sci2sci raises €1.2M pre-seed to make AI outputs trustworthy in regulated industries

sci2sci (Berlin) has raised €1.2 million in a pre-seed round co-led by Heliad and IBB Ventures, with participation from Robin Capital and Superangels. The company is building AI data verification infrastructure for regulated industries — tools that make the outputs of AI systems auditable, traceable, and defensible under regulatory scrutiny. Its initial deployment target is biopharma, where data integrity requirements are among the most stringent in any commercial sector.

The two products sci2sci has developed are VectorCat, a data cataloguing and lineage tool designed to track AI-generated outputs back to their source data and computation steps, and Integrity, which verifies that AI model outputs meet specified quality thresholds and remain within acceptable ranges across inference runs. Together, they address the core institutional anxiety about AI in regulated contexts: that outputs cannot be independently verified and reproduced, making them unsuitable for submission to regulators or use in clinical and safety-critical decisions.

The regulated AI gap

The AI governance problem in regulated industries is structural, not incidental. In pharmaceutical development, data supporting a regulatory filing must be verifiable — a clinical analysis generated by an AI model must trace back to the underlying dataset with full documentation of every transformation step, or it cannot be included in a submission. This requirement has not changed because AI tools have become available; what has changed is the scale at which AI is now generating outputs that researchers want to use in exactly those contexts.

Most existing compliance tools were designed for structured relational databases and human-authored analyses. They handle the provenance of a table or a report, not the provenance of an embedding or a vector inference chain. sci2sci's thesis is that this structural mismatch creates an infrastructure gap that the major compliance software vendors will be slow to fill, because it requires AI-specific data architecture expertise they do not yet have internally.

Biopharma is the initial vertical because the regulatory stakes are highest and the willingness to pay for compliance infrastructure is strongest. EU AI Act requirements for high-risk AI systems in healthcare add a regulatory deadline dimension: companies using AI in drug development will need demonstrable governance frameworks in place regardless of whether they would otherwise prioritise them.

IBB Ventures and Heliad

IBB Ventures is the venture capital arm of the Berlin Investment Bank — it has historically backed Berlin-based B2B software with regulatory or industrial use cases, a profile that fits sci2sci precisely. Heliad, a Frankfurt-based listed investment company with a direct and fund-of-funds portfolio, provides broader pan-European VC exposure. The combination is typical of German-ecosystem seed rounds: a regional institutional anchor combined with a wider financial-market backer.

The capital goes toward engineering headcount (building out the two products) and accelerating early deployments with biopharma clients. A €1.2 million pre-seed is small capital for a technically complex problem in a compliance-driven vertical where client procurement cycles are long. The company's 18-month window is to reach a deployment depth in biopharma that makes its data infrastructure hard to rip out before the larger vendors close the gap.