VetHubb Scribe: AI answers drawn from a trusted source, not the open internet

What it did

• AI scribe for vet practices. Scribe turned what a vet said or dictated during a consultation into clinical notes. It launched at a veterinary show in November 2025.
• Answers from a trusted source. Through a 12-month exclusive agreement with BSAVA, the British Small Animal Veterinary Association, Scribe had access to their clinical reference content. It searched that content from the patient findings and showed the relevant medical information alongside the notes.
• The vet stayed in charge. Notes were drafted as the vet worked, then the vet checked and signed them off before they went on record. Each note kept an audit trail, and the whole thing ran under an AI usage policy we wrote.

Why that matters

General AI tools answer from whatever they were trained on. Vets can't check where that came from, and nor can anyone else. Scribe worked the other way round: it searched a source the profession already trusts and showed what it found. A qualified person then decided what to do with it.

That's what we mean by grounded reference AI: answers drawn from your firm's or your sector's trusted source, with a person checking.

What VetHubb included

VetHubb ran in the browser on desktop, tablet and phone, with nothing to install. It worked alongside the practice system rather than replacing it.

• Ambient Scribe. The vet records the consultation in the browser and sees a live transcript. Scribe then drafts clinical notes in History, Examination and Plan format for the vet to review.
• Review and approval. Notes move from draft to awaiting approval to approved. The vet can edit before approving, and the draft and final versions are kept separately.
• Practice system integration. Approved notes are pushed into the practice's Teleos records against the right animal and client, with a timestamp on each push.
• BSAVA clinical reference. Searches BSAVA clinical reference content from the patient findings and shows the relevant information alongside the notes.
• Vet Assistant. Reads a patient's consultation history, lab results and treatment records and sets out risk factors, conditions to monitor and recommendations for the vet to consider.
• Treatment Board. An in-clinic whiteboard showing active cases and treatment plans, so the team can see what's in progress and what needs attention.
• Knowledge Base. Each practice adds its own protocols and preferred approaches, and the Vet Assistant uses them in its analysis.
• Client and patient records. Owners, pets and consultation history in one place, with each pet's link to the practice system shown.
• Consultation history. Every consultation for the practice, filterable by status, with the transcript, notes and patient history on one page.
• Workflow automation. n8n workflows route events, process consultations and schedule treatment reminders.
• Multi-practice set-up. Each practice's data is kept separate at database level, with roles for different staff and an admin area for clinics and users.

We also built an AI voice receptionist for incoming calls.

How AIOS uses the same approach

Most firms already have a trusted source. It's usually scattered across shared drives, inboxes and people's heads. AIOS puts it in one place, in plain files you own, and sets up your AI tools to answer from it. On every answer that matters, a named person checks before anything goes out.

For example:

• Your own policies and procedures. Staff ask a question and get an answer from your handbook, with the section it came from. They don't get a guess.
• Your price lists, terms and past quotes, so draft quotes and replies match what you actually charge and promise.
• Sector guidance. For care providers, that could be the DHSC guidance on using AI in adult social care and CQC's published expectations, alongside your own care policies.

It runs on the tools you already pay for, in your own accounts, and it starts the same way as every AIOS job: one process, scored first.