How a Convo tour
actually gets made.

Two to four weeks for a single gallery, depending on how organised your reference materials are and how many reviewers you want in the loop. The first week is drafting and your read-through; the second is voicing, translation, and a soft launch on the floor. If your materials are scattered and you want three curators to weigh in, push it to four. That is still measured against the studio-production baseline of four to nine months.

No, but it adds a few days. The first thing we do on kickoff is sit with you and pull together what you already have — wall text PDFs, exhibition essays, the curator’s working notes, the catalog entries you wrote two years ago. You don’t need to clean any of this up first. Messy is fine. We are looking for source material, not a finished script.

Two safeguards. First, the draft is grounded only in materials you uploaded — the model is not pulling from the open internet, and it is instructed to decline when it cannot ground an answer. Second, every line a visitor hears was approved by a curator in the editor. The script is not a prediction; it is an artifact your curator signed off on. If a visitor question later surfaces something off in the source, you correct the source, regenerate, and republish.

Yes. The script is a living document. Edit a stop, re-voice it across all ten languages, and republish — the change is on the visitor’s phone within a minute. There is no studio queue. This is the part of the workflow that most surprises curators who have lived under the legacy production model.

Yes. Multiple staff can be invited to the admin and work on the same tour at the same time. Convo does not yet route specific stops to specific reviewers inside the admin — most institutions handle that by assigning galleries to curators on their side and using the three-status workflow (draft, review, published) as the handoff. If you need formal per-language reviewer routing, treat that as a process step for now and let us know — it is on the roadmap.

You do. Your reference materials, your scripts, your audio, your visitor analytics. We don’t train models on your content, and we don’t share it with other institutions. You can export scripts and per-stop audio in plain formats at any time.