Helsinki‑based Gosta Labs is positioning itself as a “trusted AI operating system” for healthcare, automating documentation, structuring clinical data in real time, and supporting complex decision‑making, initially in oncology, with broader clinical ambitions. Founded in 2023 by Lauri Sippola and Henri Viertolahti (the entrepreneurs behind Kaiku Health, acquired by Elekta in 2020), the company closed an oversubscribed €7.5m seed round in November 2025 led by Voima Ventures with participation from COR Group, the Aho family, Reaktor, and angel investors, bringing total funding to nearly €10m.
The product has been used across “tens of thousands” of appointments (company claim) and is already deployed with providers in Finland, Switzerland, the Baltics and Australia, where early real‑world data (presented at ESMO 2025) suggest notable reductions in oncology documentation time and automated classification of key parameters such as Performance Status and CTCAE toxicities.
What Gosta does
Gosta describes its platform as an AI operating system rather than yet another point solution. In practice, that means three core capabilities embedded into clinical workflow: (1) automated note generation from clinician–patient encounters; (2) real‑time structuring and summarisation of longitudinal patient data; and (3) contextual guidance that links decisions to international guidelines, initially in oncology.
Where many documentation tools stop at drafting text, Gosta emphasises structured outputs that map to the data elements oncology teams actually need for tumour board decisions and pathway tracking (e.g., ECOG/Performance Status, CTCAE grading), thus making the resulting notes computable and reusable. Early results showcased at ESMO 2025 reported sub‑two‑minute median time for follow‑up documentation while maintaining structured, standards‑aligned outputs, an efficiency claim that, if sustained at scale, is materially above the typical gains reported by generalist medical scribes and dictation tools.
Technically, the company says it trains proprietary task‑specific language models for oncology and has collaborated with Finland’s LUMI supercomputer since early 2024, signals that matter when hospitals probe model provenance, data governance, and Europe‑based infrastructure.
Founders
Gosta’s founders previously built Kaiku Health, a digital symptom tracking and patient‑reported outcome platform that achieved broad European oncology adoption and was acquired by Elekta (May 2020). This is not a trivial credential: hospital procurement teams often prefer founders with regulated digital‑health execution experience. Backgrounders from Finnish AI Region and other trade outlets underscore the same storyline: repeat founders, oncology‑first domain expertise, and an intent to build medical‑device‑grade software again.
The executive bench now includes Dr Lionel Hadjadjeba (CMO) and Reetta Arokoski (COO), both cited in multiple funding write‑ups as experienced oncology/medtech operators, useful for clinical credibility during evaluations and pilots.
Oncology is drowning in complexity
Oncology’s operational complexity is rising: more patients, more tumour types, and a ballooning personalised therapy set. Multiple reports around Gosta’s funding reference an IARC projection of 35 million annual cancer cases by 2050, a figure widely used to frame the need for infrastructure‑grade tools capable of turning unstructured clinic talk into structured, guideline‑linked data.
Against that backdrop, Gosta’s pitch is that clinicians don’t need another widget, they need an operating layer that pulls together documentation, data structuring, and pathway adherence, bridging EHR silos and AI point tools. Industry commentary has started to frame this as an infrastructure‑first pivot in European health AI, not unlike how US systems are moving from single‑use scribes to workflow platforms.
From documentation to computable oncology data
Where it sits in workflow. The system “listens in the background,” drafts consultation notes, and standardises key oncology elements into structured fields. Gosta’s site and funding announcements emphasise immediate visualisation of the patient journey, suggesting a UI designed for tumour‑board preparation and longitudinal decision review.
What’s clinically material. Two aspects stand out:
- Standards‑aligned structuring: Automatically inferring Performance Status and CTCAE toxicity grades, which are essential for trial eligibility, toxicity management, and pathway choices.
- Guideline linkage: Tying decisions back to international guidelines, signalling that outputs aren’t just notes but audit‑ready artefacts for MDTs and quality audits.
Real‑world signals. The company cites deployments spanning Finland, Switzerland, the Baltics and Australia; early data showcased at ESMO 2025 suggest documentation time cut by > two‑thirds, with follow‑up notes completed in under two minutes (median). As always, we’ll want peer‑reviewed publications and multi‑site replication, but these initial metrics are at least directionally strong for workflow ROI.
Funding, investors and early commercial traction
On 27 November 2025, Gosta closed a €7.5m seed led by Voima Ventures; participants include COR Group, the Aho family, Reaktor, and several angels. This follows a €1.7m pre‑seed (2024), putting total raised “close to” €10m, with several outlets and the company’s own news page reporting consistent numbers and timelines.
