With €86 million in fresh funding, TANDEM HEALTH wants to become Europe’s clinical AI layer

Swedish scale-up Tandem Health has raised around €86 million in a Series B led by EQT’s Scaleup Europe Fund. Beyond the size of the round, the financing reflects a broader shift in category: after automating clinical documentation, Tandem now wants to operate before, during and after the consultation. Already present across 14 European markets, the company faces a challenge that is distinctly European, building a common platform for healthcare systems that remain deeply national. Against Nabla, Doctolib, Corti, TORTUS and a growing number of local specialists, the contest is no longer simply about building the best medical scribe. It is about controlling the clinical workflow.

Founded in Stockholm in 2023, Tandem Health has announced a $100 million Series B, equivalent to approximately €86 million, led by the Scaleup Europe Fund managed by EQT. Kinnevik, Northzone, Amino Collective and Visionaries also participated. The financing comes little more than a year after a Series A worth roughly €43 million, and two years after an approximately €8 million seed round. Across the three rounds, Tandem has now raised close to €138 million.

The pace of financing reflects the pace of deployment. Tandem says its technology is now used by 10,000 healthcare organisations across 14 European countries. The new capital is intended to accelerate that expansion, but more importantly to broaden the product itself. The company no longer wants to provide only a medical assistant. It describes its ambition as building an “AI-native clinic operating system”, capable of managing patient flows and taking on a growing number of clinical tasks.

This is less a product extension than a move into a different market.

The medical scribe was only the entry point

Tandem’s initial value proposition was relatively straightforward. During a consultation, its software listens to the conversation between clinician and patient, extracts the relevant information and generates a structured note that can be reviewed before being entered into the patient record.

The use case tackles one of healthcare’s most immediately monetisable problems: the amount of clinical time absorbed by documentation. Tandem says practitioners using its scribe reduce the time spent on administrative work by 29%.

But the scribe has another strategic advantage: it provides an entry point into the workflow at the moment clinical information is created.

Once installed during the consultation, the assistant is no longer dealing with speech alone. It has access to the context of the appointment, receives clinical information, creates documents, interacts with the patient record and begins to understand how the clinician works.

From there, the distance between documenting an action and preparing, supporting or executing the tasks around it becomes relatively short.

Tandem has already crossed several of these boundaries. Before an appointment, its assistant can summarise available information. During the consultation, it generates documentation and can provide contextual clinical information. Afterwards, it can produce notes and suggest relevant codes.

The progression is increasingly clear:

scribe → assistant → agent → workflow orchestration.

From clinical notes to a clinic operating system

The change may sound semantic. Economically, it is much more significant.

A medical scribe essentially sells time saved per consultation. A platform capable of preparing an appointment, documenting the exchange, coding procedures, retrieving clinical guidance, managing parts of the patient relationship and automatically updating existing systems can capture a much larger share of a practice or hospital’s software spend.

The logic mirrors what is happening in other AI verticals. The first generation automates a clearly identified task. The second absorbs several successive steps. The third tries to become the layer from which users initiate and supervise their work.

Healthcare adds another consequence. The company that becomes the physician’s daily interface may gradually position itself above the practice-management software and electronic health record, even though those systems have historically formed the core layer of clinical IT.

The issue is therefore no longer just who can produce the best clinical note. It is who can become the interface through which existing software is controlled.

This is also where Tandem’s European ambition becomes considerably more difficult.

The software can be European. Healthcare systems remain national

Behind Tandem’s presence in 14 countries are markets that remain fundamentally different. Languages vary, but so do electronic health records, coding systems, reimbursement mechanisms, procurement processes, technical infrastructures, local requirements and the organisation of care itself.

Tandem is already using different market-entry strategies across Europe.

In the Netherlands, it chose acquisition. In January, Tandem acquired Juvoly, a Dutch AI documentation company that said it was used by more than 1,500 general practices, representing around 35% of Dutch primary-care practices, and processed more than 200,000 consultations per month. Tandem retained the local team and market presence — an implicit acknowledgement that national healthcare markets cannot simply be conquered from Stockholm.

