Wooly · research agent
Ask questions. Get answers. No dashboards required.
Wooly is Weave’s research agent. Ask any question about your engineering org and get grounded, source-cited answers in seconds.
How does AI-assisted output compare across teams?
Ask- Auto
- Last 13 weeks
- All repos
Or start from one of these
- “Why did cycle time spike last sprint?”
- “Which teams have the highest revert rates?”
- “What did we spend per shipped feature?”
What the answer is built from
4- 48 pull requestsgithub
- 112 reviewsgithub
- 9,204 agent runscursor
- $12,480 of spendbilling api
Any question about your engineering org, asked in your own words.
Natural language across your entire engineering dataset. No query builder, no dashboard request, no waiting on a data team.
- Median answer
- 4.2 seconds
- Claims with a source link
- 100%
- Dashboards to configure
- None
What do you want to know?
- Which engineers are using AI most effectively?
- Analyze the strengths and weaknesses of our teams
- Where are our deployment cycles getting stuck?
Every answer is grounded in your own records, and cites them.
- Q1“Why did cycle time spike last sprint?”4.2s
- Q2“Which teams have the highest revert rates?”2.8s
- Q3“How does AI-assisted output compare across teams?”6.1s
- Q4“Which AI tool has the best cost-per-merged-PR?”3.4s
- Q5“What did we spend per shipped feature last quarter?”5.7s
Four steps between your question and a cited answer.
Wooly does not guess. It resolves the question, pulls the records, separates human from AI work, and keeps a link to every source it used.
Thinking · 4 sources · verifying data points
4.2sComparing AI-assisted output per engineer across four teams over the last thirteen weeks. Normalising by codebase component so a platform change and a mobile change count the same, then checking the result against peer organisations of the same size.
AI-attributed share of shipped output · by week
- Step 1 of 4
Parse
Wooly reads the question, resolves the teams and repos it mentions, and picks the time window.
- Step 2 of 4
Retrieve
It pulls the underlying records: pull requests, reviews, deploys, agent runs and spend.
- Step 3 of 4
Attribute
Human and AI contributions are separated on every record it touches, not estimated after.
- Step 4 of 4
Cite
Every number in the answer keeps a link back to the exact record it came from.
Answers grounded in your records, not in general knowledge.
Context-aware answers
Every answer is grounded in your organization’s own PRs, reviews, deployments, and costs. Not general knowledge — your data.
Source-cited answers
Every claim links back to the underlying records. Click through from any number to the PRs behind it.
Uncover what dashboards miss
Ask cross-cutting questions no fixed dashboard was built to answer, and get an answer in seconds instead of a data request.
One answer · every claim traced
5 of 5- 642.12 units shipped48 pull requests
- 51% AI-attributedcursor · agent runs
- 99th percentile outputpeer benchmark
- 3 reverts on mobiledeploy log
- $12,480 of AI spendbilling api
No figure appears in an answer without the record behind it.
The questions a fixed dashboard was never built to answer.
What a fixed dashboard gives you
Panels chosen months ago
Someone decided which cuts mattered before you had the question.
One metric at a time
No way to join spend against output against review load.
A request queue
A new cut is a ticket, and the ticket takes days.
Numbers without provenance
Trust the tile, or go and rebuild the query yourself.
What Wooly gives you
Any question
Asked in plain words, the moment it occurs to you.
Cross-cutting by default
Spend, output, review depth and revert rate in one answer.
Seconds, not days
Fast enough to ask the follow-up while you are still thinking.
A citation per claim
Click any figure and land on the record it came from.
- 4.2sMedian time from question to cited answer
- 100%of figures carry a link to their record
- 0Dashboards to build, request or maintain
Ask from wherever the question came up.
Wooly answers in the place you were already working, and keeps the answer shareable afterwards.
Slack
Mention Wooly in any channel and the answer lands in thread, cited, for everyone reading.
Your Editor
Ask from inside the repo the question is about, without breaking flow to open a tool.
The CLI
Pipe a question into a script, a scheduled job, or your own internal reporting.
The Workspace
The full Wooly surface, with history, saved answers and shared threads.