The group CEO of a large insurer recently told us he now has “two or three layers of junior underwriters whose roles I don’t really need going forward.” His real worry is succession: if AI does the work those underwriters learned on, who will replace his senior people when they retire?
Firms shouldn’t try to rebuild the old development path, and buying experienced teams only moves talent from one firm to another without closing the development gap. Instead, firms need a model in which AI supplies the practice and senior people coach. Early-career underwriters start in shadow mode: they assess a risk on their own before seeing the AI’s view or a senior’s, then compare. Once their calls hold up, they move to parallel mode, assessing and pricing alongside the AI, with a senior reviewing the differences. Brokers follow the same pattern for winning business: AI-prepared account briefs, negotiations rehearsed in simulation, and an early seat in real pitches.
Rather than banking all the time AI saves as headcount cuts, firms should reinvest some of it in building the technical and commercial capabilities that strengthen the judgment of high-potential people, with progress measured against the senior roles they are training for.
Senior Talent Is Leaving Faster Than It Can Be Replaced
The problem is most acute in large commercial insurance, where senior judgment is hardest to develop and the people who hold it are closest to retirement.
- Senior talent is aging: About a quarter of the London market workforce is over 50[1]. In US broking, the average age of established producers at top agencies rose from 48.6 to 49.9 between 2022 and 2024[2].
- Fewer young people are coming in behind them: London market early-careers hiring fell 25% in 2025, a further 25% fall is forecast for 2026, and under-30s could shrink from 24% of the workforce to 7% within a decade[3]. In the US, only about half of new producers ever validate, and even top agencies underinvest in new ones[4][5].
- Buying experience moves talent around without adding any: Large team lift-outs have become a feature of commercial broking, and some have ended up in court. Lift-outs can work for the buyer, but they don’t grow the industry’s senior bench, and they make every firm’s best teams a target.
The Old Development Path Was Already Disappearing
Many firms had already moved routine work such as intake, re-keying, triage and certificates to operations teams, often offshore, and early-career underwriters and brokers lost the work they learned from. AI is accelerating that shift, and this time the roles are going away rather than moving elsewhere. Chubb aims to automate about 85% of its major underwriting and claims processes within three to four years[6]. Firms that want future leaders will have to design the path deliberately.
What Future Leaders Will Need
Firms will likely need fewer senior people, and those people will need two skill sets above all:
- Technical depth: specialist and sector knowledge, and the judgment to assess a complex risk the first time it’s presented.
- Commercial skill: developing clients and markets, giving advice, and winning business on the strength of trust and breadth of knowledge.
AI is absorbing the volume work through which underwriters built the first, and the traditional apprenticeship was always slow at teaching the second: many learned to win business only by sitting next to a good senior for years. As we argued in The Underwriter Redefined, the underwriter’s value is shifting toward specialization, advice and growth, and development has to follow.
Point Development At New Business
Renewals are repeatable, data-rich and most of a commercial book’s volume, so they will likely be automated first. New business needs judgment on an unfamiliar risk and a relationship that has to be won, so it will stay high-touch for longer.
In our work with insurers and brokers, carriers that redesigned their underwriting workflow first have moved 60% to 70% of renewals onto accelerated paths[7]. But renewals were where people learned the book, so early-career staff now need a structured route into new business much sooner.
Five Moves To Build The Next Generation
- Train technical judgment in shadow and parallel mode: Use the gaps between the trainee’s call, the AI’s and a senior’s to sharpen both judgment and the referral rules. Focus the practice on new business and complex risks.
- Give early-career people a structured route into new business: Plan early exposure to prospect meetings and pitches, use AI to prepare account briefs, and rehearse negotiations in simulation. Start handing over books 18 to 24 months before a producer retires, with client introductions, joint reviews and a phased revenue split.
- Define the future senior roles, then capture the judgment at risk: Build the roles around the two skill sets, with a visible path and realistic timescales. Capture what’s worth keeping from today’s seniors as referral rules, annotated decisions and account playbooks.
- Write the AI policy into the underwriting authority manual: Set out what AI can decide alone, what it must refer, and to whom. Whatever the manual leaves to humans becomes the curriculum. And because regulators assume competent people oversee these decisions, a thin pipeline of judgment is also a compliance risk.
- Select and retain talent, and measure capability, not just output: Choose high-potential employees fairly, including experienced operations staff who already know the work and the clients. Plan to keep them, because fast-tracked talent will be in demand elsewhere. Track time to competence alongside productivity.
Questions For Today’s Leaders
- Which decision capabilities and client relationships would we lose if our ten most senior people left in the next two years?
- Which entry-level tasks have we moved to operations or automated, and what replaced the learning they provided?
- How long does it take a new hire to underwrite a complex new risk, or win a new client, on their own? Is that time falling?
Redesign Talent Development Alongside The Work
Nobody needs another generation of underwriters spending years re-keying submissions to learn what good looks like. Firms can instead give selected talent a structured path in which AI takes the repetitive work, deliberate practice replaces sheer volume, and senior people spend their time coaching rather than checking.
The CEO who no longer needs those layers of junior underwriters may already employ some of the people who could close the gap. The question is whether he develops them before his senior people retire.
Sources [1]. London Market Group, London Matters 2024 (24% of the London market workforce over 50); RSM UK, The future of insurance: an ageing workforce and a growing talent gap, 2026. [2] Big "I" and Reagan Consulting, Best Practices Study: 2022 results, Insurance Journal, September 2022; 2023 update, IA Magazine; 2024 update, August 2024. Weighted average producer age (WAPA) of validated producers at Best Practices agencies. [3] London Market Group, London Matters 2026 fact base; "Futures Academy enters fourth year amid hiring slowdown," Insurance Business, July 2026; "London Market Group reveals collapse in early careers hiring," Insurance Post, July 2026. [4] Big "I" and Reagan Consulting, 2023 Best Practices Study, producer success rate of 45.5%, as reported in Zywave, The Fast-Track Guide to Increased Producer Productivity and Validation, 2024; Reagan Consulting, Producer Recruiting & Development Study, 56% success rate, as reported by ProgramBusiness. [5] "Top independent agencies kept profit margins strong even as growth slowed," Insurance Business, August 2026, on the 2026 Best Practices Study update (net unvalidated producer payroll of 0.0% to 1.7% of revenue against a healthy range of 1.5% to 2.0%). [6] "Chubb to cut up to 20% of workforce in 'radical' AI drive," Insurance Business, December 2025; "Chubb lays out digital transformation plan," Coverager, December 2025. [7] SSA & Co. engagement observations from recent carrier underwriting workflow redesigns. Figures reflect observed engagement experience, not a market-wide estimate.