AI In Insurance: Misconceptions Are Half The Battle, Says Equisoft's Brian Carey
Misperception slows AI adoption more than the tech itself. That's the message from Brian Carey, VP of Insurance Solutions Engineering at Equisoft.
The company put AI directly into people's hands, started small, proved value, and scaled. That simple sequence turned curiosity into daily use.
Start with use, not theory
Equisoft partnered with Anthropic's Claude and kicked off a pilot with clear, low-risk use cases. The goal wasn't to "talk AI." It was to remove drudge work and show time saved.
Once teams felt the lift, momentum built. AI became indispensable across project management, business analysis, development, and more.
Company-wide adoption, real gains
After fully adopting Claude in September last year, Equisoft expanded access beyond delivery teams. Marketing, finance, even legal got involved.
The most consistent outcome: time savings. Repetitive tasks got automated or accelerated, letting people focus on higher-value work. Feedback across the org has been overwhelmingly positive.
Integrated into insurance workflows
Equisoft didn't stop at internal productivity. They're baking Claude into client-facing software to improve underwriting, claims, suitability reviews, and commissions processing-areas with long cycle times and error risk.
One practical target: NIGO (not-in-good-order) processing. If AI can reduce NIGO and speed cleanup, you cut costs and shorten issuance times.
Clear the myths first
- Prompting isn't model training. Typing instructions doesn't alter the base model.
- People hesitate to deploy customer-facing AI. The tech exists; the gap is trust, guardrails, and proof of value.
- Literacy matters, but usage matters more. Skills stick when people apply them to real work.
Equisoft's playbook you can copy
- Pick a safe, secure platform and set guardrails (PII handling, logging, access control).
- Select 2-3 high-friction use cases per function (e.g., NIGO review, requirements extraction, claims notes summarization).
- Run a 2-4 week pilot with clear baselines and success metrics.
- Train teams on practical prompting and "when to use AI" decision rules.
- Survey monthly, share wins, and build internal champions.
- Integrate into existing systems and SOPs once value is proven.
- Expand horizontally (marketing, finance, legal) after core workflows show lift.
Metrics that convince executives
- NIGO rate reduction and rework time saved
- Underwriting cycle time and first-pass yield
- Claims resolution time and exception rates
- Hours saved per role per week
- Error rates in suitability and commission calculations
- User adoption, satisfaction, and repeated use of approved prompts
Build literacy the right way
Equisoft invested in education, bringing in Anthropic to train staff and address common misconceptions. Hands-on learning made the difference: ask the model, get help, apply it immediately.
If your teams need structured upskilling, consider role-focused training and certification. It accelerates safe adoption and creates consistent standards.
Resources
- Anthropic Claude
- OpenAI ChatGPT
- Google Gemini
- Claude Certification (Complete AI Training)
- AI Courses by Job Role
What insurers should do next
Pick one workflow and prove value. NIGO clean-up, claims summarization, or producer inquiries are strong starts. Keep scope tight, measure aggressively, and share results.
Then expand to underwriting and commissions, where speed and accuracy pay off fast. The tech is ready. Your advantage is execution.
About Equisoft
Equisoft is a global provider of digital solutions in life insurance and annuities. Founded in 1994, the company is based in Montreal, Quebec, Canada.
Your membership also unlocks:
