Refract VC
Deep DiveWarmMeetings
12026-02-05Meeting: Refract VCStrong update call with Caesar and Kerry from Refract. They'd met us at Wharton and their feedback then was that UM was "too important for insurers to outsource." This call showed them how much has changed — bootstrapping, learning to code, landing Premera. Caesar loved the copilot framing and the FDE approach, called the bootstrapping story "catnip for us." Two lingering concerns: scalability of custom work vs. redeployable technology, and market size. They offered to connect us with portfolio company Cloud Gurus for integrations. Said they'd circle back internally in the next couple of days.
▶
Strong update call with Caesar and Kerry from Refract. They'd met us at Wharton and their feedback then was that UM was "too important for insurers to outsource." This call showed them how much has changed — bootstrapping, learning to code, landing Premera. Caesar loved the copilot framing and the FDE approach, called the bootstrapping story "catnip for us." Two lingering concerns: scalability of custom work vs. redeployable technology, and market size. They offered to connect us with portfolio company Cloud Gurus for integrations. Said they'd circle back internally in the next couple of days.
Meeting: Refract VC
Date: 2026-02-05 Attendees: Thomas Startz, Michael Yuan, Caesar Djavaherian (Refract), Kerry Kellogg (Refract) Type: Follow-up / Update (met previously while at Wharton) Duration: ~27 min
Summary
Strong update call with Caesar and Kerry from Refract. They'd met us at Wharton and their feedback then was that UM was "too important for insurers to outsource." This call showed them how much has changed — bootstrapping, learning to code, landing Premera. Caesar loved the copilot framing and the FDE approach, called the bootstrapping story "catnip for us." Two lingering concerns: scalability of custom work vs. redeployable technology, and market size. They offered to connect us with portfolio company Cloud Gurus for integrations. Said they'd circle back internally in the next couple of days.
Questions Asked
| # | Question (verbatim/near-verbatim) | Topic | Answer Quality | Notes |
|---|---|---|---|---|
| 1 | "Tell us... take us back from when you were at Wharton. I think our feedback was it was too important for them to outsource. Great to hear about your progress." | Traction | strong | Good narrative arc — graduated, moved in together, learned to code, built product, landed Premera. Caesar was visibly impressed. |
| 2 | "Can you go into specifics? Are you able to tell us who your partner is and what the size of the contract is, how you're going to structure the revenue?" | Traction / Business Model | decent | Named Premera, described scope (12 AI initiatives, inpatient first). But contract size = "small" and revenue structure = "haven't laid out exact terms." Honest but thin. |
| 3 | "How did you meet them and who's your champion?" | Traction | strong | Wharton professor on the board → head actuary (Dave) → head of analytics (Colt). Clear chain. CMO also excited. |
| 4 | "Directionally like size and scope of what you would do initially?" (Kerry) | Business Model | decent | "Small contract to begin with" — framed as intentional learning. They accepted it but will want more specifics next time. |
| 5 | "How much of the technology is one where you can redeploy versus are you just a consulting outsourced hiring firm customizing every build?" | Business Model / Scalability | decent | Thomas: "We want to be much more the former... a licensor type core technology thing." Acknowledged the tension honestly. Talked about templates/primitives. But no crisp proof point of what's redeployable today vs. custom. |
| 6 | "Are you able to go into what you've created so far?" | Technical / Product | strong | Live demo of copilot — chart upload, AI extraction, timeline, triage view. Caesar and Kerry clearly engaged. Caesar: "I love the idea of a copilot." |
| 7 | "Did you guys say that you worked on HIPAA compliant and SOC2?" | Technical / Compliance | decent | HIPAA yes, SOC2 Type 2 not yet, pen test needed (~$5K). Honest, they were fine with it. |
| 8 | "It sounds like software that replaces some labor... Could you remind me how much you guys raised to date?" | Business Model / Fundraise | strong | "Nothing. We've been bootstrapping." Caesar loved it — validates commitment. |
| 9 | "Is your sense you're gonna stay on the payer side and not so much on the provider side?" | GTM / Strategy | decent | Explained payer-first but will touch hospitals through payer relationships and outsourcer clients. Clear enough, but the multi-stakeholder positioning could be tighter. |
| 10 | "How much would you guys need to get to... I imagine you're looking to raise some funds now to get some breathing room?" | Fundraise | decent | $2M. Hire 2 more technical (PhDs like us) + 1 clinical/GTM (Cassie). Get to 6-7 clients. Clear and specific. |
| 11 | "If you were to only get $1 million, where do you think that would take you?" | Fundraise | decent | Through another year, one more hire, 5-6 clients. Showed he'd thought about scenarios. |
Concerns Raised
-
Scalability of custom work — Caesar's central concern. "How much can you redeploy vs. customizing every build?" He explicitly said customization "is hard to scale." Suggested self-service options and back-end optimization as the scalable path. This is the question they need to get comfortable with.
-
Market size — Caesar at the end: "We just have to wrap our heads around the market size and the opportunity." Not a rejection, but a real open question for them.
-
Original concern evolved — Their previous feedback was "too important to outsource." They acknowledged this has shifted. Caesar mentioned JP Morgan conference conversations validating that insurers are actively looking for AI help. Charter Health (Refract portfolio company) landing Blue Cross Michigan 7-figure contract further validates.
What Resonated
-
Bootstrapping + learning to code together — Caesar: "I can't tell you how much I like the fact that you both moved in together. You started coding. That's like catnip for us."
