Systems of Record Drive Vertical AI
Levi Lian, CEO of Raycaster, on why vertical AI is workflows first & chat last
Owning the files that regulators inspect turns software into core operating infrastructure. Veeva became hard to replace because Vault sits on the documents that determine whether a trial can run, a plant can release product, or a submission can get approved. That is very different from a sales system, where a buyer can tolerate downtime or swap vendors more easily. Once CRM, quality, regulatory, and content workflows live together, the vendor controls both the record and the work around it.
-
Veeva expanded from a life sciences CRM into Vault products for submissions and regulated content, then began moving CRM onto Vault. That shift matters because it pulls customer data, documents, and approval workflows into one stack, which raises switching costs and makes adjacent automation easier.
-
In biopharma, the important work is often not the customer list, it is the controlled document set. eTMF, quality records, RIM, and submission files need version history, permissions, review chains, and audit trails, because those are the artifacts inspectors and reviewers actually examine.
-
IQVIA shows the parallel path from the services side. By combining IMS Health data with Quintiles CRO operations, it built a workflow footprint from trial design through commercialization. Like Veeva, that breadth makes the product sticky because software and services are tied to the same underlying process.
The next step is consolidation around platforms that hold both the source documents and the actions taken on them. That is why Veeva is pulling CRM into Vault, and why newer AI layers are being built on top of Veeva and IQVIA rather than trying to replace them first. The winning products will sit closest to the regulated record, then automate the drafting, review, and handoff work around it.