Meet DatennaQ
Your analysts shouldn't be searching. They should be deciding.
Ask DatennaQ to
Intelligence at the speed of a conversation
Q combines the depth of Datenna's dataset with AI reasoning — turning complex China intelligence questions into instant, evidence-backed answers.
Ask in plain English. Get intelligence-grade answers.
No training, no query language, no data exports. Just ask the question your team is already asking — and get a sourced, evidence-backed answer in seconds.
Always live. Always traceable.
Every answer connects directly to Datenna's full dataset in real time. Every response comes with complete citations and confidence scores. Nothing is a black box.
From question to assessment — in a conversation.
DatennaQ doesn't just surface data. It reasons across it — drafting multi-layered intelligence assessments that are ready for analyst review, not ready to be thrown away.
Q is being built in three stages.
Here is where we are.
No intelligence agency will trust an AI to produce decision-ready assessments on day one. This is a trust progression — data first, then decision-support, then decision-ready.
Find
Ask a question. Get structured, sourced intelligence on Chinese entities — capabilities, military ties, supplier relationships, and Western connections.
Brief
Every morning, a personalised briefing on the entities and technologies your team is tracking. The analyst's job shifts from finding the needle to validating our recommendation.
Track
The AI Intel Officer. Persistent monitoring, automated triage, and draft assessments with evidence chains — ready for the analyst to review and sign off on.
Not a chatbot on top of a database.
Anybody can wrap an LLM around a dataset. Q is different because of what's underneath — and because it was built specifically for intelligence-grade use.
Explainable by design
Every answer shows its reasoning, cites its sources, and flags where confidence is lower. No black boxes — ever.
Honest about uncertainty
If Q finds 12 entities but 3 have ambiguous resolution, it says so. An AI that admits what it doesn't know is one analysts can actually work with.
7+ years of proprietary data underneath
Q reasons over 50M+ Chinese company records and a historical baseline that new entrants simply cannot replicate.
Ready for analyst review
Outputs are structured, sourced, and audit-ready — designed to go straight to an analyst's desk, not to be thrown away.
Q · Intelligence Response
Found 8 entities. 6 have high-confidence procurement linkages. 2 flagged for ambiguous entity resolution.
| Megvii Technology | Beijing | 6 | High |
| Yitu Technologies | Shanghai | 4 | High |
| SenseTime Group | Hong Kong | 3 | High |
| CloudWalk Tech* | Guangzhou | 3 | Medium |
*Entity resolution ambiguous across 2 subsidiaries. Confidence: 71%.
Illustrative output — DatennaQ in Stage 1 (Find)
What your analysts will ask Q.
Three query types. Three different intelligence contexts. One interface.
Q is only as good as the data it reasons over.
No other AI has access to this. A 7+ year proprietary archive of Chinese company intelligence — the depth that makes every answer defensible.
Why the historical baseline matters. Q can detect when an entity that was publishing openly suddenly goes quiet — fewer papers, scrubbed website, restructured ownership. Detecting the absence of information is only possible with 7+ years of baseline data. New entrants cannot replicate this.
Built for people who can't afford to be wrong.
The working-level analysts at our customer agencies are right to be sceptical of AI. Q was built with that scepticism in mind.
Every answer cites its sources
No answer leaves Q without a full evidence chain. Every data point is traced to its original source — procurement record, patent filing, corporate registry. Audit-ready from day one.
Confidence scores, always
Every response includes a confidence level. Where entity resolution is ambiguous, Q flags it explicitly. High-confidence answers are clearly distinguished from lower-confidence inferences.
Q enhances the analyst. It doesn't replace them.
Q handles data retrieval, structuring, and initial synthesis. The analyst reviews, challenges, and makes the call. The human provides the judgement the AI cannot.
“We don't convince analysts by showing the AI is smart. We convince them by showing the AI is honest about what it doesn't know.”Jaap van Etten, CEO — Datenna Product Vision, February 2026
Q is in active development.
Help us shape it.
A limited group of partner organisations are working with us now — using Q in real workflows and feeding directly into the product roadmap.
If your team is doing China intelligence analysis, we want to hear from you.