The platform
The anti-disinfo stack.
One continuous loop, from signal to correction: multi-model monitoring, claim assessment, source tracing and evidence-based corrections across 100 languages.
The loop
Detect → Qualify → Trace → Verify → Alert → Correct.
01
Detect
Multi-model querying surfaces every assertion mentioning you, your leaders or rivals.
02
Qualify
Each claim is classified — factual, contested, fabricated — against verifiable references.
03
Trace
Provenance: which sources fed the model, and how the claim was laundered.
04
Verify
Checked against primary, authoritative evidence — never opinion.
05
Alert
Drift and critical narratives trigger graded alerts to the right desk.
06
Correct
Sourced counter-narratives, ready to deploy where the record needs straightening.
Architecture
Built for breadth and proof.
- Multi-model detection
- We query supported assistants, including ChatGPT, Gemini, Perplexity, Claude and Copilot. Model coverage is defined for each monitoring programme.
- Gap measurement
- Each assertion scored against verifiable fact, producing a quantified drift you can track over time and across models.
- Provenance tracing
- We follow the claim upstream — to the sources that fed the model — and document how a weak signal was laundered into apparent fact.
- Correction generation
- Sourced, reviewable counter-narratives, with provenance attached, ready for your Comms and Legal teams.
- 100 languages
- Monitoring and correction across 100 languages — reputations do not stop at one market.
Continuous monitoring · graded alerting
See it on your organisation
See what AI says about you
A confidential audit shows the stack working on your real exposure. Findings arrive within five business days.