AITourPilot Try Henry

Accuracy your curators can check. Control that stays with you.

Your collection deserves accuracy, care and control. AITourPilot is built to keep your museum’s content and your curators’ expertise at the centre of what a visitor hears.

How a fact gets in

Everything is checked, including our own work.

A claim has to survive several independent checks before a visitor ever hears it.

  1. The work has to exist, and it has to be yours. Every object is confirmed against independent authorities, including the national records your own country keeps – before a word is written about it, and again before anything is published.
  2. What he says comes from that checked material, not from a model’s open-ended memory, and every passage can be traced back to where it came from.
  3. Then the finished text is checked back against the facts. Anything the checks raise goes to a person, and your museum has the last word.
  4. Rights are cleared before he speaks, so your museum publishes without inheriting anyone else’s.
The questions that always come

The answers your board will ask for.

  • Accuracy

    Thirteen quality gates run on every build: nine will stop a release, four flag it for review. Dozens of codified rules sit behind the guide, and each is enforced by a test.

  • Curatorial control

    You design the journey. Your experts decide what Henry may draw on, and your interpretation sets the direction he takes through the room. He supports it; he does not define it. The whole answer is its own page.

  • Sources

    Henry is built on curated, source-checked material, and your own published record is the first thing we reach for. Your curators can audit the sources behind an answer.

  • Your words, and your rights

    Henry does not copy your museum. Your own sources or public pages are read for facts and retold in his own words. Your text is never republished, and a visitor never hears it read back to them.

  • Privacy

    Conversations are kept as text. There are no audio recordings, ever.

    We do not train AI on your data, and the services we use are not allowed to either: the AI providers behind Henry are contractually barred from training on what we send them. What he draws on when he answers is text we wrote ourselves from checked facts, in his own words.

  • GDPR and security

    Everything runs under the GDPR and the data sits in the EU. There is no camera, no scanning, and no following of a named visitor around your building.

  • Governance

    Henry is a conversational AI guide, and he never pretends to be a person: if a visitor asks him, he says so.

    The rules that shape his behaviour live in the prompt behind every guide we build, and tests fail the build if anyone edits them. One of them is that he must never try to work out how old a visitor is, what education they have, or how serious they are about any of it.

Have a question about trust, data or security?