In Finland, the strategy relies more heavily on integration with existing infrastructure. Tandem partnered with CGI to connect its assistant to OMNI360, a patient-record system used by tens of thousands of healthcare professionals and covering care for around 1.5 million Finns. In August, the wellbeing services county of Western Uusimaa, LUVN, selected Tandem following a procurement process, with the aim of making the technology available to approximately 7,000 professionals by the end of 2027.

In the United Kingdom, the entry point is the NHS and its procurement mechanisms. University Hospitals of Leicester and University Hospitals of Northamptonshire signed a four-year agreement giving more than 10,000 clinicians access to Accurx Scribe, powered by Tandem, across roughly 2.5 million consultations a year. The technology was selected after several months of evaluating five suppliers.

In Spain, Tandem has entered through large private healthcare operators. Recoletas selected the company to deploy its assistant across more than 50 hospitals and clinics, following an initial pilot. The group carries out more than two million consultations annually.

Acquisition in the Netherlands, integration with a major software provider in Finland, public procurement in the UK and partnership with a private hospital group in Spain: there is still no single playbook for scaling a clinical software platform across Europe.

That fragmentation is one of the main differences between healthcare and more conventional software markets. The same technology can be available in 14 countries, but adoption remains a market-by-market localisation exercise.

Europe is therefore both the addressable market Tandem wants to capture and the complexity it must learn to master.

In France, Tandem enters an already crowded market

France is a particularly useful test case. Tandem says it is already used by close to 200 healthcare organisations, including facilities belonging to Ramsay Santé, and the country has become its second-largest European market. A local team of around 20 people now supports that expansion.

But the French market is no longer a simple contest between AI medical scribes.

The most directly comparable competitor is Nabla. The French company also started with ambient clinical documentation before gradually expanding its scope. It now describes itself as a “Clinical AI Layer”, positioned from pre-consultation preparation through coding and agentic task execution. Nabla says its technology is used by more than 85,000 clinicians, across 130 healthcare organisations and more than 20 million consultations a year.

The strategic similarities are striking. Both companies began by eliminating part of the physician’s documentation burden. Both are now trying to turn that initial position into a deeper infrastructure layer.

Doctolib is moving in the opposite direction.

Doctolib starts with the workflow, and already has European distribution

In France, Doctolib does not need to conquer the clinical workflow. It already controls several of its main entry points.

Its Consultation Assistant listens to the doctor-patient exchange and produces a structured summary. Its broader product suite already covers appointment booking, patient communications, medical records, teleconsultation and parts of billing, while the company is progressively adding telephone automation.

The strategic movement is therefore reversed.

Tandem and Nabla start with AI and move up into the workflow. Doctolib starts with the workflow and progressively adds AI.

The confrontation is no longer purely French. Doctolib says more than 570,000 healthcare professionals use its services across France, Germany, Italy, the Netherlands and the UK. In Germany, the company now sells its AI consultation assistant as a standalone product, meaning doctors do not necessarily need to use its full practice-management software.

Doctolib therefore has an advantage that AI-native companies still need to build: distribution that predates the AI product.

The relationship between the two models is nevertheless ambiguous. Tandem integrates with Doctolib and with more than a hundred other healthcare software systems. The software platform that is currently necessary for Tandem’s deployment may itself become a functional competitor.

In the emerging clinical stack, partners and competitors can be the same companies.

Vocca, Anteos and MedPair: every layer of the workflow is becoming a market

Around these broader platforms, another dynamic is emerging in France: individual stages of the consultation are becoming standalone software categories.

Vocca is targeting reception and medical secretarial work with voice agents able to handle calls, appointment booking and administrative requests. The Fondation Adolphe de Rothschild Hospital has deployed its technology to automate part of its telephone interaction with patients.

Anteos Health operates further upstream. Its assistant can conduct a voice-based pre-consultation, produce a structured summary and then act as a scribe during the appointment.

MedPair also focuses on preparation and follow-up, collecting information from patients before and after their consultation.

