.

You don't need to include everything. Email inboxes, for example, are often too noisy and too personal to be useful. Past invoices are usually not helpful for answering questions about process. You pick the documents that answer the questions new staff actually ask.

Once you've chosen what goes in, the setup takes 2 to 4 weeks. You upload the documents, test the tool with real questions, and fix the ones it gets wrong. The cost is modest—typically 200 to 400 dollars a month to run the tool—but the real cost is your time to organize and test.

Who benefits most

This works best for firms that run on documents and have consistent questions from new staff.

A strata council with 5 to 50 units and a history of decisions and bylaws sees immediate value. A professional firm—accounting, law, consulting—with junior staff who need to learn precedent and policy benefits strongly. A property management company with high turnover in field or office roles saves hours per hire.

It's less useful if you have very few documents, or if your documents are mostly confidential and can't be searched. It's also not worth doing if you only hire once a year, or if your onboarding is already fast.

Accuracy and trust

The tool is not perfect. It sometimes hallucinates—it makes up an answer that sounds plausible but isn't in your documents. It sometimes misses relevant information. It sometimes returns outdated documents if you haven't cleaned them up.

That's why the answer must always include a source, and why a person must always review before acting. The tool is a research assistant, not a decision-maker.

To keep it accurate, you need to review the answers it gives for the first month or two, and flag the ones that are wrong. The tool learns from feedback. You also need to retire old documents and update policies when they change, so the tool doesn't return obsolete information.

Where to start

  • Spend an hour listing the questions new staff ask you in their first week. Don't overthink it; just jot them down. These are the questions your knowledge base should answer.
  • Go through your file system and gather the documents that answer those questions. Bylaws, policies, past decisions, procedure guides, and client agreements are good candidates. Set aside anything too confidential or too personal.
  • Talk to your team about what they'd want to search for. You'll learn what documents matter most and what's missing. This conversation also preps them for the change.

Drowning in documents?

We turn your documents and records into something you can ask questions of, with every answer cited to the page. It’s the same technology behind SearchStrata.