The battle for the GP digital front door

For most of the last decade, the question of which platform sat between Australian GP practices and their patients was settled.

HotDoc won, comprehensively, by building better software than everyone else and by staying focused on GPs when no one else did. Its founder, Dr Ben Hurst, built the business as a doctor who understood what practices actually needed.

HotDoc had better recalls, better appointment booking, better patient forms, and a consumer app that millions of Australian patients downloaded and used. That focus generated a dominant market position: around 4500 GP practices, compared with Healthengine’s 1500 and AutoMed’s roughly 1000.

That dominance is now in question. Not because HotDoc’s software has gotten worse, but because AI is fast changing the game of access and engagement. It’s also changing the business dynamics around HotDoc in way that’s starting to cost it the thing Dr Hurst most carefully cultivated: GP trust.

The PE problem

Dr Hurst sold HotDoc to private equity firm Potentia earlier this year. He stayed on for five months. Then he left.

The Australian Financial Review, which reported his departure in July 2026, has subsequently published a series of damaging pieces on Potentia that read like a compressed history of a PE firm in distress.

According to the AFR, three lawsuits are currently flying around the firm, involving former deal makers and portfolio company founders. Fifteen staffers have left since 2023, including three well-regarded deal makers.

Potentia’s own AI labs director reportedly resigned after a Monday morning team meeting in August, five months after joining. The AFR noted pointedly that Potentia’s core business is “buying tech and software start-ups where venerating founding employees is the cardinal rule” — which makes a toxic reputation an acutely bad thing to carry.

Sources familiar with Dr Hurst’s departure say it was not amicable — that he disagreed significantly with Potentia’s direction and felt the business he had built to serve GPs was being managed for financial return rather than clinical purpose.

Dr Hurst retains shares but is no longer involved. HotDoc is now effectively run by its PE backers. In the GP community, that shift has been noticed.

Two own goals and a trust problem

The first sign that HotDoc’s GP relationships were under stress came when it launched telehealth on demand — a feature that would redirect a patient who couldn’t get an appointment with their own GP to an on-demand telehealth service, with HotDoc taking a commission on the diverted consult.

Practices saw it immediately for what it was: their booking engine competing with them for their own patients. HotDoc paused the feature and apologised, framing the problem as a communication failure. But the structural conflict was real.

The second own goal landed in late August, when HotDoc sent a notice to practices announcing a new $2.45 “platform fee” applying to “transactions completed through the HotDoc platform” from 1 October.

The wording was ambiguous enough that many practices read it as a per-booking charge across HotDoc’s 25 million annual bookings — which would have represented close to $60 million in new annual revenue.

HotDoc subsequently clarified the fee applies only when a patient pays through HotDoc’s own payments product, which covers less than 10% of bookable appointment types. But the communication damage was there. The other booking engines say that they got a new wave of defector practices following the incident.

The fee itself is not unreasonable in isolation. But on HotDoc’s Routine Requests product — the repeat prescriptions, referrals and certificates that have no Medicare rebate — the new flat $2.45 charge is close to three times higher than the previous percentage-based card surcharge on the same low-value transactions.

GP owners being woken up

Influential GP commentators noticed. And they weren’t taking prisoners.

In a recent LinkedIn post Dr Max Mollenkopf compared HotDoc practices to those passengers on the Titanic who thought the safest bet in the situation was to stay on the ship. 

Dr Mollenkopf, a Newcastle GP who posts prolifically on practice economics and digital health, has been making the core argument for months: the digital front door is the most important patient engagement point a practice has, and booking engines have systematically disintermediated practices from it.

He makes four specific arguments for practices abandoning Hotdoc:

• Your evolution is controlled by HotDoc — you become an end responder, not a leader, as patient expectations shift toward product-led services.
• Direct competition — HotDoc has its own clinicians servicing patients through the app competing with the practices paying for the platform and it is taking fees outside of your PMS for services the practice itself can charge for.
• The payment aggregation play — he thinks the platform fee is thin-edge-of-wedge toward HotDoc processing all practice payments and monetising the space between patient and clinician.
• Strategic blindness — practices don’t see the digital front door as a key strategic asset, leaving their negotiating position weak when the next monetisation move arrives.

