During SapioCon 2026, speakers from across the industry returned repeatedly to a problem that is costing life sciences organizations more than they realize. Sapio Sciences research found that 65% of bench scientists repeat experiments not because the science demands it, but because they cannot find, trust, or reconstruct the context of previous work. The Pistoia Alliance offers a structural explanation. An estimated 55% of an organization’s scientific data is effectively dark, trapped in legacy systems, unstructured repositories, and the memory of people who have since moved on.

Both organizations brought survey data to back up what many in the room already suspected: that bottlenecks in modern scientific organizations are rarely a shortage of data or ambition but a challenge with the underlying digital infrastructure.

The digital backbone as a clinical necessity

Federico Lionetti leads systems development at Navignostics, a precision oncology scale-up developing a diagnostic solution that profiles more than 40 proteins at single-cell spatial resolution to identify personalized drug combinations. He set the terms plainly: “Building the right digital backbone is as important as the science itself.”

At Navignostics, the workflow is the product. The diagnostic process moves from sample receipt through histology, staining, and acquisition to automated reporting. Treatment outcomes in clinical studies were three times more effective than standard care. Lionetti described this process as a relay race where each step hands context forward to the next runner. A unified lab informatics platform acts as a digital relay coach, not just storing data but orchestrating handovers to ensure context remains intact throughout.

The design philosophy that followed from that reality shaped every technical choice the team made. As Lionetti noted, “We didn’t want to transition from paper to digital. We wanted to start digital from day one.” By building the digital backbone before the science began, Navignostics ensured they had a solid foundation on which to build their audit-ready compliant reality.

Solving the strategic problem of fragmentation

The Wellcome Sanger Institute operates at a scale that reinforces this argument from a different angle. Amy Yeung, Head of Cellular Services in Scientific Operations, described managing nearly 55 petabytes of sequencing data at an organization that sequences the equivalent of a gold-standard human genome every 12 minutes.

At this volume, the gradual accumulation of disconnected systems and local workarounds becomes more serious than mere inefficiency. Data that cannot be found or connected to the decisions it once informed stops being an asset; it becomes a structural problem.

Wellcome Sanger aims to transform operational data from a “byproduct of science” into a strategic input. By moving to a unified digital platform, the Institute is establishing the structured, trustworthy foundation required for advanced cross-program analytics. Getting this data foundation right is the essential strategic work that enables all future scientific growth.

Transitioning from passive records to active advisors

Once the digital foundation is in place, and the data infrastructure is governed and trustworthy, the informatics platform can move from a passive record to an active participant in the science. This is where integrated artificial intelligence (AI) earns its place in the workflow rather than being bolted on around it.

Rob Brown, VP and Head of the Science Office at Sapio Sciences, framed the shift directly: “An AI lab notebook should not replace the scientist. It should sit in the workflow, add context, connect the evidence and help teams make better decisions.

Brown illustrated this transition with a real-world example: a medicinal chemist querying 1,500 molecules across three assays for dopamine transporter (DAT) selectivity. Traditionally, this would require exporting data to external tools for manipulation. Within an AI-enabled informatics platform, the scientist receives a focused analytical report, candidates ranked, calculations shown, statistical context provided, all without leaving the workflow. The reasoning stays inside the record, ensuring the scientist does not have to break the workflow to do the thinking.

The foundation of smarter science

The common thread across these sessions is that smarter science is not a technology you buy; it is an operating environment you build. The requirement is the same whether you are a scale-up like Navignostics building clinical precision from the first biopsy or the Sanger Institute unifying operations across petabytes of data. The capability of your AI and analytical tools will always be capped by the integrity of your digital backbone.

A “Digital from Day One” approach is not just a philosophy about technology. It is a strategy for making science faster, more reproducible, and ultimately more impactful, and the foundation on which everything that follows is built.

Key Points

  • The cost of inaction: 65% of scientists repeat experiments due to lost context, while 55% of industry data remains dark and inaccessible.
  • The platform as a relay coach: Success requires a lab informatics platform that orchestrates context across the entire workflow lifecycle, ensuring data does not lose its meaning between teams.
  • Digital from Day One: Building a digital backbone simultaneously with scientific processes is the only way to ensure audit-ready compliance and the turnaround times required for personalized medicine.
  • Foundation over pilot: The value of scientific intelligence is proportional to the trustworthiness of the underlying data foundation; getting the data right is the work that makes operational science possible.
  • Active informatics: The shift from passive repositories to active platforms allows interpretation and decision-making to happen within the governed record, preserving critical institutional knowledge.