Podcast #1 - Modern MLR
- Jonathan Olsen
- Aug 7
- 2 min read
A four to six month launch window. A four to six week MLR cycle. An average of two and a half rounds of rework.
That math does not close, and most organizations are still running review models built twenty years ago.
Lou Pillai reframes it: MLR was never about citations. It is a governance model for your risk posture.
The leverage is upstream in modular content, not in speeding up the review meeting.
For a small team commercializing for the first time, that arithmetic is fatal. The launch slips, and it slips for reasons that have nothing to do with the science.
On the episode, Lou Pillai made the point that reframes the whole problem. MLR is not really about citations and technical content. It is a governance model for an organization's risk posture. Which is exactly why bolting AI onto the review meeting misses.
The industry's instinct has been point automation. Cut the review time. Lou called that noble, and then said the harder thing. By the time content reaches review, it is already too late to make changes cheaply, so you go back and redo. The leverage is upstream, in content strategy. Modularize the content. Break claims into components you can recombine into whatever you need, in whatever market, in whatever language. Very few people think that far upstream, and that is where AI actually has power.
Vendors are moving here too. Veeva's Falcon MLR is claiming it can eliminate 70% of manual MLR labor within five years. That is a vendor projection, not a result, and worth treating as such. But the direction is right. Let AI clear the unambiguous work so the cross-functional room spends its time on the genuinely contested calls, where a slightly progressive read of the label needs a human to weigh the consequence. That is risk management, and it is uniquely human. It should be what your reviewers are doing at the top of their license.
Which raises the question most teams have not answered yet. If AI drafts the content and does the first pass on review, who signs off? What does the operating model look like, and where does accountability actually sit?
That is a change management problem wearing a technology costume. And Right First Time, a manufacturing principle, becomes a launch requirement when you have one shot at the window.
Listen to this part of the conversation here: https://vimeo.com/1210362697?share=copy&fl=sv&fe=ci#t=1583.
Comments