# How an Australian AI scribe won over Canada’s doctors > Source: > Published: 2026-07-30 15:59:49+00:00 If you’ve ever waited weeks for a medical appointment or struggled to find a family doctor, it’s easy to assume Canada’s healthcare challenges come down to a shortage of physicians. But the administrative demands placed on doctors are also putting serious strain on the system. “I think the standardized use of AI in health care is not a matter of if, but when, and how.” Dr. Tom Kelly, Heidi Doctors are ordering more lab tests, referrals and prescriptions per patient than they did a decade ago, [according to research](https://www.annfammed.org/content/24/3/185), and the associated paperwork now totals nearly 20 million hours each year. That’s time that could otherwise go to care, and contributes to career burnout. Half of family doctors report [high job stress](https://www.cihi.ca/en/news/canadas-family-doctors-are-under-pressure), and more than one in five say they plan to stop practicing regularly within three years. “These lost hours are the bottleneck of the Canadian healthcare system, and the result is a workforce that is chronically overworked and under-resourced,” said Dr. Tom Kelly, co-founder and CEO of Melbourne-based Heidi, one of the world’s most widely used AI scribes. It’s no wonder doctors have been among the earliest adopters of AI. Heidi’s newly released [Canadian Impact Report](https://www.heidihealth.com/en-ca/blog/impact-report) suggests AI scribes have gone mainstream across the profession, citing Canadian Institute for Health Information data that shows 46 percent of Canadian doctors now use AI for notetaking and documentation, 13 points above the global average. The report also offers a closer look at what those changing habits can mean in practice. Across five three-month pilots, 333 clinicians used Heidi to document 111,770 patient visits. The results indicate that the tool saved an average of 13 minutes per visit, adding up to more than 25,098 clinical hours—the equivalent of 3,137 eight-hour workdays and an estimated $2.51 million in clinical time. Now the race is on to see whether AI can take on much more of the work surrounding patient care, without introducing new risks. “I think the standardized use of AI in health care is not a matter of if, but when, and how,” said Kelly. “We know that AI is already having profound impacts in healthcare. What we need now is strong governance and regulations to ensure we’re safely and ethically deploying AI tools to solve the biggest challenges facing the industry.” ### User-led innovation Healthtech has long been considered a top-down market, with governments or institutions choosing new tools before clinicians get a say in what they really need. Kelly said Heidi’s experience challenges that dynamic. Adoption has largely been driven by word of mouth, helping to make Canada the company’s second-largest global market. “Canadian clinicians are choosing us rather than the other way around.” Across pilot regions noted in Heidi’s recent impact report, every clinician surveyed said they would recommend the platform to a colleague. At one Ontario health team, 80 percent kept using it after the pilot ended, with no obligation to continue. The company sees that as evidence that designing around users’ day-to-day needs leads to more lasting adoption—and, ultimately, better outcomes for both clinicians and patients. That philosophy has also influenced what Heidi builds next. A former physician himself, Kelly said the company works closely with doctors and clinical teams to understand healthcare’s biggest frustrations, using those conversations to develop new products and features. One tool, Heidi Evidence, now gives clinicians access to trusted guidelines and research within the platform, rather than sending them to general-purpose AI tools that can cite outdated, irrelevant or even made-up information. Since launching earlier this year, Heidi says Canadian clinicians have submitted more than 1.4 million to Evidence. The company has also integrated its scribe with electronic medical record systems, including Accuro, Canada’s [largest EMR platform](https://www.heidihealth.com/en-au/blog/qhr-technologies-integrate), and moved into hardware with [Heidi Remote](https://www.heidihealth.com/en-ca/hardware), a clip-on device that captures clean audio as a clinician moves around an exam room or through the wards. Those efforts are part of a broader vision Kelly calls an AI Care Partner: a connected platform that supports clinicians before, during and after an appointment, rather than simply documenting it. ### Next steps for health tech But expanding AI’s role in healthcare also raises the stakes. To really work well depends on more than just what the tools can do, but on whether they can earn the trust of clinicians, patients, and regulators. Ontario’s Auditor General recently found that some [AI scribes produced inaccuracies in testing](https://www.hrreporter.com/focus-areas/automation-ai/ai-systems-for-ontario-doctors-hallucinate-auditor-general/394430), and that procurement processes relied largely on vendors’ own accounts of their security and privacy controls: eleven of the twenty submitted no third-party audits at all. Kelly said that is the gap Heidi set out to close on its own terms, building the regulatory and privacy standards of every country it operates in from the start rather than waiting to be asked for them. In Canada, that means storing and processing data domestically, complying with national and provincial privacy laws, and holding independent certifications for data security and responsible AI governance—verified by outside auditors rather than self-declared. In April, Ontario added [Heidi to its Vendor of Record](https://betakit.com/why-regulatory-rigour-is-becoming-a-competitive-edge-in-clinical-ai/) list. “The province has already cleared the runway, allowing providers to adopt Heidi, knowing the critical due diligence has been settled at the highest level,” he said. Kelly believes Canada is well-positioned to help shape the next phase of AI in medicine, pointing to the federal government’s [$200-million](https://ised-isde.canada.ca/site/ised/en/canadas-national-artificial-intelligence-strategy-ai-all) commitment as recognition that the field has the potential to improve not only clinician productivity, but the healthcare system more broadly. “I want to ensure that AI adoption is understood and welcomed in Canada’s healthcare systems,” he said. **PRESENTED BY** Discover why Canada’s clinicians [chose Heidi](https://bit.ly/3RKXEcj).