-
Copilot framing — Caesar: "I love the idea of a copilot." He immediately saw the go-to-market advantage — sell without deep integrations, nurses keep their workflow.
-
Landing Premera — Caesar: "It's incredible." The fact that it was on their 2027 roadmap and we pulled it to 2026 landed.
-
Catalyst framing — Caesar paraphrased it back: "So basically you're acting as a catalyst to help them deploy their own vision of their processes and systems using AI." Understood it perfectly.
-
Timeline extraction / demo — Caesar: "I love that it can extract a timeline and all the things that nurses spend a lot of time on just reading."
-
Two smart guys, still committed — Caesar: "We absolutely love you guys... driven, graduated, still doing this, moved in together, doubled down. Really strong signal."
-
Payer pressure narrative — Thomas's point about headwinds not abating, government not being generous, utilization not going down. Caesar nodded along.
Our Commitments
None explicit from this call.
Their Next Steps
- Caesar and Kerry to circle back internally ("next couple of days")
- Offered intro to Cloud Gurus (portfolio company) for integration acceleration
- Implied: need to get comfortable with market size thesis before moving forward
Investor Insights
Firm context:
- Refract VC, Caesar Djavaherian (partner), Kerry Kellogg (partner/associate)
- Continuing to deploy current fund. Three exits last year — good fund performance signal
- Refining healthcare investing approach, getting deeper into insurance marketplace
- Closer with insurance executives, understanding defensive posture against provider AI tools
What they care about:
- Founder commitment and grit (bootstrapping, coding, living together)
- Scalability path — not just consulting/customization, need redeployable tech
- Market size and opportunity scope
- Insurance companies as buyers (validated through their own portfolio + JP Morgan conversations)
Portfolio signal:
- Charter Health — landed Blue Cross Michigan, 7-figure annual contract. Selling to insurers works in their experience. This is a positive comp for us.
- Cloud Gurus — integration/deployment help. They offered to connect us. Investing in the relationship.
Their thesis evolution:
- Previously skeptical that insurers would outsource UM
- Now seeing (via JP Morgan, portfolio, market) that insurers are desperate for AI help and don't have talent
- Payer market conviction increasing
Process signals:
- No formal process described
- "Circle back in a couple of days" suggests relatively fast internal discussion
- They engaged for full 27 min, offered portfolio intro — not a courtesy call
Interesting Moments
"Catnip for us" — Caesar's genuine reaction to the bootstrapping/coding story. Strong founder-market fit signal.
Caesar paraphrasing our thesis back to us — "So you're acting as a catalyst to help them deploy their own vision." When an investor articulates your thesis better than you, they get it.
"Every single payer C-suite discussion is revolving around what tools to implement" — Caesar sharing his own market intelligence. Validates our timing thesis from their LP/portfolio perspective.
Kerry mostly quiet but engaged — Asked the practical questions (how did you meet them, size and scope). Probably the diligence-oriented partner.
Scalability advice — Caesar's suggestion about self-service options and back-end optimization wasn't just a question — it was him telling us what the scalable version looks like. Worth taking seriously.
Charter Health as internal comp — Caesar referenced their own portfolio company landing Blue Cross Michigan. He's pattern-matching us to a company they already believe in. Positive signal.
Raw Transcript
[See above — 27 minutes, Feb 5 2026]
Top Concern
Scalability + market size
Overview
- Stage focus: Seed / early
- Check size: TBD — need to confirm
- Healthcare thesis: Getting deeper into insurance marketplace. Seeing payer C-suites scrambling for AI tools. Portfolio includes Charter Health (Blue Cross Michigan, 7-figure contract) and Cloud Gurus (integration/deployment).
- Key people: Caesar Djavaherian (partner), Kerry Kellogg (partner/associate)
Relationship History
| Date | Type | Notes |
|---|---|---|
| Pre-May 2025 | Intro (at Wharton) | Liked us, liked domain expertise. Feedback: UM "too important to outsource." Passed at the time. |
| 2026-02-05 | Update call | Strong reconnect. Impressed by progress — bootstrapping, Premera, product demo. Concerns: scalability, market size. Offered Cloud Gurus intro. |
Status
- Current stage: Re-engaged / Update → Pending internal discussion (OVERDUE)
- Next step: Follow up with Caesar. Said "couple of days" on Feb 5 — it's been 8 days.
- Temperature: Warm (upgraded from their previous pass, but silence is a flag)
What They Care About
- Founder grit and commitment (bootstrapping is "catnip")
- Scalability — redeployable tech vs. pure consulting/customization
- Market size of the payer AI opportunity
- Insurance companies as viable buyers (validated through their own portfolio)
- Self-service / back-end optimization as the scalable path
Concerns to Address
- Scalability — Caesar's main question. Need to articulate what's redeployable today, what the platform becomes, how custom work decreases per client over time.
- Market size — "We just have to wrap our heads around market size and opportunity." Need TAM framing.
- Contract specifics — They asked and we were vague. Next time, have Premera structure clearer.
Materials Sent
- Deck (not sent yet — they have it from before, may need updated version)
- Two-pager
- Product brief
Notes
- They came to us previously at Wharton, passed, now re-engaging. This is a warm re-open, not a cold intro.
- Caesar shared real market intelligence (JP Morgan conference, Charter Health). He's invested in the thesis, not just evaluating us.
- Kerry is the quieter, more diligence-oriented partner. Will likely drive the detail questions if they advance.
- Cloud Gurus intro could be valuable regardless of investment outcome.
- Three exits last year — fund is performing, they have capital to deploy.