5 days ago
Zifo's Global Survey Reveals Early Momentum for AI in Biopharma, But Data Readiness Remains Key Hurdle
RALEIGH, N.C., CAMBRIDGE, England and CHENNAI, India, July 24, 2025 /PRNewswire/ -- A new survey of scientists and informaticians reveals that while investment in artificial intelligence (AI) and machine learning (ML) is rapidly accelerating across the R&D, manufacturing, and clinical value chain, persistent data silos and integration headaches are stalling progress -- raising crucial questions about whether science-focused companies are truly ready to harness the full power of AI.
Zifo Technologies Logo
The Data Readiness Survey, conducted by Zifo Technologies, polled scientists and informaticians from over 30 science-driven companies, revealing both enthusiasm for AI's transformative potential and persistent challenges in data management and integration.
While nearly two-thirds of organizations have begun investing in AI and ML across their value chains, only 32% of respondents express high confidence in their company's ability to leverage scientific data effectively for AI initiatives.
A striking 70% report that accessing the data needed for AI projects is either difficult or somewhat difficult, underscoring a widespread struggle with data accessibility. Data is not harmonized in terms of storage or metadata, the report noted, highlighting the lack of standardized practices that hampers progress.
Data silos, interoperability, and automation gaps emerge as major obstacles. Nearly half of organizations find it "somewhat difficult" or worse to pipeline and integrate data from lab instruments, with aging infrastructure and a lack of common standards complicating seamless data exchange. Automation of data capture is growing, but 26% still rely primarily on manual processes, and 10% use no automation at all. Adoption of standardized data formats and ontologies is mixed, with 39% agreeing their organization has them, and an equal proportion either unsure or disagreeing.
The report also noted that a critical gap exists in current data management solutions for High-Performance Computing (HPC) environments. Most systems, such as Electronic Lab Notebooks (ELNs), are not designed to handle the petabytes of unstructured data generated during complex analysis. While the initial capture of instrument data and the final storage of analyzed data products are well-automated, the crucial intermediate processing stage on HPC systems remains poorly supported.
As Zifo's Chief Scientific Officer Paul Denny-Gouldson observes, "Data management is fundamental to ensuring data reuse and data retrieval, because that is the lifeblood of what enables FAIR [Findable, Accessible, Interoperable, Reusable] data".