Investor commentary highlights a team able to “deliver value clinicians feel immediately”, a refrain that tracks with common buyer feedback on documentation‑first AI when it’s tightly coupled to structured oncology data rather than free‑text alone.
PitchBook’s high‑level profile (paywalled details) aligns with the seed‑stage status, 2023 founding date, and Espoo office location; public elements also describe the platform as a clinical documentation assistant built on LLMs.
Hiring and operating signals
Open information across company materials and secondary outlets suggests a lean team with oncology product, data science, and med‑software experience, growing around the seed close. Basic company directories (e.g., Craft) list Espoo HQ at Otakaari 5, which aligns with the Finnish startup ecosystem hubs around Otaniemi. (Directory data can lag; use it as directional.).
Medical scribing & ambient documentation (horizontal):
Abridge, Nabla Copilot, Ambience, Suki, Nuance Dragon Ambient eXperience, these tools dominate generalist documentation and ambient capture. Their strengths are breadth and EHR integrations; their weakness in oncology is often limited structuring of disease‑specific variables beyond templated fields. Gosta’s claimed differentiation is oncology‑specific structuring and guideline linkage rather than just note drafting. (Vendor‑agnostic landscape perspective; Gosta’s oncology‑first claims are supported by its own materials and multiple funding reports.).
Oncology‑native AI & data infrastructure:
Corti (DK) is renowned for real‑time audio AI (not oncology‑specific), useful as a Nordic comparator on safety and deployment pace; PathAI and Tempus/Flatiron sit closer to pathology/genomics/EHR data layers, while Caisis/mCODE ecosystems try to standardise structured oncology data. Gosta appears to aim at the clinical front‑line OS layer that sits above pathology/genomics feeds but inside the consultation workflow, converting encounter talk into computable, pathway‑linked records. (This positioning reflects Gosta’s own public claims and investor framing.).
Bottom line on differentiation: If Gosta can repeatedly translate conversations into structured oncology variables and keep those aligned with guidelines across tumour types, it earns a defensible wedge: workflow + data quality. If not, it risks collapsing back into the broader scribe category where incumbents have scale and integrations.
Patents and IP posture
A broad sweep of Google Patents, WIPO PATENTSCOPE, EPO/espacenet, and USPTO search portals did not surface any publicly visible, explicit Gosta Labs patents at the time of writing. This is not unusual for early‑stage workflow AI companies that prioritise trade secrets, model weights, datasets, and SaaS moat over filings in year one–two; any future medical‑device modules or model‑training innovations could still move towards protection. (Readers should note that patent databases update continuously; absence of evidence is not evidence of absence.)
Where it’s being used (and what’s next)
Gosta cites usage with public and private providers across Finland, Switzerland, the Baltics and Australia. Several reports attribute early real‑world results to ESMO 2025, including automated classification of Performance Status and CTCAE and the sub‑two‑minute follow‑up documentation median. These are precisely the kinds of metrics procurement teams ask for in pilot‑to‑scale decisions; time saved, data quality uplift, and adherence to internal documentation standards.
Why it matters for the Nordics
The Nordic region has an unusually co‑operative healthcare and data ethos, but also strict expectations around safety, privacy, and public value. A European‑built AI stack trained for real‑world clinical use and operated on European infrastructure is politically and operationally resonant. If Gosta continues to demonstrate tangible admin‑time reductions while upgrading data quality for oncology MDTs, it supports both access (clinicians spend more time with patients) and research‑readiness (structured data across sites). That’s a Nordic sweet spot.
What CIOs/CMIOs and oncology heads will ask
- Can you prove durable time savings beyond pilots?
Expect demand for multi‑month, multi‑site time‑and‑motion studies and error‑rate audits. Gosta’s ESMO data is a start; scaling proofs across tumour types and languages will be key. - How do you operationalise guideline linkage?
Buyers will want to know which guidelines are mapped, how updates propagate, and whether the logic is transparent to clinicians. Gosta’s public claims emphasise international guidelines; documenting the update process will matter in procurement. - Model provenance and data protection (EU focus)
European hospitals will probe data flows, hosting, and model training sources. Collaboration with LUMI and European infrastructure are helpful, but buyers will want DPIAs, logging, and retention policies. - EHR integration burden
Even the best AI fails if the integration tax is high. While Gosta positions as plug‑and‑play, CIOs/CMIOs will ask about FHIR compatibility, identity, consent, and how the structured outputs land in the EHR without breaking clinician workflows. - Clinical governance and change management
Oncology heads will look for MDT champions, training, and audit trails. The structured note and guideline‑linking claims are promising if coupled with transparent oversight.
