Premera Blue Cross

Engagement

Premera wants to expand AI capabilities in utilization management but is constrained by talent and capacity. They've been building AI systems for ~2 years and have production deployments (auto-auth, ADT feed processing, prioritization models). They don't need us to teach them AI — they need us to augment their team and accelerate their roadmap. FDE/consulting model, not SaaS.

The specific opportunity: they had a 2-year AI roadmap and moved year-1 items to year-2 due to capacity constraints. Colt explicitly floated "could we accelerate year-2 to year-1" with our help. Domains mentioned: imaging (ortho), lab auto-auth, insourcing to remove vendor expense.

Working Team

Daisy AI: Thomas Startz, Michael Yuan Connection: Brad Flugel (Wharton healthcare professor)

Premera — On the proposal:

• David Braza — executive sponsor, our entry point
• Colt Courtright — Runs UM AI domain. ~12 projects. Technical leader. Most substantive counterpart so far.
• Romilla Batra — connected us in, briefed Chad, based in LA
• John Hauser — on proposal (role TBD)
• Nathan Crock — on proposal (role TBD)

Premera — Key but not on proposal:

• Chad Murphy — Chief Clinical Officer (pharmacist). Runs UM operations. Business leader. Haven't met yet — this is the key meeting.
• Jamie Halstead — Runs ops alongside Colt. AI/automation path. Haven't met yet.

Premera — Other:

• Kristina Rose (kristina.johnsonrose@premera.com) — invited to Jan 5 call, couldn't attend
• Beth — Dave's admin, scheduling
• Mark — vendor management (contracting goes through him)
• Talhah/Tal — CFO (funding decisions)
• "Caroline" — current vendor for advanced imaging reviews (disruption candidate)

Current Status

Phase: Warm handoff to Chad/Jamie/Colt operational team Last interaction: Jan 28 — short call confirming interest, Dave routing us to Chad/Jamie Next: Call with Chad Murphy, Jamie Halstead, Colt (Beth scheduling)

Dave and Romilla have confirmed they want to work with us. We've had the demo (Jan 5) and the follow-up (Jan 28). Now it's about getting in front of Chad (CCO) who runs the UM roadmap. That call determines what we work on and how we contract.

What We Know About Their UM Operation

Three UM milestones (from Romilla):

1. Initial prior auth — patient arrives, admit/observe/not covered? Runs through InterQual. Their #1 priority.
2. Concurrent review — day 5 check (DRGs cover days 3-5, bandwidth limits earlier review)
3. Daily reviews — contract-specific daily authorization

What they've already built (from Colt):

• Live ADT feed processing, updating ~every 15 min (very chatty HL7)
• Fully automated non-inpatient reviews (frees bandwidth for inpatient)
• Auto-auth: eliminates 25-50% of manual work on all requests
• AI accuracy ≥ human clinicians for approvals
• Can automate approvals end-to-end (including letters) but NOT denials (regulatory)
• Summary prediction for prioritization across admissions in WA and AK
• 80-90% of admissions arranged where they have ADT feeds; CCDs much sparser
• Claims data gives broad clinical history (diagnosis/therapy, not allergies/family)
• 2+ years of building and deploying to production

Key constraints:

• Must use evidence-based criteria (InterQual/MCG) — can't use AI-generated criteria (legal risk)
• Work within their environment and tools — may not need our UI at all
• Regulatory: can automate approvals, not denials

What Matters

• Chad meeting is the gate. He runs the UM roadmap with Jamie. If we land well, we're in.
• Year-2 → Year-1 acceleration is the pitch. Colt framed it explicitly. Specific items: imaging (ortho), lab auto-auth, insourcing from vendors.
• We're augmenting, not replacing. They have sophisticated systems already. Our value is capacity and speed.
• Initial prior auth > concurrent review. Romilla was clear. We demoed concurrent; they want prior auth first.
• API/backend > UI. Colt suggested we may not need a UI. Work within their existing tools.
• Business model alignment is strong. All three agreed: FDE/consulting > SaaS. Dave: "a lot of vendors are going to be disrupted."
• Contracting goes through Mark (vendor mgmt), funding through Talhah (CFO).

Open Questions

See questions.md. Priority for Chad/Jamie call:

• Which year-2 roadmap items can we take on?
• What does a contract look like through Mark/vendor management?
• What's the working relationship — embedded in their environment?
• Who's day-to-day counterpart?
• What access do we get (ADT feeds, their dev environment)?

Scope

Near-final proposal defines scope: chart summaries, data ingestion (HL7/PDF/file share), template alignment, UM flag extraction, analytics dashboard, training. InterQual/MCG integration explicitly out of scope. 12+ week phased timeline (contracting → discovery → integration → evaluation). BAA required.

See scope.md for complete proposal.

Commercial

Status: Not formally discussed with Premera. Proposal lists implementation fee and ACV as TBD. Contracting path: Through Mark (vendor management), funding through Talhah (CFO) Their framing: capacity augmentation, accelerate roadmap, insource from vendors

See commercial.md

Risk

• Chad meeting is the gate — haven't met the decision maker yet
• They're already sophisticated — we need to show we add value beyond what they already have
• We demoed concurrent review; their #1 priority is initial prior auth (which we haven't built yet)
• InterQual/MCG out of scope intentionally — strategic decision, don't rush
• Contracting could be slow (vendor management process, CFO approval)
• Working in their environment adds technical dependency we can't control

Documents