“I need you to fix carrier partner restrictions. And by that, I mean every possible scenario.” — Lead Engineer, Feb 2, 2026

Emergency Redesign: Rebuilding Partner Authorization in 2.5 Weeks

With homeowners insurance launching in 8 weeks to 1,300+ partner agencies, Obie’s authorization system was broken. I reframed the data model, led a rapid redesign, and shipped a self‑service system before the deadline hit.

RoleSole product designer
TimelineFeb 2 – Feb 18, 2026
Scope1,300+ agencies, 6 programs
The Emergency
The authorization system was breaking the launch

Obie was preparing to launch homeowners insurance to 1,300+ partner agencies. The existing authorization model created 300+ database records per partner and couldn’t handle program‑level restrictions, staged rollouts, or compliance requirements.

Constraint: Homeowners launched on 4/6. We had 2.5 weeks to redesign the entire authorization system or the launch would fail.
Scale
1,300+
Partner agencies
Legacy Model
306
Records per partner
Timeline
2.5 weeks
From emergency to approval
Impact
99.7%
DB complexity reduction
The Reframe
From “restrictions” to “allowances”

Instead of tracking everything a partner couldn’t do, we inverted the model: define baseline capability, then layer explicit exceptions. That single shift aligned with how admins think and unlocked the scalability engineering needed.

Old system: carrier_partner_restrictions
- 1 partner × 51 states × 6 carriers = 306 records
- Massive table joins
- No program-level restrictions

New system: partner_programs (JSONB)
- 1 partner = 1 record
- Store what partners CAN do (allowances)
- Missing = restricted
- Rules-engine compatible
“That is the high level.”Lead Engineer, confirming the reframe
AI‑Accelerated Prototyping
From six weeks to one morning

Instead of a traditional Figma → handoff cycle, I used Claude Code to build working prototypes stakeholders could interact with the same day.

Result: Compressed a 6‑week design cycle into a 2.5‑week emergency build with real testing in the first morning.
Feb 9 — 9:30 AM
Prototype v1 shipped (“claude and i are COOKING”)
Feb 10
v2: geo regions, program pills, clone restrictions
Feb 17
v3: multi‑select, search for 1,300+ agencies
Feb 18 — 8:09 AM
Final approval (“A‑M‑A‑Z‑I‑N‑G”)
Stakeholder Facilitation
Resolving conflicting mental models

Mid‑project, I discovered the AD team and Engineering had opposing interpretations of how restrictions should work. I facilitated a conversation that reframed the conflict as layered systems, not competing requirements.

AD Team model
- State licensing = access
- Programs managed separately
- “Why program-level state restrictions?”
Engineering model (Swiss cheese)
- Licensing, moratoriums, program rules
- Any layer can block access
- Needs state-by-program granularity
Principle unlocked: Make simple things simple, make complex things possible.
“If you can build this while making it easy to manage for my team, I don’t want to be a blocker. Deal.”AD Team lead, after alignment
The System
Progressive disclosure for complex rules

The final system supported bulk operations, state‑level overrides, and a clear audit trail without overwhelming admins who only needed simple enable/disable workflows.

Outcome
Shipped before launch with measurable impact
Database
306 → 1
Records per partner
Speed
6+ weeks → 2.5
Design cycle reduction
Launch
On time
Homeowners rollout
Scale
1,300+
Agencies enabled
“I think that looks fantastic.”Lead Engineer, final approval
Reflection
What this demonstrates

When the lead engineer asked for “every possible scenario,” I didn’t stall or scope‑reduce. I reframed the system, built working prototypes fast, and aligned cross‑functional teams around a model that could scale. This is the kind of work I love: high‑stakes, high‑ambiguity, and deeply technical.