Another week, another “game-changing” agent protocol drop. Cool – but plumbing isn’t purpose. Since OpenAI's MCP, we've seen frameworks from Google, IBM, Microsoft, and specialized protocols like NANDA. The AI community debates each, as if picking the "winning" protocol determines success.
The Insurance Reality Check
Insurance headcount grew from 2.8M (2016) to 2.98M (2023), with BLS projecting 3.08M by 2033. Behind that modest 3.4% growth lies a massive hiring challenge.
The industry needs to hire 450K people total—100K for growth plus 350K retirement replacements. Where's the growth happening based on the net numbers from BLS data?
- Knowledge work is exploding: data scientists (+33%), actuaries (+21%), operations research (+23%)
- Sales still dominates: 63K new frontline sellers needed
- Tech talent stays scarce: 30K+ computer/math roles
You can't hire your way out. Perfect terrain for Agentic AI to augment existing experts.
Think of the competition for talent. Only 4% of millennials find insurance appealing. Attracting digital native Gen Z will be nearly impossible. The real challenge is skills evolution and talent competition.
Let's look at this data point: 762K professionals in business and financial roles (26% of workforce, 1.5% growth) already understand insurance's complexities. Instead of hiring externally, use agentic AI to handle their routine tasks and elevate them into higher-skilled roles. You get the contextual knowledge while alleviating the talent acquisition burden.
Stop Debating Platforms, Start Solving Problems
Insurers are embracing AI, Celent's research shows they’re Early Adopters (14% of the market), 28% already deployed GenAI by April 2024, with nearly half expecting production by 2025.
But while we debate AI platforms, protocols and models, we're missing the fundamental point: adoption and scale require industry collaboration, not individual platform choices.
Consider what IMOs and carriers spend collective millions on annually:
- Agent onboarding and training (Carrier specific training across multiple IMO relationships)
- Compliance monitoring (50+ state regulations tracked separately)
- Customer acquisition (duplicated lead generation efforts)
Meanwhile, agents struggle with inconsistent tools, duplicated data entry, and fragmented experiences across carrier relationships.
Service Design First, Technology Second
Focus on the use case and service design, then choose technology to deliver it.
From the latest BLS data, the industry will need net new 60K people for sales-related roles. How do we make this function attractive and reduce friction to improve efficiency and productivity?
Take agent onboarding as an example. Start with the real challenge: new agents typically take months to become productive, facing inconsistent training, and fragmented systems.
Ask instead:
- What could cut agent time-to-productivity by half?
- Where do new agents get stuck, and why?
- What do top-performing agents wish they had on day one?
The answer isn't a technology platform—it's service design that spans IMO-carrier boundaries. Here's how to build it
The Roadmap to Success with Agentic AI
Start with service design, not technology selection:
1. Map agent skills and knowledge: What expertise do your top performers have? What decisions do they make?
2. Identify knowledge gaps: Where do new agents struggle? Where is critical data scattered or hard to access?
3. Catalog tools and handoffs: Which systems require human intervention? Where do workflows break down between departments?
4. Design the augmented experience: How could an AI agent handle routine decisions while escalating complex ones to humans?
5. Then choose enabling technology: Platform decisions become obvious once service design is clear. With AI evolving monthly, avoid vendor lock-in. Choose solutions that let you pivot without rebuilding your entire service design
The companies mapping skills and knowledge first will discover that agentic AI doesn't just automate tasks—it augments human expertise at scale. The difference? You'll choose it because it solves your talent challenges, not because it's trendy
The competitive advantage isn't just your AI platform—it's how quickly you solve real problems while competitors argue about infrastructure.
