Stacklink
Stacklink

Use Case

See what your team knows — and what they don't

Track search patterns, identify documentation gaps, and measure knowledge health across your entire organization.

Read-only accessEU hostedEncrypted in transit & at rest

Search analytics

See the top 100 queries, which ones return results, and which ones hit dead ends. Fix gaps before they cost you.

Usage patterns

Which teams search most? Which docs are most referenced? Which sources are never touched? Data-driven knowledge management.

ROI measurement

Calculate time saved, support tickets deflected, and onboarding days reduced — real metrics for leadership.

You can't improve what you can't measure. Knowledge analytics turns tribal knowledge into organizational intelligence.

Capabilities

How it works for knowledge analytics

01

Gap analysis dashboard

Queries with no results or low-confidence answers reveal exactly where your documentation falls short.

02

Team-level insights

Compare search behavior across engineering, product, support, and sales. Each team has different knowledge needs.

03

Trending topics

Spot emerging questions before they become bottlenecks. A spike in "deployment process" queries means your docs need updating.

04

Source health scores

Each connected tool gets a freshness score. Stale Notion pages drag down your overall knowledge health.

05

Weekly reports

Automated email to knowledge managers with key metrics: searches, results quality, gaps identified, and docs updated.

06

Export & API

Pull analytics data into your existing BI tools. CSV export and REST API for custom dashboards.

Under the hood

The shape of a safe query.

Three guarantees, made visible. Reads run on the role you already trust. Writes pause for a person. The query is the receipt.

01

The write-gate, demonstrated

Two paths for two intentions — reads ship, writes wait.

A SELECT executes the moment it parses. Anything that mutates state pauses for a person on the other side of the gate. Same query surface, two doors.

Read pathexecutes immediately
SELECTquery token
parse
plan
execute
rows

Result rows return inline. Audit-logged, replayable, but the agent does not stop to ask.

Write pathhuman approves
UPDATEquery token
parse
plan
gate
approve
execute

INSERT, UPDATE, DELETE — all stop at the gate. A named human reviews the exact statement, then approves or denies.

02

The query is the citation

Question in, annotated SQL out, rows attached, rerunnable.

A natural-language question becomes SQL the analyst can read, with the schema-aware decisions called out. Result rows return with the exact query attached — replay it and you get the same answer.

What is MRR by plan for the last 6 months?you
1SELECT date_trunc('month', period_start) AS month,
2plan,
3SUM(mrr_cents)/100.0 AS mrr_eur
4FROM subscription_periods
5WHERE period_start >= now() - interval '6 months'
6GROUP BY 1, 2
7ORDER BY 1, 2;
monthplanmrr_eur
2025-07Pro€198,420
2025-08Pro€204,910
2025-09Pro€212,340
3 rows · 320 ms · audit-logged
03

Where the 1.4 seconds goes

Schema, plan, execute, cite — accounted for, segment by segment.

Most BI questions sit in queues for days. Here, the whole round-trip is under a second and a half — and every millisecond is labelled.

Stages of a single queryrelative cost
13%
50%
25%
13%
  • Inspect schema

    ~13% of the round-trip

  • Plan the SQL

    ~50% of the round-trip

  • Run on the warehouse

    ~25% of the round-trip

  • Cite + render

    ~13% of the round-trip

Median over a six-month rolling window. Warehouse exec varies with the query — schema and citation costs do not.

Ready to unify your knowledge?

Join the waitlist to be first in line when we launch.

No credit card requiredSetup in 5 minutes