Skip to main content

Case A Expand I: Meridian — Discovery and Problem Reframe

· 2 min read
AI Playbook author

Case A · Expand I (parent: Meridian overview). From market signal through discovery to a written problem reframe.


1. Pre-opportunity context

  1. Market sensing — insurers want GenAI; fear harm, privacy, hallucination, regulation, legacy, weak governance.
  2. Marketing — responsible-AI report + insurance roundtable → senior conversation, not a product pitch.
  3. Account — relationship partner with COO and CRO on priorities, not demos.

SE lesson: content that surfaces risk and operating-model questions creates better opportunities than feature webinars.

2. Qualification

LensFinding
NeedReduce handle/search time; improve consistency
UrgencyBoard plan ~3 months
SponsorshipCOO + CRO
FundingLikely; not line-itemed
FitInsurance + AI + risk depth
RiskCustomer-facing GenAI too early; independence clear

Classifications considered: Qualified · Conditionally qualified · Nurture · No bid.
Outcome: Pursue. Decline single-phase “chatbot live in 90 days.”

Pursuit investment: insurance partner, AI solution lead, data architect, risk, change, commercial.

3. Discovery methods and findings

Methods: stakeholder interviews, knowledge walkthroughs, sample claim/process journeys, architecture skim, risk workshop.

Findings:

  • Agents waste time searching policy and process knowledge.
  • Answer quality varies by tenure; escalations spike on edges.
  • Customer-facing automation touches advice-like wording → higher risk tier.
  • No golden set, weak audit design, no clear knowledge owner.

4. Problem reframe

Weak askEngineered problem
Deploy a GenAI chatbotCut avoidable agent search time; improve answer consistency with human verification; build governance for future customer-facing AI

Logic: business objective → problem → root causes → capabilities → operating-model change → technology—not tool first.

5. Exit criteria to Expand II

  • Written reframe signed by COO/CRO sponsors in working session.
  • Risk tier agreed for Phase 1/2 (employee-facing assist ≠ customer advice).
  • Initial RAID: knowledge ownership, data quality, eval gap, change capacity.

Discussion

Comments

Share feedback or questions about this page. No account required.

Loading comments…