Skip to content
AI.Engineering
WorkOperations

Siloed wiki → 3 hours saved per employee per week

Engineering, ops, and customer success each had their own wikis. Nobody could find anything; people pinged each other instead of searching.

Industry

Operations

Outcome

3h saved/week

Detail

per employee

Stack

4 tools

The problem

Engineering, ops, and customer success each had their own wikis. Nobody could find anything; people pinged each other instead of searching.

The solution

A semantic search engine across all three corpora with permission-aware retrieval, plus a Slack bot that answers natural-language questions with citations.

Approach

Permission-aware retrieval was non-negotiable: HR docs can't leak into Engineering channels. Every chunk is tagged with its ACL and filtered at query time.

Detail

The Slack bot rewrites vague questions into precise queries before retrieval, which lifted answer quality more than any model upgrade did.

Sample evaluation harness

// evals/extraction.eval.ts
import { runEval } from "./harness";
import { invoiceCases } from "./cases";

await runEval({
  name: "invoice-extraction",
  cases: invoiceCases,
  judge: "strict-json-schema",
  passThreshold: 0.95,
});

Let's build

Have an AI feature that needs to ship without falling over?

Tell me what you're trying to automate. I'll come back with whether it's a 2-week build, a 2-month build, or honestly not a good fit.