Ask any vet about documentation, and the answer is usually some version of the same thing: it is the part of the job they like least. Squeezed between patients during the day, finished from memory at the end of it. Late, tired, incomplete.
It is not a personal failing. It is just the way the clinical day is structured. There are always more patients than there is time, and documentation, essential as it is, will never feel as urgent as the animal in front of you.
On 3 June, in partnership with the Hong Kong Veterinary Association, Heidi brought together a panel of clinicians to talk honestly about what that pressure actually looks like, and what AI can do about it. Dr. Albert Ip, Chief Medical Officer (Asia) at Heidi, was joined by Dr. Joao Loureiro, a veterinary cardiologist at the City University of Hong Kong Veterinary Medical Centre, and Dr. Jamie Fu, a residency-trained small animal surgical registrar at CityU VMC. Over the course of an hour, the conversation moved from the structural pressures facing Hong Kong's vets to the practical realities of using AI in a busy referral clinic and a surgical practice.Here are some of the key points from the discussion.
Veterinary knowledge is outpacing the clinicians expected to keep up with it
Hong Kong has over 200 veterinary clinics and more than 1,100 registered veterinarians. The caseload is growing. The complexity is growing. The time available to each clinician is not.
Dr. Albert framed the scale of the problem simply. Medical knowledge now doubles in under 73 days, and veterinary knowledge is no different. A vet finishing their training today will see what they learned effectively replaced many times over before they retire. The idea that any one person can keep pace with that, while also running a full caseload and documenting every visit, is not a reasonable expectation.
And yet that is what the system currently asks of them.Dr. Joao was candid about his instinct as a specialist, to trust his own knowledge base before reaching for an external source. But he described what draws him to Heidi Evidence when he does use it. "What I like about the evidence is that it is referenced," he said, "and so it allows me to choose if something was mentioned in a webinar like ours, or if it is published, peer-reviewed information. I really enjoy having access to that immediately as you scroll through the data."
With clinical knowledge moving at that pace, no individual clinician can keep pace by reading alone. Tools that surface referenced, peer-reviewed evidence at the point of care do not replace clinical judgment. They take some of the weight off the clinician expected to carry it all.
Two vets, two workflows
A valuable part of the session came from Dr. Joao and Dr. Jamie speaking directly about their day-to-day experience, not what AI documentation can do in theory, but what it looks like in practice.
Dr. Joao is a cardiologist running a referral service. His cases arrive pre-loaded with history: hundred-page files from referring vets, imaging reports, medication lists. Before seeing a patient, he loads that material into Heidi and generates a structured summary built around what a cardiologist specifically needs: murmur grade, heartworm prevention status, recent medication. The visit happens. Heidi listens throughout. Afterward, the echo reports and blood results go in as context, and the referral letter comes out the other side.
"My notes look much better," he said. When he looks at his internal medicine colleagues' notes next to his own, he no longer feels the gap. It releases the stress and anxiety of being at 8 o'clock writing a referral letter, a task that previously sat unfinished at the end of most days.
Dr. Jamie's experience is different. He is a surgeon. His sessions are dominated by owner conversations about procedures, risks, what to expect. He runs Heidi during visits for ambient scribing, then opens a new session immediately after surgery, speaks through what he did while it is still sharp, and has a complete surgical report in two or three minutes. Discharge notes, post-operative calls, handover summaries, all documented the same way.
What is striking is how differently AI helped each of them. Both adapted it to how they already practice, rather than the other way around.
What Hong Kong's legal framework currently requires
For any practice working through the practical and ethical questions around AI documentation, Dr. Albert laid out the current position in Hong Kong clearly.
The minimum requirement is verbal consent from the client and documentation of that consent. AI documentation sits in a lower-risk category than clinical decision support or treatment planning tools because the function is administrative: transcribing and organising what was said, with the clinician reviewing and approving before anything enters the formal record. The clinician remains the final arbiter throughout.
On data privacy, the position is clear. Heidi does not retain audio. There is no recording to retrieve. What exists is a transcript, locked to the clinician's account, auto-deletable, and never used to train the model. These were the questions Dr. Joao worked through before committing to the tool.
"One of the things that drew me to Heidi is the certifications. Seeing that it was used by the NHS allowed me to feel some confidence that the information being shared would be safe. The fact that the sessions do not train the model, that made me feel very comfortable."
Hong Kong's multilingual reality creates a real clinical gap
There is a dimension of clinical documentation in Hong Kong that does not come up much in global conversations about AI, but matters considerably here.
Dr. Jamie, originally from Hong Kong, sees owners in English, Cantonese and Mandarin, sometimes all three in a single visit. A study cited during the session documented how multilingual interactions in Hong Kong veterinary settings, which often involve the owner, a nurse, and sometimes an additional translator, create real risk of information being lost across the chain.
AI documentation that transcribes and generates owner letters and referral notes across languages closes a real clinical gap. Some clinicians in Hong Kong are already generating owner letters in Japanese. In a city where a patient's family may not share a language with their vet, giving them a written summary they can actually read is part of what thorough care looks like.
Where this leaves us
The notes should not be the hardest part of the job. For too many vets, they still are, written between patients, finished from memory, or quietly left until tomorrow.
What the session made clear is that this is starting to change. Not through any single tool, and not without the clinician staying central to every decision. But the weight of documentation, the hours after the clinic, the fragmented records, the owner's concerns that never made it to paper, is not a fixed cost anymore.
Dr. Jamie put the practical side of it plainly: "If we work out how much time we're saving a day, that's a bit of a no-brainer."
This webinar was hosted by the Hong Kong Veterinary Association as a CPD session on 3 June 2026. Dr. Albert Ip is Chief Medical Officer (Asia) at Heidi. Dr. Joao Loureiro is a Specialist in Veterinary Cardiology at CityU VMC. Dr. Jamie Fu is a Residency-trained Small Animal Surgical Registrar at CityU VMC. All views expressed are those of the individual speakers.
