Lukas Saari remembers picking up the phone with trepidation.
Over a few intense weekends in 2023, Lukas and his co-founders, Oliver Åstrand and Oscar Boldt-Christmas, built the first version of Tandem: not just as a product, but as an entry point into something larger. They’d started putting the product into the hands of doctors to gather real-time feedback and establish clinical authority behind what they were building.
At this stage, every piece of feedback was crucial, and every doctor’s opinion was instrumental. That’s why, when the phone call came from the first senior doctor who had just started using their product, Lukas felt his heart beat quicken. But the feedback he received speaks for itself.
“Not only did he love the product, he was so impressed with it that he wanted to invest his own money in the business,” Lukas recalls, beaming.
It’s feedback like this that has been one of the most rewarding aspects of Lukas’ journey to date. The speed at which customers fall in love with Tandem Health, and at which they turn from users to advocates.
This cuts to the very core of what differentiates Tandem. “Many healthcare solutions come from the decision maker perspective rather than the end user perspective,” Lukas says.
For those end users, Tandem doesn’t promise “superpowers” or “superhuman efficiency”; it promises something much more valuable—time and freedom. And like a new breed of technologies, its success lies in becoming invisible, fitting into existing workflows rather than forcing new ones.
Solving healthcare’s biggest inefficiencies
Many of us will remember that first time we experienced technology that we could only describe as “magic”. For Lukas, that was in 2022, when he first tried out ChatGPT. While others were writing love poems or satirical covers of pop songs, Lukas was becoming quietly obsessed with this new AI frontier.
“It felt like the largest platform shift of our lifetime.”
As a problem-solver by nature and a consultant by profession, Lukas quickly started scoping out problems that could be quickly addressed using generative AI. Leading McKinsey’s Healthcare Analytics practice in the Nordics at the time, he thought deeply about opportunities for bringing together healthcare and AI.
One such problem stood out: clinician workload, and its inextricable link to burnout. One problem stood out: clinician workload, and more broadly, a system built around administration instead of care. It was a dysfunction that saw medical professionals becoming overworked, overstretched, and burned out, the latter of which is believed to impact over half of clinicians (with around 20% reporting depression). The result is potentially colossal, pushing doctors to breaking point, shift reduction, and in many cases, to quitting altogether. In a recent survey of the NHS, 15% of clinicians referred to taking serious steps towards leaving the NHS as a result of burnout.
So where were doctors wasting their precious time? Where was the system losing critical minutes and hours that could be the difference between life and death?
Given Lukas’ technology background, this of course zoomed in on digitization and what he refers to as “healthcare’s broken tech stack”. And after going straight to the source, Lukas and his co-founders quickly unearthed one prominent problem: documentation. Specifically, the hours spent every day writing up medical notes during and after every appointment.
Lukas saw a possibility—to make a huge impact on the lives of doctors by unlocking hours of their time. This, in turn, could transform healthcare as we know it.
More time practicing medicine, less time practicing administration
Almost instantly, Lukas knew that generative AI could practically eradicate clinicians’ time spent on medical documentation. Just as notetakers or scribing tools had taken other industries by storm, they were equally well placed to transform healthcare.
Over the next few weekends, Lukas and his co-founders, Oscar and Oliver, set about building a Tandem MVP. A tool that listens during consultation, transcribes conversations, and automatically drafts medical notes. Doctors would simply need to review and approve these drafts.
There was a beautiful simplicity to this, especially in the complex realm of healthcare innovation. Not only did Tandem promise productivity, it could also plug into existing workflows and integrate with current systems. And best of all, it avoided one of the biggest bottlenecks to healthtech.
“By starting with documentation, we could move quickly while still building for the realities of clinical safety and regulation.”
That said, they had to make sure that their end users were interested in using it. So just weeks after they started, they were already shipping the product into the hands of doctors around Sweden. They had a simple core belief that they had to prove to be true—why would doctors spend 10 minutes writing notes from scratch when they could spend minutes or less reviewing and accepting drafts?
Positive indicators emerged immediately—clinicians going from consultation to reviewed notes in around a minute, with adoption quickly spreading across clinics. “Clinicians could spend more time practicing medicine, and less time practicing administration,” Lukas says.
On a macro level, they began to glimpse the bigger opportunity at stake: to unlock huge efficiency gains across healthcare systems that were stretched thin by staffing shortages and ballooning healthcare challenges. This was a global problem, and demanded global ambition.
“Luckily, we dared to be ambitious early on,” Lukas says.
