From data to curated information to new product

Setting the Scene

The scientific information industry is undergoing a significant transformation. With the exponential growth of data in life sciences, materials science, and chemistry organizations face challenges in managing, integrating, and deriving meaningful insights from vast and diverse datasets. Traditional data management systems often struggle with the volume, variety, and velocity of modern scientific data, leading to inefficiencies and delayed discoveries. 

CAS, renowned for its comprehensive chemical substance database, recognized the need to evolve its product offerings and infrastructure. The new requirements were for interdisciplinary research. Pharmaceutical companies and academic researchers now want gold standard, curated, biological research as well as chemical data. 

A new data hub, new infrastructure, and a new way of working were needed to rapidly meet this demand.  

• Internal operations: modernising curation tools, some over 30 years old 

• Customer-facing platforms: enabling semantic search and data-driven insights 

• Customer AI initiatives: through co-developed, agile solutions 

 This AI-driven transformation is supported by CAS’s strategic investments across the entire value chain—from operations to product development and customer enablement.

The Challenges

From legacy to agility

Despite its industry leadership, CAS was constrained by legacy systems and fragmented workflows that were no longer fit for purpose in a data-intensive, fast-moving research environment. 

Key obstacles: 

• High data volume: Content scattered across many disconnected sources;
• Inconsistent standards: Lack of harmonization across formats and taxonomies;
• Poor compatibility: Diverse systems and tools blocked integration and insights;
• Slow delivery timelines: New content could take 18+ months to become productized;
• Rigid processes: Agile ceremonies had lost meaning, creating bottlenecks instead of speed;
• Cultural silos: Siloed roles and documentation-heavy workflows fostered a blame culture rather than collaboration.

CAS also faced a unique pressure: to compress 80 years of chemical curation experience into just 2–3 years to cover new disciplines like biology and the new discipline of biological research. 

The Solution

Technology, process, and cultural change

THE TRIANGLE OF SUCCESS

To drive innovation, CAS adopted a foundational model they call the Triangle of Success, recognizing that impactful AI transformation requires three distinct capabilities: 

• Domain Experts: Who bring scientific context and define ontologies 

• Content Experts: Who ensure proper data modelling and scientific harmonization 

• Tech/Algorithm Experts: Who build scalable, intelligent systems 

This triangulated expertise enabled effective semantic modelling, entity extraction, and relevance-driven curation. 

This transformation was not limited to technological change. As Rodney Fulford from CAS highlighted, true innovation came from aligning people, processes, and platforms, introducing a cultural shift as much as a technical one.

 
TECHNOLOGY MODERNISATION WITH MARKLOGIC AND AWS 

 “Big data makes the hard things possible, but it makes the simple things hard. MarkLogic gave us visibility and speed to do the simple things well.” 
— Rodney Fulford, CAS 

CAS selected Progress MarkLogic, hosted on AWS, to replace five siloed data stores with a single operational data hub.  

The platform supports both structured and unstructured content management, combining hierarchical XML with triple-based metadata to enable advanced search, enrichment, and content reuse. Importantly, the solution was not solely focused on semantics, it was designed to manage rich XML content in tandem with RDF triples to meet the diverse needs of CAS’s content architecture.
 
MarkLogic offered: 

• Real-time visibility into hierarchical and unstructured data 

• Support for multi-model content (XML, JSON, RDF, SQL) 

• Built-in semantic search, triple stores, and enrichment capabilities 

• Superior speed, performance, and change agility 

ROADRUNNER: 90-DAY PROOF OF CONCEPT 

To validate the new architecture, CAS launched Project Roadrunner, a 90-day POC focused on the core knowledge management pipeline: 

• Unified ingestion, curation, discovery, and enrichment 

• Replaced ETL-heavy workflows with agile, semantic pipelines 

• Enabled content visibility within 15 minutes of ingestion 

• Demonstrated "Stage Gate" checks for fast-cycle quality control 

Key milestone: reduced cycle time for content from over a year to <75 days

PHOENIX: A NEW CURATION ENGINE 

To validate the success of Roadrunner laid the foundation for Phoenix, CAS’s next-generation platform for scientific content creation. Phoenix brings flexibility, feedback-driven iteration, and test data subsetting, all powered by unified semantic models.

ODH: CREATING A GOLDEN RECORD FOR SCIENTIFIC DATA 

Building on Phoenix’s foundation for curation, CAS initiated the Operational Data Hub (ODH): a modern data management architecture that serves as a central point for accessing, integrating, and processing scientific data at scale. Designed for performance, the ODH enables fast, structured access to curated content, solving complex data integration challenges while laying the groundwork for advanced enterprise strategies like cloud scalability and future technology adoption. 

Unlike a raw data lake, the ODH produces a “golden record” through multi-stage semantic enrichment. It ingests structured data from multiple external vendors, transforms it into a unified CSA XML format, and links key biomedical entities such as targets, diseases, drugs, and interactions. Datavid supported the initiative by contributing to metadata modeling, UI development, and foundation standardization, ensuring downstream platforms can reliably consume high-quality, linked content. This trusted data layer later became essential to powering products like BioFinder.

PRIME WITH DATAVID: FROM POC TO BIOFINDER V2 LAUNCH 

To drive product innovation, CAS launched Project PRIME (Product Innovation Mechanism) with Datavid as its strategic co-development partner. The initiative began as a 90-day proof of concept (July to November), where Datavid played a central role in rethinking the BioFinder platform’s architecture.

Working closely with both product and content teams, Datavid enabled agile development workflows that accelerated delivery and reduced risk. Following the success of the PoC, CAS proceeded with full-scale implementation - culminating in the launch of BioFinder V2 earlier this year. 

The result: a modernized, scalable product delivered on time and built through close collaboration with Datavid.

Don’t just take our word for it.
WATCH THE DEMO!

The Outcomes