Omni Know-How
August 30, 2026
I've been on the front lines of healthcare data long enough to have watched the entire arc — from the early days of fragmented, low-capture data assets barely stitched together, to the sprawling, signal-rich environment we navigate today. I've seen how life sciences organizations have evolved in how they use that data to reach and teach HCPs. And I've developed a deep, genuine curiosity — professionally and personally — about where this is all heading.
AI is forcing me to ask those questions harder than ever. And I think anyone in healthcare marketing data strategy who isn't a little unsettled — and deeply energized — by what's coming isn't paying close enough attention.
The Evolution I've Witnessed
Early in my career, the conversation was almost entirely about access. Prescription data here. Claims there. A survey from six months ago. A rep call note no one fully trusted. Capture rates were low, linkage was manual, and a unified HCP view was a north star most organizations could gesture at but never reach.
Then things changed — incrementally, then all at once. Data sources multiplied. Digital engagement signals piled up. The challenge shifted from not enough data to far too much to make sense of fast enough. That shift is exactly what makes AI a structural necessity, not a novelty.
The Central Shift: AI Will Make Better Data-Driven Decisions Than We Can
Let me be direct about something that requires intellectual honesty from those of us who've built careers on analytical judgment: AI will be better than humans at making complex, multi-variable, data-driven decisions. Not eventually — soon. In some domains, arguably already.
The cognitive constraints on human decision-making are real. We hold a limited number of variables in working memory at once. We're prone to confirmation bias. We optimize for what we expect to see. AI doesn't have those constraints — it can process thousands of variables simultaneously, identify non-obvious interaction effects, and surface patterns that would never emerge from human analysis. Not because we aren't capable, but because we're not built for it at that scale.
What this means in practice for HCP marketing is significant, and it plays out in two specific ways I watch closely.
From national treatment to segmentation to 1:1 — where it makes sense. The industry has already made meaningful progress on this journey. We moved away from treating every physician identically at a national level, and segmentation was the right answer for that moment — and in many ways still is. Brands are built on segmentation. It gives field teams a workable framework, helps us understand behavior at a macro level, and remains a valuable lens for how we think about physician populations. But AI opens a new chapter beyond it. At the individual HCP level, AI can synthesize engagement signals, prescribing behavior, clinical context, and channel preferences to surface not just a profile but a recommended action — what to say, through which channel, at what moment. The question is no longer whether to segment, but knowing when a segment-level view is sufficient and when a truly individualized approach will move the needle. AI gives us that choice in a way we've never had before.
From reactive signals to a smarter combination of reactive and predictive. Data latency has been the quiet enemy of relevance for years. By the time a prescribing signal reached our analytics layer and translated into an outreach, the moment had often passed. But that's changing. As AI improves data interoperability and accelerates the pipeline from signal to action, real-time and near-real-time triggers are becoming genuinely useful — and in many situations they arrive in time to matter. Reactive triggers aren't going away; they'll continue to play an important role.
What AI adds on top of that is predictive capability beyond what we have seen in the past— and I believe we're on the verge of a step-change here. As data that once lived in isolated silos gets processed across disparate sources simultaneously, reactive and predictive models will increasingly work together. Identifying which physicians are likely to encounter an appropriate patient before that encounter occurs. Mapping HCP knowledge gaps through content engagement signals before they show up in prescribing patterns. Spotting emerging KOLs as influence is forming, not after the fact. Flagging infection rate hot spots months before they surface in public health reporting. The future isn't reactive or predictive — it's a connected intelligence layer where each makes the other smarter.
The Foundation Has to Come First
None of this trust in AI becomes warranted without doing the hard, unglamorous work first. The models are only as good as what we feed them and how we train them.
That means the near-term agenda for anyone serious about this is largely unsexy: auditing data quality, resolving HCP identity fragmentation, cleaning and harmonizing feeds never designed to work together, and building governance structures that give AI the right foundation to operate from. The organizations investing in this now will compound that advantage as capabilities continue to grow. The ones that skip it will be building on sand.
Human judgment doesn't disappear in this future — it moves up the stack. We need to ask better questions, set better objectives, build better guardrails, and develop the organizational conviction to act on AI outputs that may challenge our instincts. That last part is as much a cultural challenge as a technical one.
Where This Leaves Us
The gap between the data we have and the insight we can actually act on has always been larger than it needed to be. Too much latency. Too many silos. Too little synthesis.
AI doesn't just make us faster at closing that gap — it fundamentally expands what's possible. It enables capabilities we literally couldn't achieve before: true 1:1 personalization at scale, predictive intelligence drawn from sources that once were isolated, decisions informed by variables our brains simply cannot process simultaneously.
The organizations that lead through this well won't just have the best technology. They'll be the ones who did the foundational work to make their data trustworthy, and built the conviction to follow where it leads.