It’s like one of our 350 internal briefs that could only have been written if you suffered through and lived for ten years decade. This is one of the most comprehensive assessments I’ve read. You hit all the primary challenges with precision. Adherence alone is 50% of worldwide treatment failures. Of 8.3B it’s 1 in 3 with multiple morbidities. Poly pharmacy is a huge issue and with gradually declining liver and kidney function while aging, response to treatment changes over time. Early identification requires more than an app or point solution (device). And the AI meeting reality is a big one. Longitudinal data is important, but which data? Everyone has a data point or two they claim is the gold standard, but for long term chronic disease management and early ID, the correct layers and insights matter. There’s another big challenge beyond capturing utilization and engagement that doesn’t drop off. You need to be able to drive sustained behavioral change. That is far more difficult and requires a tall stack of techniques, skills, capabilities, and coordination that challenges the economics as much as the efficacy. These are the most difficult metrics. So is time. Years to build POC. Years to refine and optimize before scale. Years to validate efficacy. Not a 2 year old dated IRB study. It needs to be “here are thousands of patients under daily support, let’s see the data and outcomes in real time. We just caught a medication contradiction. This patient at this moment is having suicidal ideation right now.” Once validated at adequate enough scale, various payer approvals may take another 1-3 years. Your VC/PE firms start burning out at 4 years. big hyper scalers lose focus and remain with essential business models. Finally, head count is a sort of it. This cannot be a pure SaaS model. I used to run into the guys at Sequoia and Summit Partners. They told me if you’re adding headcount in your growth model you’ll never get funded and the business model will never succeed. People expect SaaS level margins. I told him once optimization was finished we’d achieve it. They said I was crazy. About three months ago, Sequoia published their new investment thesis and suddenly shifted their tune. They said Services built over the AI as the proper model. For every one dollar you receive in Software you receive six dollars in Services. We learn over time. We ultimately had to start building many years ago. We didn’t have time to wait for venture capitalist to figure out what we believed we already knew. So $165 million later here we are. We own it. We’re shipping. But it was so damn complicated I would never go back and do it all again. :)