With €30m, BIOLEVATE wants to turn AI into a compliance tool for drugmakers

Biolevate has raised a €30 million Series A to deploy its artificial intelligence platform across regulated life sciences operations. The Paris-based company no longer presents itself merely as a medical-writing tool. It wants to shape how drugmakers search, document, verify and validate scientific knowledge. That ambition places it up against established regulatory platforms, and at the centre of the debate over sovereign AI in healthcare.

Artificial intelligence promises to shorten the time between a scientific discovery and a treatment reaching the market. In laboratories, it can already scan publications, connect datasets, prepare first drafts of documents and identify signals in regulatory intelligence. But in this industry, producing an answer is only the beginning. Its source must be established, the method understood, the underlying data checked and, above all, the decision assigned to an identifiable expert.

That less visible part of the process, yet one that is critical to clinical development and drug approval, is where Biolevate wants to position itself. The startup has raised a €30 million Series A, co-led by RAISE France and Orange Ventures, with participation from EQT Ventures, the MSD Global Health Innovation Fund and Station F. It raised €6 million in a seed round in November 2024.

The funding is intended to accelerate product development and international expansion, including the opening of a Boston office. Aymar Hénin, entrepreneur and investor through RAISE France, becomes chair of the board, while Jérôme Berger, Orange’s group strategy and venture capital director, joins it. The make-up of the round signals the company’s next phase: the task is no longer simply to finance a technology, but to embed it inside large life sciences organisations.

The bottleneck begins after discovery

AI in healthcare is usually told from the front end of the value chain: better identification of therapeutic targets, faster-designed molecules, better-selected clinical trials. This part of the chain attracts the headlines and much of the capital. Yet the friction continues long after the research lab.

A therapy must be documented across thousands of pages of publications, protocols, trial reports, responses to health authorities and submission dossiers. Regulatory teams must follow different rules across jurisdictions, maintain consistency across documents produced over years and demonstrate that every conclusion is supported. The potential time savings are considerable. So are the risks of poorly controlled automation.

Biolevate first entered the market through medical writing, the preparation of scientific and regulatory documents. It is now extending its proposition across research, clinical trials, health technology assessments, regulatory submissions and drug development. This shift matters: Biolevate is no longer seeking simply to help a writer work faster. It wants to move controlled knowledge across several functions within a drugmaker.

Agents, but above all an evidence trail

The platform claims a knowledge-based approach. It aggregates scientific, regulatory and company-specific sources, then uses them to power workflows in which results remain linked to the documents supporting them. Its agents can search, extract, compare and synthesise; experts retain control over review and validation.

The distinction is more than semantic. In a regulated environment, the challenge is not merely to reduce a model’s erroneous answers. It is to reconstruct the full path from a source to a statement, from a statement to a document, and from a document to a decision. An unsourced answer may be inconvenient in a search engine; it becomes an operational risk when it enters a dossier submitted to a health authority.

Biolevate says it can complete literature reviews up to sixteen times faster and generate an 80% efficiency gain in regulatory intelligence and compliance. It also says it can operate 100,000 agents simultaneously. These are company-reported performance indicators. Their significance will depend on the use cases concerned, the degree of automation involved and the extent of human validation that remains in the workflow.

Production deployments are the funding round’s real signal

Biolevate says it now has more than ten large enterprise customers in production, after establishing partnerships with Sanofi, NVIDIA and Microsoft. Those deployments give the round its real significance. In healthcare, a pilot is relatively easy to launch, often funded through an innovation or digital department. A production deployment entails a very different discussion: data security, supplier qualification, integration with existing systems, quality control and accountability for business teams.

The startup now employs 50 people and plans to double its headcount again. That growth also raises a familiar question for enterprise software companies selling to large drugmakers: when does deployment stop relying on a substantial mobilisation of specialists and become a product that can be replicated from one customer to the next?

Biolevate has also filed seven patent applications, two of them already lodged in Europe. The portfolio does not yet amount to an insurmountable barrier. It does, however, signal an effort to protect technical components and methods in a market where underlying models evolve quickly, while validation rules, connectors and domain knowledge take much longer to build.

Competitors already own either the records or the workforce

Biolevate is entering a market that is far from empty. In medical and regulatory writing, Yseop, Certara and Indegene already offer automation, document generation and support for writing teams. Yseop, also French, is the closest comparable on the promise of traceable AI capable of producing clinical and regulatory documents.

Further up the chain, Veeva and ArisGlobal have a structural advantage: their platforms already manage product information, document versions, approvals and exchanges with health authorities at many drugmakers. Both are now adding GenAI and agent capabilities to their own environments. Clarivate, meanwhile, owns regulatory and scientific intelligence databases used by laboratories in their decision-making.

Then there are the large service providers, from Indegene to IQVIA, which still employ the teams that carry out a substantial share of writing, pharmacovigilance and regulatory affairs work. Their incentive is straightforward: to turn expertise sold by the day into software workflows they control themselves.

Biolevate’s role, then, will not be to replace every one of these players on its own. It must show that it can become the layer connecting knowledge with operations, sitting above existing systems of record without creating yet another tool for teams to manage. That is a harder proposition to sell than a writing assistant. If it works, it is also far more defensible.

Strategic for France, on certain conditions

Biolevate’s strategic value does not stem simply from applying AI to healthcare. France already has startups working in drug discovery, medical imaging and hospital data. It lies in the part of the value chain Biolevate is targeting: organising and validating sensitive knowledge within industries that remain heavily dependent on foreign software, data and models.

Such a platform could give European drugmakers the ability to deploy agents on their own corpora, in controlled environments, with explicit policies for sources, data and the people who validate outputs. Sovereignty is therefore not limited to the choice of a language model. It also concerns the application layer where business rules, access rights, control procedures and the memory of past decisions are embedded.

That description of Biolevate as a strategic startup remains conditional. It requires the company to retain and deepen control over workflows, evaluation and integration, rather than becoming an interchangeable interface built on models developed elsewhere. It also requires decision-making, intellectual property and the capacity to serve European customers to remain anchored in France and Europe, even as the company pursues the US market.

From that standpoint, Boston is close to unavoidable. The city brings together drugmakers, biotech companies and regulatory expertise that could turn Biolevate into a global supplier. The challenge will be to expand there without allowing the company’s centre of gravity to leave Paris, a familiar risk for European startups once large-scale commercialisation begins.

An open path between independence and consolidation

This Series A does not open an immediate exit scenario. It gives Biolevate the means to attempt to establish itself as an independent platform. To do so, the company will need to turn its early deployments into long-term contracts, deepen its integrations and show that its technology can apply across therapeutic areas and regulatory regimes without multiplying bespoke projects.

If that trajectory succeeds, several families of acquirers could emerge. Regulatory software vendors may want to add a more advanced reasoning and evidence layer. Service providers and CROs may seek to industrialise part of their operations. A model or infrastructure provider could also view Biolevate as an entry point into a regulated vertical, provided it preserves the technological neutrality that drugmakers are likely to require.

That final possibility will not depend solely on model performance. It will depend on Biolevate’s ability to become an asset that is hard to rebuild: a team that understands regulatory affairs, customers that use the platform in practice and processes whose outputs are accepted by quality functions.

The real question, then, is not whether AI can write a scientific document faster. It already can. It is whether Biolevate can become the place where a drugmaker organises evidence, validation and accountability. At that point, it would stop being another productivity tool and become strategic infrastructure for healthcare.