Daisy AI — Deck
Seed deck, January 2026
Slide 1: Title
DAISY AI
AI-powered services for healthcare operations
Thomas Startz & Michael Yuan thomas@daisyai.com | michael@daisyai.com
Slide 2: One-liner
We combine deep healthcare expertise with AI to transform how health plans handle complex administrative work.
Starting with utilization management.
Slide 3: The Problem
Healthcare back-office is massive and broken.
- Utilization management is a multi-billion dollar function across payers
- Incentives are misaligned between payers and providers
- Nurses and case managers are drowning in complexity
- Denials, appeals, rework — pure administrative waste
Most startups that try to fix this fail.
- They think technology solves everything
- They automate symptoms, not root causes
- They don't understand the economics or the relationships
Slide 4: The Insight
We see what others miss.
| Surface view | Real view |
|---|---|
| "Automate denial letters" | Payer-provider relationship dynamics |
| "Prior auth automation" | Incentive misalignment |
| "Tech problem" | People + process + trust problem |
The moat isn't what we know. It's how we see.
If you solve symptoms, you build band-aids. We understand the real issues.
Slide 5: The Solution
AI + expert services for healthcare operations.
- Not pure SaaS — services wrapped in software
- Not replacing humans — augmenting them
- Not generic AI — deep healthcare expertise
What we do:
- McKinsey work: strategy, problem identification, process mapping
- Palantir work: build the software, run parts of the business
- Teach customers to build — create AI capability inside their orgs
Slide 6: First Product
Inpatient Status Assessment
Decision support for inpatient vs. observation determination
The gap:
- Status decision happens at T+4-8 hours
- No one applies criteria at that moment
- Physicians decide without guidance
- Case managers review 14-24 hours later — too late
What we built:
- Pulls clinical data automatically (vitals, labs, diagnosis)
- Applies Two-Midnight reasoning in real-time
- Shows what supports the status, what's missing, what to document
Impact: $2,000-$5,000+ per case. Fewer denials. Less rework.
Slide 7: Why Now
Three things converged:
1. AI crossed a threshold
- LLMs can handle judgment-intensive healthcare tasks
- But you need domain expertise to apply them correctly
- Most tech companies don't have that — we do
2. Infrastructure caught up
- Claude Code, Vercel, Neon, GitHub
- Small team can ship what took 10 engineers five years ago
- We're not under-resourced — we're leveraged
3. Payers are under pressure
- Humana, United stock prices down
- Medicare/Medicaid cuts coming
- Can't attract tech talent
- They know they need to change — now
Slide 8: Traction
Pull, not push.
| Customer | Type | Stage |
|---|---|---|
| Major West Coast health plan | Commercial | Contracting (Spring) |
| AppriseMD | Outsourcing | Late stage |
| HURC | Outsourcing | Mid stage |
| Behavioral health groups | Payer-adjacent | Early conversations |
Health plan context:
- UM automation is one of their three AI pillars
- 12 projects in scope
- Was on 2027 roadmap — we pulled it to 2026
Slide 9: Why Payers?
Yes, they're slow. Here's why we're there anyway:
The pain is on their side
- They employ the UM nurses
- They set the rules providers must follow
- They bear the cost of misalignment
They're under unprecedented pressure
- Stock prices declining (Humana, United)
- Medicare cuts removing the government backstop
- Commercial margins compressing
They can't attract tech talent
- Not sexy enterprises for engineers
- We fill the gap they can't fill internally
Our approach: pull, not push
- Find what's 1-2 years out on their roadmap
- Accelerate it for them
- Don't sell — partner
Slide 10: The Team
Rare combination: healthcare PE + engineering + both build
Founder 1
- Princeton engineering
- Warburg Pincus healthcare PE
- Bridges business and technology in the same mind
Founder 2
- Healthcare investing background
- Built technical fluency through the partnership
- Ships production software daily
Why this matters:
- Most healthcare startups fail because tech people don't get healthcare
- Or healthcare people can't build
- We have both — in both founders
Slide 11: The Flywheel
The tools caught up to the ambition.
Coding in 2025 ≠ coding in 2015:
- Work with AI, not write from scratch
- Pick the right platforms, ship fast
- Taste and judgment matter more than algorithms
What this means for us:
- Two founders shipping production software
- Not under-resourced — leveraged
What this means for customers:
- If we can develop this capability internally, we can do it for them
- We teach them to build, not just use
- That's stickier than any software
Slide 12: The Ask
Raising $2-3M
Use of funds:
- First FDE hires (expand delivery capacity)
- Accelerate customer acquisition
- Extend runway through major closes
Milestones this round:
- Close health plan contract (Spring 2026)
- Prove model scales across 2-3 customers
- Build pipeline for Series A
What we're looking for:
- Partner who understands services + software
- Healthcare expertise
- Concentrated portfolio = real support
Slide 13: Summary
Daisy AI
- Healthcare back-office is broken — we understand why
- AI + expert services, not just software
- First product: inpatient status decision support
- Real traction: major health plan contracting
- Team bridges the gap that kills healthcare startups
- Raising $2-3M to scale
We're building the company that health plans wish they could build internally.
Slide 14: Contact
Thomas Startz thomas@daisyai.com
Michael Yuan michael@daisyai.com
New York, NY
Design Notes
When creating slides:
- One idea per slide
- Big text, minimal bullets
- Dark on light (clean, professional)
- No stock photos
- Traction slide and team slide = most scrutinized
Slide count: 12-14 for send, can expand for live
Key slides to nail:
- Slide 6 (First Product) — this is the "what do you actually do" answer
- Slide 8 (Traction) — proof it's real
- Slide 9 (Why Payers) — gets ahead of the objection
- Slide 10 (Team) — founder-market fit story