Taken separately, these products may appear to address different categories. Together, they describe the same workflow:

prepare → collect → receive → document → code → direct → follow up.

This is the chain that Tandem, Nabla and Doctolib are progressively trying to aggregate.

The strategic question is therefore changing. Will healthcare organisations buy several specialised agents, each optimised for a single task, or prefer a platform capable of coordinating the entire process?

Tandem’s €86 million round is also a bet on the second scenario.

From Copenhagen to London, other European models are emerging

Europe’s competitive landscape is not simply an extension of the Tandem-Nabla-Doctolib triangle.

In Copenhagen, Corti is pursuing a more infrastructure-led strategy. Rather than focusing only on an application for clinicians, the Danish company offers an Agentic Framework designed for healthcare companies that want to build their own AI agents. When it launched the framework in February, Corti said more than 100 companies were already using its infrastructure across more than 50 use cases, ranging from documentation and coding to care coordination and clinical decision support.

French company Posos, for example, has integrated its certified medication intelligence into Corti’s infrastructure, allowing agents built on the platform to access specialised information on contraindications, drug interactions and prescribing risks.

The model differs from Tandem’s. Tandem wants to become the platform used by clinical teams; Corti wants to provide part of the infrastructure on which other platforms build their agents.

In London, TORTUS represents a third strategy, one much more tightly integrated with the British NHS. The company says it has now been used across more than 2.5 million consultations, with a presence in more than 1,000 NHS organisations when practices reached through its partnership with X-On are included.

In June 2026, TORTUS also secured Class IIa UKCA certification, explicitly positioning its product as clinical decision-support software rather than simply a transcription tool.

Three models are therefore beginning to emerge in Europe: own the interface, as Tandem aims to do; provide the infrastructure for agents, as Corti does; or embed deeply inside one national health system, as TORTUS has done with the NHS.

It is still too early to know which of these strategies produces the strongest economies of scale.

When AI stops documenting and starts influencing clinical work

The most important dividing line may not be between these commercial strategies at all. It may lie between what AI documents and what it begins to interpret.

Producing a consultation summary does not raise the same issues as suggesting a medical code, identifying a risk signal or answering a question that may influence patient management.

Tandem has deliberately moved towards that regulatory frontier. Its Coding Assistant, AI Scribe and Clinical Decision Support module have each secured Class IIa certification under the European Medical Device Regulation. Its decision-support module can answer contextual questions during a consultation using sources including national guidelines, NICE recommendations or protocols uploaded by a healthcare provider.

In France, e-sensia illustrates another version of the same shift. Its technology analyses medical calls to provide clinicians with information that may help identify critical situations, while keeping the final decision in the hands of the healthcare professional.

The closer assistants move towards clinical decision-making, the greater their economic potential. But requirements increase at the same time: clinical validation, traceability, certification, model governance, human supervision and liability.

Europe adds another complication. Tandem is building under EU MDR, while TORTUS operates under UKCA.

European expansion is therefore not only multilingual. It is also increasingly multi-regulatory.

Europe’s moat may be integration, compliance and distribution

Under these conditions, the quality of the AI model itself may not become the main competitive advantage.

Foundation models continue to improve rapidly. Speech recognition, summarisation, information extraction and document generation are becoming accessible to a growing number of companies. Parts of that technological layer are therefore likely to commoditise.

Differentiation is shifting towards what remains difficult to reproduce.

First, integrations. Tandem says it supports more than one hundred EHRs and clinical software systems. Behind each integration are different APIs, coding systems, authentication methods, data structures and workflows specific to each country or vendor.

Then comes compliance. In healthcare, the product does not simply need to work. Its supplier must convince medical leadership, IT teams, security officers, lawyers, regulators, procurement departments and clinicians.

Next is distribution. Once an assistant is deployed across thousands of professionals and embedded in the daily processes of a healthcare organisation, replacing it becomes more expensive. The switching cost is no longer only technical; it becomes organisational.

