loading...
Tandem's AI scribe reaches 98% of UK GPs and 10,000 European care orgs – now raising $100M to become a full clinic operating system.
ENTRY ANGLES
Apply ambient AI documentation to legal consultations and other regulated professions · Build inspection report generation for AEC (architecture, engineering, construction) · Create audit-trail documentation tooling for financial advisors and compliance officers
VERTICALS
CAPABILITIES
Ambient audio transcription and summarization, Domain-specific clinical or professional terminology handling, Deep EMR or case management system integration, Regulatory certification (EU MDR, FDA, or vertical equivalent)
Clinical documentation was supposed to improve when hospitals moved from paper to electronic records. Instead, the transition created a new category of work: structured data entry that happens after the patient leaves the room, consumes the equivalent of a third of a GP's working time, and contributes nothing to care quality. The European GP shortage is, in part, a paperwork problem.
Tandem Health was founded in 2023 by Lukas Saari, Oliver Åstrand, and Oscar Boldt-Christmas – a team combining McKinsey healthcare analytics, Google X and Uber AI engineering, and twenty years as a senior McKinsey European healthcare partner. The starting product was ambient transcription: the system listens during a consultation, generates a draft clinical note, and pushes it directly to the patient's EMR. The clinician reviews and approves; the documentation burden drops by 29%, with 30% improvements in job satisfaction and administrative stress measured in a July 2026 study published in JMIR Medical Informatics.
Tandem now operates across 10,000 care organizations in 14 European markets, with NHS trusts, Ramsay Santé, and Humanitas among its named customers. Its partnership with Accurx – covering 98% of UK GPs – is one of the largest AI scribe deployments in healthcare worldwide. A joint programme across University Hospitals of Leicester and Northamptonshire adds 10,000 clinicians and approximately 2.5 million outpatient appointments annually. The platform connects to 130-plus medical record systems across Europe, and its scribe, coding assistant, and decision support modules each hold Class IIa medical device certification under the EU Medical Device Regulation. The $100 million Series B, led by EQT's Scaleup Europe Fund and closed in September 2026, brings total funding to $160 million.
The scribe is the wedge. Ambient transcription solves a visible, painful problem with minimal behavioral change – the consultation continues normally, the AI observes – and that low-friction entry is what allows Tandem to reach 10,000 care organizations without a lengthy procurement cycle.
But documentation is only the most visible fraction of clinical administrative work. Triage, appointment scheduling, referral management, and care coordination all run on the same EMR infrastructure, all require the same clinical coding logic, and all represent equivalent time burdens on clinicians who are already operating near capacity. Tandem's architecture – 130 EMR integrations already in place, three separately certified AI modules, presence in 14 markets – is the foundation for extending into each of those workflows without requiring new regulatory clearances or integration work from scratch.
The JMIR publication matters beyond its marketing value. In European healthcare procurement, where NHS trusts and health ministries evaluate software purchases through formal clinical evidence assessments, a peer-reviewed outcome study is a competitive asset that a white paper cannot replicate. It also creates a reimbursement pathway: documented clinical impact is what allows AI tools to move from pilot budgets to line items in national health service contracts.
Nuance (Microsoft), Suki, and Abridge dominate the US clinical documentation market. Tandem's European position – 130 EMR integrations across 14 markets, regulatory certifications under EU MDR rather than FDA, NHS relationships at national scale – is not easily replicated by US-first competitors expanding to Europe or by EU-native startups starting from scratch.
Clinical documentation is one category in a broader class: high-value professional workflows where the documentation obligation is legally non-negotiable, the expert's time is expensive, and the gap between what the expert produces and what AI can generate has recently collapsed.
The legal profession has the closest parallel. Lawyers produce time-billed documentation – counsel letters, court filings, compliance reports – that is subject to audit, attracts malpractice risk, and currently occupies a large fraction of billable hours. Ambient transcription of legal consultations and automated draft generation for standard document types are less developed than their clinical equivalents, despite the structural similarity.
Architecture and construction have a version of the same problem in inspection documentation: site visits produce reports that are mandatory, expensive in professional time, and highly formulaic once the inspection is complete. Voice-and-vision AI models are now capable of generating those reports from a site walkthrough without manual transcription.
The entry angle with the strongest near-term characteristics is any regulated profession where documentation is mandatory, time-consuming, and subject to audit – and where a single large institutional buyer, whether an NHS trust, a law firm partnership, or a construction conglomerate, could validate a pilot within a defined scope before a full rollout. Tandem's path through large NHS hospital systems is the template.