The advantage of building in Sweden, Lukas adds, is that the size of the market means that you are thinking internationally from day one. In European healthcare, he knew that speed alone wasn’t enough: you have to scale with trust.
Expansion and keys to success
And so, in 2024, Lukas, Oliver, and Oscar made a huge leap. They launched into the three biggest European markets—the UK, France, and Germany—simultaneously.
There was no way they could leave a strategic expansion of this scale to chance. For success, they knew they had to set a very high bar for talent. Lukas recalls some of the best advice he received: “Identify the competency gaps you have as quickly as possible, and hire excellent people to solve them.”
He nods first and foremost to his fellow co-founders, whose skill sets deftly balance the requirements of a high-growth startup. Oliver, his CTO, was his number one pick for technical co-founder, bringing extraordinary academic credentials in engineering and machine learning, plus experience launching AI startups. Oscar, his CCO, was Lukas’ colleague at McKinsey, having been a senior partner at the firm’s healthcare practice leading around 1000 consultants.
And for each geographic expansion, Tandem hired experienced operators to lead, individuals Lukas dubs “a Tandem CEO in each country”. Take Dr. Carol Wildhagen, their DACH country director: a medical doctor by background, with seven years at BCG focused on healthcare systems before launching her own health tech company. Hiring leaders of this caliber, Lukas says, means he can entrust others with handling both 0-to-1 scaling and 1-to-10 scaling.
More broadly across the team, they’re looking for people with a “track record of exceptionalism”, individuals who are highly ambitious in everything they do and thrive in challenging environments.
Lukas’ high bar for talent is matched only by the incredibly high bar that he sets for the platform itself. By design, the team is structured to sit incredibly close to their users, gather ongoing feedback, and drive continuous improvement into the product.
Tandem’s go-to-market team actually doubles as a product team, combining clinical knowledge with technical know-how to ensure the best possible prompts and outputs for their AI.
They also host regular workshops with the CIOs of their largest customers to hear their feedback, pain points, and understand how Tandem can incorporate this into their solution. It helps that more than 80% of their commercial team has a medical background. “They’ve been in the shoes of our users,” Lukas says. “That gives us a big advantage.”
Surfing the AI wave
While it was ChatGPT’s transformational potential that caught Lukas’s eye, Tandem Health has never been about building AI for AI’s sake. His motivations have always been about using technology to solve specific but universal problems. Tandem’s AI capabilities matter only as much as its ability to solve problems for its end users.
For that reason, Tandem has always been “LLM agnostic”. Whatever’s the fastest, cheapest, most capable model available—Tandem adopts it.
“This means our users get the benefit immediately, rather than waiting for us to develop it.”
Naturally, bringing AI into a highly regulated and complex industry hasn’t been without its challenges. Trust isn’t something added later; it’s built into the system, especially given the sensitivity of the patient and clinician data that they are handling. Every new Tandem product must meet strict clinical safety standards, given that its business is quite literally a matter of life or death.
Regulation remains an ongoing source of complexity. “What surprised me most is how ambiguous the regulatory landscape is,” Lukas says, witnessing how different regulators (let alone the clinics themselves) will interpret the same rules differently.
The combination of these factors creates an ongoing tension, where the pace of AI doesn’t match the pace of healthcare. “Much of healthcare is publicly funded. That means public procurement frameworks, which extend timelines quite a bit.”
But challenges aside, the scope of opportunity remains dizzying, as does Lukas and his co-founders’ ambition. Tandem is already used by over 5000 healthcare organizations across Europe. Their NHS partnership via Accurx reaches 200,000+ NHS staff, including 98% of GP practices in the UK. Lukas describes 2026 as the year Tandem transitions from “AI medical scribe to complete AI medical assistant”; down the line, he aspires for Tandem to be the AI operating system for clinics.
Lukas’s aspirations may be zealous, and yet also make perfect sense. An AI assistant that can support before (triaging patient requests before appointments, summarizing medical history for clinicians), during (providing real-time clinical guidance in consultations), and after (automatically handling referrals, prescriptions, and follow-ups).
It’s this that makes Tandem’s a particularly striking vision: not because of the potential of the technology, but because of what it means for the humans who benefit from it most. It’s a vision where the best technology is not overbearing; it’s invisible. It disappears into workflows, creating more time for the human part of medicine and offering the best possible care.
So if Lukas succeeds in this vision, the most important thing about Tandem won’t be what its software can do; it will be about what its users no longer have to think about.
.png)
