For insurance executives, artificial intelligence is no longer a side project in the innovation lab. It is becoming a core driver of how policies are sold, underwritten and serviced - and new research suggests it could already take over a significant share of the work your organisation pays for today.
MIT’s Project Iceberg, a largescale simulation of the U.S. labour market, finds that current AI tools are technically capable of performing tasks worth 11.7% of total wage value, or about US$1.2 trillion annually. The model represents around 151 million workers across 923 occupations and more than 32,000 skills.
Insurance is squarely in the zone of highest exposure: documentheavy, ruledriven, and dense with administrative and analytical tasks that can be codified and automated.
Project Iceberg doesn’t simply speculate about “jobs of the future”. It compares, task by task, what people actually do with what existing AI systems can already do.
Each worker in the simulation is treated as an “agent” with a bundle of skills and tasks, mapped against thousands of deployed AI tools.
The central metric, the Iceberg Index, measures technical exposure: the share of an occupation’s wage bill tied to skills where AI has already demonstrated usable performance in at least one context. The authors stress that it “captures technical exposure, where AI can perform occupational tasks, not displacement outcomes or adoption timelines.”
For insurance employers, that means:
Public attention has focused on AI writing code and replacing some technology roles. In Iceberg’s terms, that’s just the “surface”.
Looking only at current AI adoption in computing and technology occupations, the researchers find exposure of about 2.2% of wage value — roughly US$211 billion — which they call the “Surface Index”. They emphasise this is “only the tip of the iceberg.”
The real mass of capability “extends far below the surface through cognitive automation spanning administrative, financial, and professional services,” accounting for the full 11.7% and around US$1.2 trillion in wages.
In other words, AI is already strong at exactly the kinds of tasks that dominate insurance operations:
One summary highlights high Iceberg scores driven by “cognitive work—financial analysis, administrative coordination, and professional services.” For insurers, substitute policy servicing, claims handling and underwriting support, and the exposure becomes obvious.
The Iceberg report doesn’t list insurance job titles, but the task patterns it highlights map cleanly onto common insurance workflows.
Tasks with the highest nearterm automation potential include:
The authors note that financial institutions already deploy AI for “document processing and analytical support,” while healthcare systems automate “administrative tasks.” Insurance operations combine both: financial logic sitting on top of dense medical, legal and technical documentation.
This doesn’t mean whole roles vanish overnight. It means:
How rapidly that shift happens is now a commercial decision.
Project Iceberg was designed so governments can “identify exposure hotspots, prioritize training and infrastructure investments, and test interventions before committing billions to implementation.” For insurance employers, it effectively asks:
If AI can already perform 10–15% of the work you pay for, how will you use that capability?
There are three broad paths.
Use AI mainly to reduce headcount in operations, claims and admin, driving down the expense ratio. This is the fastest route to shortterm margin improvement, but carries risks:
Hold overall headcount relatively flat while:
This approach aims to improve combined ratios and NPS at the same time — but it requires deliberate redesign of roles and incentives.
Use AI as a catalyst to rethink your entire operating model:
Iceberg doesn’t tell you which path to take. It tells you you’re already standing at the fork.
To make this research actionable at executive level, insurers can take four concrete steps.
Ask each major area — underwriting, claims, operations, finance, distribution — to:
Even a firstpass estimate of the percentage of wage spend on highexposure tasks will reveal where the business case for automation is strongest — and where workforce risk is highest.
Use those exposure maps to run 3–5 year scenarios:
For each scenario, estimate impact on:
This is exactly how Iceberg is meant to be used — to support “evidence-based planning as AI capabilities expand across the economy.”
Technology teams will naturally focus on tools and platforms. The executive team needs to decide, in parallel:
Without these decisions, automation will drift towards piecemeal costcutting, weakening longterm competitiveness.
Finally, elevate this beyond operational efficiency. Boards and risk committees should be seeing:
The MIT team behind Iceberg is blunt: “The window to treat AI as a distant future issue is closing.” For insurance employers facing margin pressure, regulatory scrutiny and rising customer expectations, that window may be shorter than most realise.
AI is already capable of doing a meaningful share of the work across your value chain. Whether that becomes a story of leaner, more customercentric growth — or of abrupt cuts and eroded trust — will depend on the strategic choices you make now, while the iceberg is clearly visible and still some distance ahead.
