Dan Wilson — Feb 13, 2026

Attendees: Thomas, Michael, Dan Duration: ~32 min

---

1. PREMERA UPDATE

Thomas shared the outcome of yesterday's Premera meeting (with their head of analytics and AI data team):

• Landing on FDE consulting work. 3-6 months of embedded "PhD work" — learning their systems, understanding their needs.
• Rate proposed: $150/hr for ~40 hrs/week. Proposal sent, expect response early next week.
• Enterprise product contracting in parallel. Premera said even if we wanted to, enterprise procurement takes 6-9 months through security, review, permissions.
• Two specific workstreams revealed:
  • Auto-auth: Accelerating automated authorizations (MCG/United have versions of this)
  • Appeals processing: Premera is ~12 months behind on the provider appeal AI wave. Getting slammed with AI-generated appeals. Need help making the process more efficient.
• Access level: Will be permissioned like employees. Significant data and system access.

Dan's reaction:

• "That's a great outcome for you guys honestly. You're going to get paid to go to school."
• Contractor access vs. vendor access is "180 degrees different" — as a vendor they'd expect you to have all the answers.
• $150/hr is fair/market. If they counter at $130, take it. Even $100 is still $200K/year per person.

---

2. TAX & CORPORATE STRUCTURE ADVICE

• As a C Corp, consulting income gets double-taxed. Consider an S Corp or LLC subsidiary for the consulting revenue.
• S Corp / LLC: pay yourself a minimal salary ($40K), dividend the rest quarterly — much better tax treatment.
• Talk to your tax person about whether this makes sense.
• "It might be complicated but it might be a situation where you want to do some kind of subsidiary that's taxed differently."

---

3. STOCHASTIC / CORAL HEALTH — APPEALS CONVERGENCE

Dan is building an appeal letter generator with Coral Health through his Stochastic work:

• Getting all payer decisions to tune the model
• This is the opposite side of what Premera wants help with (provider attacking vs. payer defending)
• "There might be a scenario where we collaborate around that"
• Don't want to waste time rebuilding if IP can be repurposed
• Some of the IP could be built outside the Premera relationship — need to figure out what makes sense

---

4. THE VISION — TRANSACTION CLEARING

Dan articulated the end-state vision clearly:

• Think of auth like financial transactions. If both sides are on the same toolset, "for the love of God, can we just clear the transaction more efficiently?"
• But: "What I learned with Moxe is you can't build the end to end and expect both sides to adopt it together."
• Each side has to be valuable independently first.
• Finance analogy: get enough buyers and sellers, then clear the transaction and take both sides of the spread.
• Advice: Keep going at payer side, forget provider for now. But make architectural decisions with the eventual bridge in mind. Build a business model around the idea that you'll eventually get there — "don't screw the other side of the transaction. That'll put you on the right side of history."

---

5. THE NEWCO CONCEPT (Evolved from Feb 11 Merger Idea)

Dan's thinking has evolved from a straight merger:

• Envisions a "NewCo" that's an FDE services business
• Partners with Stochastic for platform/tech — when DaisyAI hits a technical ceiling, Stochastic builds the abstracted capability
• DaisyAI focuses on workflow, agents, domain-specific stuff — not building pipelines for 75 document types
• Targets RCM opportunities on both sides of payer/provider
• Key tension he's wrestling with: one company vs. distinct businesses?
  • All-in-one = suboptimal cost allocation, can't amortize tech the same way
  • If Stochastic isn't the best tool for a specific job, you need flexibility to use something better
  • "Intertwined economics but independent operations"
• Not finalized with Glenn at Stochastic — "this is the head of Dan"
• Needs more time to "turn more cards" — what Thomas and Michael are doing with Premera buys that time

---

6. TEAM OF 5 UPDATE

• They're going to shut down or sell in the next ~2 months
• Open to acqui-hire: a few million, mostly earnout + vested equity
• Dan's thinking: structure a deal where you get the IP + them long enough to integrate, but don't overindex on them being the forever anchor team
• "I'm gun shy given the fact that they are quitting because they gave themselves a year"
• Andrew is the commercial guy, plus a product guy and clinical data engineers
• If Premera scope grows, could potentially hire them as contractors through DaisyAI and arbitrage (bill $150, pay $75)

---

7. DAN STARTS AS CTO OF SUMMA HEALTH (Monday!)

• 3-month contract as outside CIO of Summa Health (hospital system in Akron, OH)
• Brought in by General Catalyst through the Percepta relationship
• GC/Percepta has been frustrated: months of work and they can't get a single LLM provisioned within Summa's infrastructure
• Dan's role: go inside Summa and "clear the way" for Percepta's work to actually happen
• Reports to CEO and CFO of Summa, but brought in by GC
• Thomas connection: Abhinav (Percepta, close friend from school) and Lipsa are both at Summa doing call center work
• Dan: "In about three months from now, I think all of us are going to have collectively a lot more insight"
• Will test whether the Palantir-esque FDE model works in healthcare

---

8. FUNDRAISING PUSHBACK

Dan pushed hard against raising right now:

• "Why would you fundraise right now if you've got the FDE thing? What do you need the money for?"
• At $150/hr you can pay yourself properly — that's the insurance you need
• Better time to raise: After coming out of Premera with a product built on real learnings, in use at a health plan, with a pipeline. "Here's our pipeline — that would be a much more successful capital raise."
• AI startup space is crowded. Investors proxy on revenue. Without revenue, it's just a call option.
• Fundraising takes minimum 3 months — "you're going to be trying to do that while really what you should be doing is 100% focused on crushing the Premera opportunity"
• "Leave nothing on the table over there"
• Alternative: Try to get more money from Premera — expand to 4-5 person scope, get expense reimbursement for dev platforms, bundle operating costs in
• If fundraising is easy and pricing is good, "then take the money"
• Offered to invest: "Send me the deck and I may invest"

Thomas's response: "I don't think we've gotten the feedback to not do the fundraise yet, but I think given we just sent this contract off this week... maybe the reality of revenue coming sooner than we expected might give us more optionality."

---

9. IP PROTECTION REMINDER

• "Make sure you've got the right protections around IP residuals"
• Use separate computers for Premera work vs. DaisyAI work
• Everything in your head is your IP if developed without Premera IP
• Define background IP / foreground IP clearly in the contract
• "You want to have a place to do stuff that's super clean"

---

Action Items

• Review Premera contract for IP protections (background IP, residuals, separate dev environments)
• Talk to tax advisor about S Corp / LLC subsidiary for consulting income
• Seriously evaluate pausing fundraise to focus on Premera
• Thomas: say hi to Abhinav at Percepta/Summa
• If Premera scope grows, explore bringing team of 5 as contractors
• Stay tight with Dan through Summa stint — convergence in ~3 months
• Dan offered to invest — send deck when ready

People Discussed

Strategic Takeaways

• Revenue > fundraise right now. The Premera contract changes the calculus. Dan's pushback is the strongest signal yet to deprioritize the raise.
• The NewCo concept is still half-baked but directionally interesting. DaisyAI as FDE services + Stochastic as platform partner. Need Glenn's buy-in.
• Appeals is a live opportunity on both sides. Dan building provider-side appeal letters, Premera wants payer-side appeal defense. Collaboration potential is real.
• Summa/Percepta will be a case study. If GC's FDE model can't even get an LLM provisioned, that says a lot about the market reality.
• The team of 5 has a ~2 month window. If we want the IP, the clock is ticking. But the Premera work comes first.