Dr Mollenkopf points out that now patient thinks of HotDoc when they want to see a GP, not the practice. The reminder SMS is branded HotDoc. The repeat prescription request flows through HotDoc. HotDoc captures the commercial value.

Dr Mollenkopf is now in a position to act on this from inside a large corporate, having joined Ochre Health as both a GP and a form of strategic advisor.

Whether Ochre listens to Dr Mollenkopf’s logic or not, other major corporates are starting to think the same thing independently.

ForHealth has its own app and uses HotDoc. Although there is no indication they will jump, you’d have to think that the people who’ve transformed this business from when they bought it through a focus on better patient engagement and doctor wellbeing, would be looking at the dynamic Dr Mollenkopf is pointing out.

Oh, and ForHealth is also PE. It likes to make money as well and maybe feels that all that extra patient control and external fees would be better on the inside than the outside.

The AI layer is about to make all of this more urgent. Scribes like Heidi and Lyrebird are building agentic features that will sit between patient and practice as well – recalls, reminders, booking, and follow-up to name a few.

Best Practice’s own booking module and its Lyrebird partnership are building that capability inside the practice management software.

If agentic AI embedded in the PMS eventually takes over the upstream patient engagement functions that HotDoc, Healthengine and AutoMed currently perform, the booking engine market faces a structural threat from below as well as competitive pressure from above.

Best Practice has 80% of the GP PMS market. It’s a classic digital platform distribution choke point if it needs to be and has always represented an existential threat to the booking engines in this way.

Is it booking engines per se or mainly HotDoc?

 

Dan Stinton, Healthengine’s CEO for the past two years, is candid about where his company has come from.

“It’s fair to say that historically, Healthengine had not invested enough in our patient engagement software,” he told this publication.

“But over the last three years, it has been where the vast majority of our investment and engineering focus has gone.”

Healthengine has significantly improved recalls and reminders, automated waitlist functionality unique in the market, custom new patient forms, and an AI voice receptionist currently deflecting around 50% of incoming calls into practices using it. But in a manner that’s just a ticket to properly play.

Where Mr Stinton thinks the group can thrive is partnering with practices using AI to share the data and monetary benefits of the GP digital front door.

Healthengine now has an agentic recalls product booking patients directly into the PMS through a conversational AI exchange, achieving booking conversion rates of over 40% against an industry standard of around 10% from a standard SMS – that’s according to Healthengine.

“We are literally saving them hours and hours and hours,” Mr Stinton said.

“In some cases, days a week.”

Healthengine’s bigger play is on positioning.

“We understand the importance of practices maintaining that direct relationship with their patients through a digital front door,” Mr Stinton said, on record.

“We are here to support that, not replace that. And that is our focus.”

Healthengine is building features designed to make the practice the visible front of house -practice branding, dedicated phone numbers, patient communications coming from the practice rather than from Healthengine.

“Healthengine’s [position is to] empower the practices that use our payment software to decide what is an appropriate fee to charge.”

Healthengine still has a fairly big marketing problem: despite serving 10,000 practices across GP, dental, allied health and specialists, the GP market still largely associates the brand with the internet directory it was a decade ago.

“They think of HotDoc,” Stinton admitted, “and if they don’t like HotDoc, they go to AutoMed or us.”

Mr Stinton’s working hard on the “us” proposition based on his group being on the outside with more patient access digitally.

Healthdirect: the pragmatic government outlier

 

Bettina McMahon, CEO of Healthdirect Australia, says that the question is no longer who controls the digital front door. It is whether the health system’s fundamentals are represented at the table of whoever does control it.

