The order of operations in Frank's account matters more than any single number in it. The enterprise data platform and the Clinical Data Repository were built during the pandemic, before generative AI became an executive priority, and Frank is explicit that the predictive and proactive work would not run at all without the data organised first [10][11]. Everything downstream, including nurse summarisation and the outreach engine, draws on a bill settled years earlier. That is also why the newest-looking piece is the least durable: Aetna says it was the first payer to deploy Rich Communication Services, with engagement above 80 percent and opt-outs under half the SMS level [14][15], and a messaging channel is procurable by any competitor inside a quarter.
The usage figures deserve a harder read than they will get. Frank, speaking to Forbes, puts use below 10 percent under the old cost-estimation model and near 70 percent "among people using the capability before a procedure" [8][16]. Those are not self-evidently the same denominator, since the second is conditioned on members who already reached for the tool. That comparison could be read as flattering the rebuild, but Aetna reports running A/B tests and then tying digital interactions to care costs, prior authorisation automation and outcomes [9], which is the measurement discipline that would eventually settle the question. The version that travels in a board deck, ten to seventy, stays incomplete until the base is stated.
Where the nurse figure came from is more interesting than its size. The prior workflow had a nurse navigating multiple systems and hundreds of pages of clinical documentation before a conversation with a member facing a cancer diagnosis [12]. Ninety minutes is a bit under a fifth of an eight-hour shift [2], and it was going into retrieval. Summarisation before the call, ambient listening during it and organisation afterwards amounts to a search problem solved at the desk, in the same spot where the clinician had been doing the searching by hand [12], rather than a clinical intervention in itself.
The economics explain why this agenda should survive a cost review, which is a different question from whether it is popular. Frank says that when he took the role about four years ago, peers at other insurers were putting their biggest bets on utilisation, cost management and cloud transformation, and he chose member experience instead [6]. But the scorecard he describes asks whether a feature helped someone obtain preventative care or avoid an emergency room visit [9], which is utilisation management approached from the consumer side. The segment carries $143.4 billion of revenue against 26.6 million medical members, roughly $5,390 per member per year [3][5], and it accounts for about 36 percent of CVS Health's 2025 revenue [2][4]. When the payer holds the medical cost, removing the member's homework and lowering the claim are the same project, and the alignment is what makes the spend defensible in a lean year.
An operator with a comparable estate can copy the sequence in Frank's account long before copying any of its outcomes, because the consolidation work sets the ceiling on everything reported after it.
