AI agents booking doctor’s appointments: Inside PocketHealth’s drive to improve patient care

It’s a critical time for Canada’s economy and its business community. The country needs innovation and growth more than ever, yet businesses across a wide range of industries are facing trade disruptions, rising costs and technological change, among other challenges. This series profiles forward-thinking Canadian companies that are growing their markets by re-examining their business practices and trying new things.

When Harsh Nayyar injured his ankle playing tennis while living in California’s Bay Area in 2014, he received his MRI images on a CD.

For Harsh and his brother Rishi, both technologists, the experience seemed shockingly outdated. Tech platforms such as YouTube and Netflix offered instant online streaming, yet medical images still needed to be burned onto physical discs.

The brothers decided to tackle the problem together, launching the Toronto-based software development company PocketHealth in early 2016 to help patients access, share and understand their medical imaging records and reports – all online, from any device. Rishi is chief executive officer and his brother Harsh is chief technology officer.

Ten years in, PocketHealth’s platform has reached more than three million patients across Canada and the United States and operates at more than 900 sites across more than 250 health care organizations, such as UHN, Sinai Health, Trillium Health Partners, Mackenzie Health, William Osler Health System and Niagara Health. Its annual revenue, the company says, is in the $10-million to $50-million range, growing more than 25 per cent year over year.

Rishi says PocketHealth’s medical image sharing services help patients discover the results of a biopsy or MRI quickly and efficiently, plus patients can more easily share findings from an emergency department visit with their family physicians.

The company also has examples of patients who spotted items missing in their report that they were then able to get corrected, he says. “It changed the path of their care.”

Now, the company is pursuing its second act, expanding beyond medical-image access into agentic artificial intelligence that can automate administrative tasks in hospitals and clinics, such as processing referrals and scheduling appointments.

This innovation is a response to health care organizations grappling with disconnected record systems and demanding administrative work that takes time away from patient care, Rishi says.

“We are in a heavily overburdened health care system.”

A cure for fragmentation

Dr. Andrew Pinto, a physician and director of Upstream Lab, a non-profit research lab based at MAP Centre for Urban Health Solutions, St. Michael’s Hospital in Toronto, says that one of the big problems with Canada’s health care system is that it is enormously fragmented. Individual hospitals, clinics, primary care organizations and public health have their own record systems, meaning there isn’t a single electronic health record for the country.

As a result, each of these systems has a different way of defining data, Dr. Pinto says, and data can’t flow easily from one system to another. While PocketHealth came up with a solution that solved this fragmentation, ideally the fragmentation wouldn’t exist, he adds.

“They are bringing a new perspective to a very old problem and they are showing us what technology can do,” Dr. Pinto says, adding that Canada needs legislation that would mandate interoperability.

Automating time-consuming tasks with AI

PocketHealth’s agentic AI platform, PocketHealth Conductor, launched in late 2025. Rishi explains that Conductor has a coordinated network of AI agents that can complete non-clinical tasks autonomously.

For example, when a physician sends a referral, Conductor processes it, calls the patient and immediately books an appointment live over the phone. It’s also able to read faxed and uploaded documents.

The aim is to automate the health care industry’s time-consuming administrative workflows so providers can spend more time on patient care. And PocketHealth says Conductor is live in production today at “dozens” of hospitals and imaging providers in Canada and the U.S.

But adoption isn’t without its challenges, Rishi says.

“No matter how incredible [AI] is, every health care system, every clinic we deploy in has its own unique way of performing tasks,” he explains. In other words, a company can’t install an AI system without adapting it. Also, there’s a lot of ongoing effort to make agentic AI systems feel human and improve – rather than impede – the care journey.

“The obstacles are that it still requires a lot of configuration work, a lot of testing, which means you have to test very carefully, you have to roll out slowly, and I think we have very, very ambitious plans,” he says.

From Dr. Pinto’s perspective, AI has great potential. But he points out that the health care industry needs to address equity concerns around the technology, such as whether it’s lifting everybody up, whether it’s accurate and whether it works well for groups of people that may not have the digital literacy or devices to use it.

Leveraging a technology in its ‘infancy’

Investors are interested in what PocketHealth is doing. The company raised $45-million in Series B financing in March 2024, bringing their total equity funding to $75-million. The Series B financing was led by Round 13 Capital, with Deloitte Ventures and Samsung Next joining as new investors and additional investment from Questa Capital and Radical Ventures.

Rishi says that at least for now, PocketHealth plans to concentrate on Canada and the U.S. as it expands its AI offerings, rather than enter new international markets.

He hopes the technology will enable patients to reach out and schedule their own exams, and better understand the preparation needed for procedures instead of just receiving a phone call or a letter in the mail.

“I want it to be that they can call in and they’re never on hold and they get their questions answered,” he adds. As well, agentic AI can help new immigrants get the same quality and access of care in their first language as English speakers do.

“Agentic AI is still in its infancy,” Rishi says. “It needs more case studies, more deployment stories. We’re not even in inning one.”