She suspects that the main controlling element may end up being the large language models – ChatGPT and Claude – and this probability now needs to be carefully thought through.

“If you can’t fight them, join them,” she said this week.

“Our model isn’t commercial. We don’t have to get eyeballs on Healthdirect. We just have to get quality health information in the hands of Australians. The channel doesn’t matter so much.”

Ms McMahon, who once ran the Australian Digital Health Agency, is a very creative thinker.  She is working on how to insert the role and data of Healthdirect into the protocols of the AI platforms that Australian patients are already using. And she’s working closely with the Department of Health, Disability and Ageing, to get them across the line on something a typical government department would see as a governance and regulation nightmare.

When a patient asks Claude or ChatGPT a health question she wants to make sure that they get directed toward credible clinical services and evidence-based information, rather than toward seductive and dangerous health misinformation that is already proliferating on these consumer platforms.

Ms McMahon acknowledges it is not easy.

US technology companies operate under different constraints than Australian government-funded bodies.

But she is raising the sense of urgency internally and finding support.

“If we continue the current trajectory,” she said, “someone else will be controlling who comes to see us, whose interests aren’t aligned with the Australian health system or with Australian consumers.”

This approach seems to stand in contrast to the Australian Digital Health Agency’s primary current focus: Provider Connect Australia, the agency’s ambitious national provider directory program.

It’s an all in-house play. 

All government firewalled infrastructure, very little external engagement, and certainly none contemplated so far with the big LLMs.

The Agency has a trump card now in trying to pull this master directory off: it’s taking data directly from Services Australia and AHPRA about every provider including their IHIs. That’s going to give the directory a reasonable spine and the opportunity to hook this data directly to Medicare is important. Providers will need to opt out, like consumers need to opt out of the My Health Record, if they don’t want to be part of it.

But if you’ve talked to  people who have built large internet directories before, including this writer (I managed smart people who did) Provider Connect still has the air of an early 2000s solution to a 2027 problem.

Those of us who built internet directories in the early 2000s on the back of Google and SEO watched Google systematically extract that data and use it for its own purposes, killing the independent directory model in the process of taking all the money as well.

The LLMs are going to do something structurally similar to any government-built provider directory that doesn’t actively integrate with the platforms patients actually use.

There’s also a lot of other work for the agency to get this piece of infrastructure done: it needs legislation, it’s relying on buy in from multiple tribal medical sectors and software vendors, it’s being built for a plan to build a national Health Information Exchange, which may also now be redundant in the brave new world of network effects being seeded by new AI ventures.

The vital thing they will be able to do is hook the provider data, which is generated by Services Australia on a regular basis via the need for providers to stay registered, to Medicare and presumably even to billing, and then other services…maybe even in an Agency built app for patients.

As usual its pretty good thinking.

But they need to think just a little about how the My Health Record went off the rails.

The My Health Record was, when it was conceived, a good idea but it was very quickly outdated by the rapid evolution of the digital world from centralised data management to distributed management via new cloud technology.

When it became obvious that the agency should pivot, it instead doubled down on its efforts to make the project work, largely because of political pressure and the enormity of the sunk costs.

And then it wasted a further $1.5 billion and about five years of infrastructure development. 

Important additional data for a functional national provider directory exists, distributed across HotDoc, Healthengine, AHPRA, Healthdirect, HealthShare and others. All the LLMs will have much of this data and more eventually. And because the LLMs will be used near universally by both patients and providers, the LLMs have the means and money to keep the location and engagement protocols of all our providers updated, especially if we were helping them.

But so far the Agency doesn’t appear to be contemplating how the LLMs might effect how patients or providers behave in an AI world.

The Agency is decidedly light on AI thinking.

But AI is to them what the cloud was to the My Health Record developers around 2004.

It needs to be contemplated in the bigger picture, like McMahon is contemplating how it affects her organisation with respect to the provider digital front door.

