Lens Intelligence
If it can't cite it, it doesn't say it
Lens Intelligence answers questions using only the project's own documents, request responses and notes — and every answer names the sources it drew on, so a reviewer can check the claim instead of trusting the tone.
Grounded Q&A
Ask the deal, not the internet
Ask what the target's release cadence looks like, where key-person risk sits, or which security controls are evidenced. The answer comes back with the documents and responses behind it, and a clear statement when the evidence is thin.
- Named citations on every answer, linked back to the source in the dataroom.
- Confidence labelling, with explicit gaps rather than confident guesses.
- Scoped context — you choose which workstreams and entries are in play.
Source readiness
You can see what the AI could see
Before you rely on an answer, Lens shows how much of the intended source set is actually indexed and readable — so a confident answer over 12% of the dataroom is never mistaken for a complete one.
- 01
Scope
Select the workstreams, folders, request responses and journal sections the model may read.
- 02
Check readiness
See coverage across the selected sources before generating anything.
- 03
Generate
Produce an answer or an insight module, versioned and timestamped.
- 04
Review
A human accepts, edits or rejects. Nothing publishes itself.
Insight modules
Structured output, not a wall of prose
Highlights
The short read for a partner with four minutes.
Key findings
Material observations with workstream and evidence.
SWOT
Structured position from the assembled evidence base.
Risks & signals
Emerging concerns, including what the data does not show.
Evidence
The source set behind each module, versioned.
Follow-ups
Suggested information requests to close the gaps found.
Human + AI
The model drafts. The team decides.
Lens Intelligence compresses the reading, not the judgement. Every generated artefact is attributable, reviewable and version-controlled, every answer names the documents behind it, and the human reviewer is recorded alongside the output.
How we work →See grounded AI on your own evidence
A 30-minute walkthrough of diligence, integration and value creation in one system.