Chris Dailey

The model is the easy part. The harness, the APIs, the connectors, the write-back logic, that is where it breaks. Rami Debs and AJ Wijesinghe actually counted: 30 APIs for one workflow. Worth the read.

89

3 Comments

This morning, Ben Torben-Nielsen, PhD, MBA at Roche pointed out a massive structural shift happening in enterprise tech: we are moving past the era of adding "AI wrappers" to existing tools, and moving into an era where AI is the center of gravity. He is exactly right. And in highly sensitive, fragmented functions like global market access, this distinction is critical. You cannot apply enterprise AI to a fragmented, offline process. Nor should you risk your data security by interacting with a dozen different embedded AI agents scattered across your tech stack. During our last funding cycle, I was repeatedly pressured to build a fragile "ChatGPT wrapper" into CCX. We refused. Instead, we built the secure, centralized Strategic Planning Layer that captures the exact data our clients' internal AI initiatives now need to wield. As the shift Ben describes takes hold, that architecture is now acting as a massive force multiplier. I just published a briefing on why the "AI Company" is a structural failure, and why true enterprise infrastructure, paired with AI infrastructure, will win the decade Foundation, then acceleration. #MarketAccess #EnterpriseArchitecture #AIStrategy

19

2 Comments

Everyone is misreading the Boltz–Pfizer deal. This isn’t drug discovery. It’s infrastructure. Boltz is running the Red Hat playbook—selling the operating system for the biotech century. In my latest blog post, I explain why this is pharma’s Linux Moment—and why it signals the end of “model tourism” and the start of standardization. (Link to blog in first comment)👇 #TechBio #DrugDiscovery #AI #Pharma #OpenSource

243

11 Comments

The compliance and auditability framing here resonates strongly. Prior authorisation is essentially a regulated workflow with deterministic rules, documentation requirements, and audit trails — and that description will sound very familiar to anyone working in GxP environments in pharma and biotech. What's striking is how closely the validation and governance requirements mirror what we see in computer software assurance for clinical and safety systems. Traceability, change control, audit-ready infrastructure, and integration with authoritative data sources — these aren't healthcare-specific challenges, they're the baseline expectations for any AI/ML system operating in a regulated context. The FHIR interoperability point is the equivalent of what we'd call data integrity and system integration validation in a GxP setting. The hard problems are never the model itself — they're the controlled data pipelines, the documented decision logic, and the evidence that the system performs as intended within its defined context of use. Claude for Medical operating at this layer suggests Anthropic understands something important: in regulated industries, adoption doesn't follow capability. It follows compliance confidence. And that's exactly where robust validation strategy becomes the differentiator.

9

2 Comments

Samsung just dropped $280 million on a single facility in Maryland. Most companies would call that expensive. Samsung calls it strategic. Here's why this acquisition of GSK's Human Genome Sciences facility is a masterclass in market expansion: Samsung Biologics didn't just buy a building - they bought market access, regulatory compliance, and 500 experienced employees who already know how to navigate US manufacturing requirements. The real genius? They're not starting from scratch. Instead of spending years getting FDA approvals and building relationships, they acquired: → Two fully operational cGMP plants → Established regulatory pathways → A trained workforce with deep US market knowledge → Immediate manufacturing capacity on American soil This move solves three critical challenges at once: 1) Supply chain resilience. Having US-based manufacturing eliminates international shipping risks for life-saving therapeutics. 2) Market credibility. American healthcare systems trust domestically produced biologics more than imports. 3) Regulatory advantage. Operating within US borders means faster response times to FDA requirements and market changes. The lesson for any business looking to expand internationally? Sometimes the fastest path to a new market isn't building from the ground up - it's acquiring the right foundation and scaling from there. Samsung didn't just enter the US market. They positioned themselves to dominate it. What's the boldest market entry strategy you've seen a company execute?

3