When we founded Lyrebird in 2024, we started with a simple question:
What if AI could take care of the documentation, so clinicians could get back to what really matters?
2024 was about possibility.
2025 was about reality, and what clinicians taught us.
2025 was the first year we saw AI scribes used at real scale. By mid-2025, nearly a quarter of Australian GPs surveyed reported using an AI scribe in their practice, a shift that would have seemed improbable not long ago.
For us, 2025 became a year of deeper learning. Time spent both inside clinics and in research settings helped us see more clearly what genuinely shifted once scribes became part of routine practice, and what remained unchanged.
Impact was measured in more than minutes
The first change we expected to hear was time saved. It’s one of the most widely discussed benefits of AI scribes, and our research showed the same pattern.
But when we spent time in clinics and asked people about their experience, time didn’t dominate the conversation. It came up, but only briefly. What surfaced far more strongly was how the work felt different, something harder to quantify, but raised repeatedly.
In many of the practices and services we spent time in, conversations often returned to similar themes: attention, mental load, and the sense of having more space to think. One allied health practitioner put it simply: “Lyrebird has helped me be more present with my patients.” At Gold Coast Health, a doctor described the same shift physically, there was “no longer a keyboard between us.”
Patients noticed the difference too. Reporting more eye contact, smoother interactions, and a greater sense of being heard. In the Gold Coast trial, 68% said their doctor spent more time directly engaging with them. This was one of the most consistent and meaningful signals from across all research sites.
In a study conducted with King’s College London in Dubai, participants shared that the structured support “helped me organise the information more clearly” and allowed them to “think more about the patient, not the form.”
We also began quantifying this using the NASA-TLX, a tool designed to measure cognitive workload. In this study, the change was substantial: total workload dropped from 461.3 to 25.0, with marked improvements across five of six subscales, including a 50-point reduction in temporal demand. Eighty-seven per cent of clinicians reported that the tool reduced their cognitive load.
This year brought into clearer focus the parts of the administrative burden that are harder to quantify. Cognitive space, presence, and how the day actually feels. Through our research, we were able to start putting numbers around those shifts, and the conversations we had showed how meaningful the change can be when even some of that cognitive weight begins to lift.
What the time made possible
Across the practices we spent time in, people described time in terms of the small parts of their day that had returned. Dr Margaret Hall told us: “Since using Lyrebird, I’ve not only been able to relax and unwind in the evenings. I’ve had time to watch a movie mid-week which used to be a luxury.” A practice owner told us his new year’s goal was to buy a Harley. He was able to double my Practice revenue and buy three. Allied health teams talked about finishing their summaries during the day instead of on weekends.
These customer success stories gave us something rare: a clear view of what actually changes when AI scribes become part of everyday clinical practice.
What stayed with us wasn’t the scale of the change, but what it meant to each person. People didn’t talk about reclaimed minutes; they talked about reclaiming parts of their day they had lost.
That empathy is our engine.
What we’re still learning
While scribes eased the burden of note-taking this year, the rest of the administrative work; care plans, assessments, letters, referrals, stayed the same.
These tasks still required the same number of steps, the same switching between systems, and the same end-of-day effort. And for many teams, a scribe simply became one more tool in a sea of other solutions.
Clinicians pointed this out quickly. Once note-taking became lighter, a different set of questions started to appear:
- “If the information is already there, could it help with the care plan?”
- “Why do I still need three systems to complete one consultation?”
- “Does it have to take this many steps?”
With deeper integration in place, we were able to explore some of these questions alongside partners. Chronic care plans were one example. These plans often required multiple screens, manual template work, and were frequently pushed to evenings or weekends. Using the consult transcript alongside a patient’s history in Best Practice, the plan could be populated in real time and saved directly back to the record.
For many clinicians we spoke with, this meant a task that once took 15–20 minutes was completed before the patient left the room, rather than becoming end-of-day or weekend work.
This work is still early, but it surfaced an important shift in perspective: once a single part of the administrative burden changed, it became natural for people to question the rest. Tasks had simply been accepted as “the way things are,” and what else might be possible if the system around them changed too.
Looking ahead to 2026
2025 showed us that AI scribes work in real clinical settings, but it also showed us that documentation is only one slice of the administrative burden.
AI scribes will evolve from capturing the consult to shaping the workflow around it.
Clinicians described the biggest value not in the draft itself, but in the steps that follow: completing forms, closing loops, preparing letters, and ensuring information lands where it needs to. The next wave of tools won’t just record the encounter, they’ll help move work forward.
Structured data will become the quiet engine behind clinical workflows.
As documentation and administrative requirements continue to grow, health services are placing increasing emphasis on data that is reliable, traceable, and immediately usable. Notes will still matter, but the structured elements around them will become the starting point for automation and care coordination.
Expectations around reliability and governance will rise. With national reviews underway, the conversations this year made it clear that accuracy is only one part of the equation. Organisations are looking for systems that behave predictably, handle data transparently, and provide confidence that they can be audited and trusted in high-stakes environments. This is a positive step and it raises the standard for everyone.
Integration will become the differentiator, not an add-on. Clinicians consistently told us that the tools which made the biggest difference were the ones that removed steps, reduced switching, and lived inside their existing workflows. In 2026, integration will shift from an efficiency preference to a core requirement.
Clinician wellbeing will become a design principle.
Across our research sites, one of the most striking themes was how deeply documentation shaped the emotional and cognitive load of clinical work. As organisations prioritise sustainability and retention, AI tools that meaningfully reduce this burden will carry greater weight.
2026 won’t just be about better notes. It will be about building the infrastructure that allows clinicians to spend more time in the part of their job that only humans can do.