Risks and open questions
- Evidence depth. Early results are conference‑presented; peer‑reviewed, multi‑centre studies would bolster credibility.
- Scope creep vs focus. Expanding beyond oncology too fast could dilute the domain‑specific advantage that differentiates Gosta from horizontal scribes.
- Competition. Incumbents in ambient AI are integrating structured fields and oncology templates at speed; Gosta must keep the structure‑first edge.
Editorial verdict
Gosta Labs is a credible Nordic bet on turning clinical conversation into computable oncology data, with a founder team that knows the terrain and early‑stage results that, if validated, translate directly into saved clinician time and higher‑fidelity records. The company’s choice to invest in structured outputs and guideline linkage and to build on a European infrastructure narrative fits Nordic buyers’ sensibilities. The next 12–18 months should be about replication (sites, languages, tumour types), publishing, and making integrations boring.
If Gosta sustains measurable gains at scale, it earns the right to call itself an operating system rather than another app. If not, the gravity of the ambient documentation incumbents will pull it back toward commodity territory. For now, the signals are decidedly promising.
Sources (APA )
- Gosta Labs. (2025, November 27). Gosta Labs completes €7.5M oversubscribed seed round to scale its AI Operating System for oncology professionals worldwide. https://www.gostalabs.com/news/gosta-labs-completes-eur-7-5-million-seed-round-to-scale-its-ai-operating-system-for-oncology-professionals-worldwide[gostalabs.com]
- Gosta Labs. (n.d.). Company website. https://www.gostalabs.com/ [gostalabs.com]
- Chesnokova, S. (2025, November 27). Gosta Labs lands €7.5M to bring AI precision to cancer care. Tech Funding News. https://techfundingnews.com/gosta-labs-raises-7-5m-to-scale-ai-oncology-precision/[techfundingnews.com]
- Tech.eu. (2025, November 27). Gosta Labs completes €7.5M seed to scale its AI OS for oncology worldwide. https://tech.eu/2025/11/27/gosta-labs-completes-eur75m-seed-to-scale-its-ai-os-for-oncology-worldwide/[tech.eu]
- Finnish AI Region (FAIR EDIH). (2025, December 1). Finnish Healthtech Startup Gösta Labs Secures €7.5 Million to Transform Cancer Care with AI. https://www.fairedih.fi/en/2025/12/01/finnish-healthtech-startup-gosta-labs-secures-e7-5-million-to-transform-cancer-care-with-ai/ [fairedih.fi]
- EU‑Startups. (2025, November 27). As cancer cases surge toward 35 million by 2050, Finland’s Gosta Labs raises €7.5 million to ease oncology workloads. https://www.eu-startups.com/2025/11/as-cancer-cases-surge-toward-35-million-by-2050-finlands-gosta-labs-raises-e7-5-million-to-ease-oncology-workloads/ [eu-startups.com]
- Voima Ventures. (2025, November 27). Voima Ventures leads €7.5M Seed Round in Gosta Labs. https://voimaventures.com/voima-ventures-leads-e7-5m-seed-round-in-gosta-labs-to-scale-its-ai-operating-system-for-oncology-professionals-worldwide/ [voimaventures.com]
- Craft. (n.d.). Gosta Labs company profile. https://craft.co/gosta-labs [craft.co]
- PitchBook. (2026). Gosta Labs company profile. https://pitchbook.com/profiles/company/628311-25 [pitchbook.com]
- WebProNews / HealthRevolution. (2025, November 27). The Operating System for Oncology: Inside Gosta Labs’ €7.5 Million Bid to Rewire Cancer Care. https://www.webpronews.com/the-operating-system-for-oncology-inside-gosta-labs-e7-5-million-bid-to-rewire-cancer-care/ [webpronews.com]
- Google Patents. (n.d.). Search and read the full text of patents. https://patents.google.com/ [patents.google.com]
- WIPO. (n.d.). PATENTSCOPE: Search International and National Patent Collections. https://patentscope.wipo.int/search/ [patentscope.wipo.int]
- European Patent Office. (n.d.). Searching for patents. https://www.epo.org/en/searching-for-patents [epo.org]
- USPTO. (n.d.). Search for patents. https://www.uspto.gov/patents/search [uspto.gov]