Finally, there is the workflow itself. The more a platform understands how an organisation operates, the more it can move from automating one task to automating an entire sequence.

From this perspective, the AI model could gradually become the most replaceable component of the stack. Integrations, certifications, distribution and accumulated clinical trust may prove far more defensible.

Tandem’s acquisition of Juvoly in the Netherlands can also be read through this lens. The company did not merely buy technology. It acquired local distribution, integrations, knowledge of the Dutch healthcare system and an installed user base.

If the market consolidates, similar transactions are likely to follow.

The EHDS could gradually change the equation

Europe’s fragmentation will not necessarily remain unchanged.

The European Health Data Space, or EHDS, entered into force in March 2025. Its objectives include improving the movement of health data across member states and creating a more harmonised market for electronic health-record systems. The first major obligations covering patient summaries, e-prescriptions and e-dispensations are due to apply from March 2029, followed by a second phase covering areas including medical imaging, laboratory results and hospital discharge reports from 2031.

For companies such as Tandem, the effect could work in both directions.

Greater interoperability could reduce some integration costs and make it easier for products to move across national healthcare systems. At the same time, it will increase requirements around standardisation, data governance and compliance.

Over time, the EHDS may therefore reduce some of the fragmentation that currently protects national incumbents while favouring platforms already capable of operating according to European standards.

Tandem’s ability to work today across heterogeneous environments could become an advantage as those environments gradually begin to converge.

Can the Scaleup Europe Fund finance a continental champion?

The source of Tandem’s €86 million also gives the deal an industrial and political dimension.

The Scaleup Europe Fund is targeting around €5 billion in capital. The European Commission has committed €1 billion, alongside European institutional and financial investors, while EQT was selected as the independent investment manager. Its mandate covers digital systems, industrial technologies and life sciences.

The vehicle directly addresses one of the structural weaknesses repeatedly identified in European technology: the difficulty of financing large late-stage rounds without forcing the continent’s most promising companies to rely heavily on US or Asian capital.

Its first investment this summer was in Finnish radar-satellite company ICEYE, as part of a €1 billion financing round.

Weeks later, the fund also co-led, alongside Samsung Electronics and PSG Equity, Mistral AI’s €3 billion Series D.

With Tandem, it is now entering AI-powered healthcare for the first time.

The choice is significant. The fund is not financing a technology that is inherently European. Voice recognition and generative models could be developed anywhere. It is financing the possibility that a European company can aggregate a fragmented continental healthcare market before US companies or incumbent software vendors capture the position.

Tandem’s Series B therefore becomes a particularly useful test of Europe’s new scale-up doctrine: deploy larger pools of capital into companies that have already demonstrated initial traction, so they can finance the expensive phase in which a startup attempts to become a continental platform.

The next battle is for the doctor’s interface

The first generation of healthcare software structured records and data. The next digitised appointments, patient communications, billing and telemedicine.

AI agents could change that hierarchy once again.

A clinician may eventually interact less directly with five or ten different applications and more with a conversational layer able to prepare an appointment, retrieve information, listen to the consultation, document it, suggest a code, answer a clinical question, update existing systems and trigger subsequent actions.

That interface does not necessarily replace the electronic health record or hospital software. It may become the layer through which they are used.

This is the position Tandem Health, Nabla and Doctolib are pursuing through different strategies. Corti is betting more heavily on the infrastructure used to build agents. TORTUS shows that an actor can also establish a strong position by embedding deeply inside a single national health system. Vocca, Anteos, MedPair and e-sensia demonstrate that each intermediate component can still become a standalone market.

The competition is therefore no longer really about who builds the best medical scribe.

It is about two questions: how many functions can one company aggregate, and how successfully can it reproduce that integration from one European healthcare system to another?

For Tandem, Europe’s fragmentation is currently its biggest obstacle. It could also become its strongest asset. If the company learns to industrialise that complexity faster than its competitors, every additional country will not simply represent another market. It will also add another barrier to entry for those trying to follow.

With another €86 million to deploy, that is now the thesis Tandem Health has to prove.