The ADHA should absolutely keep building the unglamorous but necessary infrastructure it has been building – script exchange, FHIR standards mandating, the data spine work underpinning clinical interoperability, including a form of directory, but with the LLMs role considered.

But the smarter play is to federate what already exists, work with the existing directory providers who have massive patient engagement today already, work out the privacy and safety frameworks with the AI platforms, and work with the LLMs if they can to get Australian health system logic into the information channels that patients are already using.

HealthShare, which runs the best private specialist referral directory in the country and is deeply integrated into Best Practice, has indicated it would work with government on this.

Its engagement of its specialist product is proof that the data in that group is gold.

What this means if you are a GP

The forces described in this piece are not abstract. They will hit the average GP practice – big or small, urban or rural, corporate or otherwise – in concrete ways over the next two to three years.

Your booking engine is no longer just a booking engine.

If you use HotDoc, you are embedded in a platform that is currently focusing on building itself into the payment and transaction layer of your practice, capturing the commercial value of your patient relationships, and in some cases directing those patients elsewhere.

Whether that trajectory continues or accelerates under PE ownership is not yet clear. It might pivot, as Healthengine says it is.  But this direction is visible now.

Switching is probably going to be hard despite what ForHealth says… if it was easy, it would probably have done it.

Notably, IPN, owned by Sonic, did a couple of years ago. Sonic tends to be a quiet achiever. It is smart about data and engagement if you look at the assets it owns. It seemed to have seen this problem the earliest and decided to take the reins on the relationship itself.

HotDoc’s consumer app install base is its most important moat.  

Patients have HotDoc downloaded, they use it habitually, and a practice that moves to Healthengine or AutoMed has to manage the transition of patient behaviour, not just change a software subscription.

HotDoc is a brand, and in the minds of patients it’s still a pretty good one.

This is rubber hits the road type friction if a practice or corporate wants to leave and it probably shouldn’t be underestimated.

A lot of things can happen now.

HotDoc’s owners might already be rethinking their position.

Smart operators like ForHealth might also be as well.

Let’s talk about what a group like HotDoc, Healthengine and AutoMed need to do for us not to leave them. That might be a good middle ground that saves both parties potential transitional heartache.

The clock is ticking though.  

Practices that wait until the economics are clearly worse will find the transition much harder than those who move now and either map a different path or start to negotiate.

If you are a larger practice or a corporate group, the question Dr Max Mollenkopf is asking is, do you want to own your digital front door or rent it from a PE firm?

The technology to do more yourself exists and is improving.

AutoMed’s model, which sits inside the practice rather than competing for the patient’s attention from outside, is one starting point.

Healthengine’s partnership model, for all its brand perception problems, is closer to what practices actually need than HotDoc’s current direction.

And the agentic AI features being built into practice management systems and being built out rapidly by other key players like Lyrebird and Heidi will, within two to three years, begin to perform many of the upstream patient engagement functions that booking engines currently own.

Lastly, the AI platforms are coming whether you plan for them or not.

Australian patients are already using Claude, ChatGPT and Healthdirect when they want to understand a symptom or find a service.

As those platforms get better at directing people to specific providers, and they will, your practice’s presence in those systems, your data accuracy in the directories that feed them, and your ability to receive and respond to patients who arrive through those channels will matter enormously.

A practice with a well-managed digital presence, accurate and complete information across all relevant platforms, and a direct relationship with its own patients through its own app or booking system, is a practice that will benefit from this shift.

One that outsourced its digital front door entirely to HotDoc five years ago and hasn’t thought about it since is one that may find itself increasingly invisible.

Your digital front door is changing. Big time.

GP practices that understand what is at stake and act on it now will be better positioned than those who wait for someone else to sort it out.

The GP digital front door and patient engagement will be debated live at Burning GP, Noosa, 26 September 2026. Program and tickets HERE –  use promo code BGP_30 for a 30% discount.

The post The battle for the GP digital front door appeared first on Medical Republic.