Government Uses Transparent AI to Build a More Complete, Repeatable View of the Life Sciences Sector
• Government
A broader, more detailed and repeatable evidence base for official statistics
The Office for Life Sciences (OLS), working with Technopolis and Glass.AI, introduced an AI-enabled methodology to identify and classify UK life sciences companies.
↗ Key Impact
~50% cost savings with increased quality and transparency of data
7,600 companies classified through AI-driven open web research and bespoke models
Established a proven and repeatable method that can be mirrored across other strategic sectors
AI-driven research expands company discovery beyond traditional classifications, established datasets and predefined lists.
Analysis of company activities improves classification against the BaHTSS taxonomy, complementing SIC codes and official data.
A repeatable, human-in-the-loop methodology enables comparison between releases and monitoring of sector change.
Background
Life sciences is strategically valuable to the UK economy and anchors the Industrial Strategy, making a comprehensive and current understanding of the sector increasingly important. Mapping UK life sciences is challenging because the sector evolves alongside established and emerging technologies. Standard Industrial Classification (SIC) codes capture only part of it and may overlook diversified businesses, smaller companies and start-ups. Existing sources can also be affected by time lags and provide limited evidence about company activities.
Previous Bioscience and Health Technology Sector Statistics (BaHTSS) collection relied heavily on manual collation, classification and validation. OLS needed a more comprehensive, repeatable approach that could broaden discovery while remaining transparent, auditable and compatible with official data and expert judgement.
Solution
OLS commissioned Glass.AI and Technopolis to develop and refine a repeatable data-collection methodology.
Glass.AI created a broad discovery universe using company websites, open-web sources, previous BaHTSS records and information from OLS data partners. Its technology identified potential companies, extracted evidence about their activities and used linguistic, semantic and statistical models to assess whether they met BaHTSS criteria.
Rather than relying solely on industry codes, the AI models examined what companies actually do. OLS and Technopolis experts reviewed samples, refined the scope and fed observations back into the models to improve accuracy and manage false positives. Applying the methodology across the 2023/24 and 2024/25 BaHTSS releases allowed the process to benefit from previous results and refined sector criteria.
Impact
The methodology gives OLS a broader and more granular view of UK life sciences while reducing reliance on manual discovery and processing. Around 50% cost savings were achieved, alongside increased quality and transparency of the resulting data, with 7,600 companies classified through AI-driven open-web research and bespoke models. This supports analysis of sector trends, clusters, geographical hotspots and market segments, strengthening evidence for policy and cross-government decision-making.
Greater visibility of companies and stakeholders also enables more targeted engagement and a better understanding of emerging challenges and opportunities.
Crucially, the approach provides a repeatable foundation for sector intelligence. Comparable releases can show not only which companies are active, but how the sector changes as technologies, markets and business activities evolve. It demonstrates how specialised AI models can extend discovery and focus expert effort while retaining the validation and official-data linkage required for trusted statistics.
Want to learn more?
Download the full case study for a closer look at how Glass.AI built a repeatable AI process for the OLS to monitor the critical life sciences sector in the UK.
Want insights no one else can see?
Discover how Glass.AI can transform your research process and unlock intelligence hidden across the web. Let's discuss your specific needs.