Trust
Built for institutions where the number has to be right
Banks and investment platforms work with Maya to track how they appear in AI answers. In these institutions a visibility figure is questioned before it is used. That scrutiny shaped how Maya collects, stores, and protects data.
Financial institutions working with Maya




Collection
A clean collection environment
Every tracked question is collected once a day on each AI platform. The question is sent exactly as written: Maya adds no system prompt, no brand hint, and no earlier conversation.
ChatGPT is collected through its consumer interface in an anonymous session with no signed-in account, from the target country and in the target language. No user memory or personalization shapes the answer.
The full conditions are in the methodology.
Freshness
Uncached data for financial institutions
For banking and investment projects Maya collects without a cache. Every daily measurement is a raw answer collected that day, not a stored answer shown again.
Every record also states whether its answer came from the platform interface or the platform API, so the origin of a number is never a guess.
Evidence
Every number traces back to an answer
Maya stores the full answer text and its source links with the question, platform, and date. A mention rate can be opened down to the answers behind it.
Unsuccessful and empty platform results are left out of the denominator instead of being counted as “not mentioned.” Competitor names are counted only when they appear in the answer text.
Security
Where your data lives
| Layer | Where |
|---|---|
| Database | PostgreSQL on AWS RDS, Frankfurt (eu-central-1) |
| Application | Vercel |
| Network and archive | Cloudflare |
| Rate limiting | Upstash Redis |
The database is encrypted at rest, including the credentials of every connected integration. Customer data is not sold and is not used to train AI models. Access is scoped to the project owner and explicit team members, API tokens are stored hashed, and every MCP tool call is written to an audit log.
Controls, sub-processors, retention, and incident response are on the security page.
Deployment
Dedicated and local deployment
For customers with stricter requirements, Maya offers dedicated deployments on local servers, including locally hosted language models.
Stored data can be kept in the country where the brand operates. In a local deployment it can stay on local servers and never be placed in the cloud. Scope is agreed per engagement.
Certifications
What we do not claim
Maya is not yet SOC 2 or ISO 27001 certified. A Data Processing Addendum is available on request, and security questionnaires are returned within five business days.
Next step
Talk to us about your requirements
Send your due-diligence questions to [email protected] or book a demo.