The European Insurer's Execution Gap
A 24-month pivot to the AI-insurer — perspective for CUOs, CCOs, CIOs and COOs
European insurance carriers have spent the last five years writing strategy decks that promise a digital, AI-native future. The operating models underneath those decks have barely moved. The gap between stated ambition and realised capability is widening — and the window to close it is now roughly 24 months, not five years. DORA, the EU AI Act and FIDA, combined with shifting distribution economics and accelerating pressure from embedded insurance and InsurTech challengers, will compound this gap — not forgive it.
The diagnosis — same disease, three dosages
Across our benchmarking of tier-1 Western European and Nordic carriers, mid-tier CEE carriers, and smaller regional players, the pattern repeats. Only the severity differs.
- Tier-1 carriers have the cash but are trapped by scale. IT spend sits at a healthy 2.5–4.0% of GWP, cloud workloads have crossed 50% and run-to-build ratios approach best-in-class (~55/45). But transformation is diluted across geographies, product lines and legacy cores that are 15–25 years old. AI is piloted in 30+ places and scaled in two or three.
- Mid-tier carriers have the clarity but not the foundation. IT-to-DWP ratios look comparable to tier-1s on paper (2.0–3.0%), but the spend is over-indexed to keeping the lights on — run is consistently 65–75% of total IT spend, digital FNOL shares hover at 10–20%, and cloud penetration is frequently below 30%. Technical debt, not ambition, is the binding constraint.
- Small and regional carriers have neither. They are effectively SaaS renters with no platform spine, no AI-ready data layer, and no capability to absorb DORA, Solvency II recalibrations and the AI Act simultaneously. They will consolidate or be disintermediated.
The benchmark view
| Benchmark | Tier-1 WE / Nordics | Mid-tier / CEE | Best-in-class target |
| IT spend % of GWP | 2.5–4.0% | 2.0–3.0% | 3.0–4.5% |
| Run vs Build mix | ~60 / 40 | 65–75 / 25–35 | 55 / 45 |
| Cloud workload share | 50–70% | 15–30% | 60–80% |
| Digital FNOL share | 30–55% | 10–20% | 50%+ |
| STP on priority journeys | 40–60% | 15–30% | 60%+ |
| Automated CI/CD pipelines | 50–70% | 20–35% | 70%+ |
| Cybersecurity % of IT | 8–12% | 5–8% | 10–14% |
Source: Cross-carrier IT operating model benchmarks, 2024–25 (tier-1: Helvetia Baloise, If/Topdanmark cohort; mid-tier: Euroins cohort).
Where strategy is failing execution — five honest failure modes
1. Run-heavy IT economics. When 65–75% of the IT budget is locked into legacy run, transformation is funded from the margin — which is exactly the margin combined ratios are now squeezing. Strategy is not funded strategy.
2. Project funding masquerading as product funding. Annual, siloed project funding still dominates, even in carriers that claim a “product operating model.” Without persistent, outcome-funded product teams around underwriting, claims and distribution, AI capability never compounds.
3. Pilot purgatory in AI. Most carriers have 20–50 GenAI pilots and fewer than five in production. The blocker is never the model — it is the absence of a governed data product layer, platform teams, and a deployment pipeline that can carry a pilot through compliance, security and underwriting sign-off.
4. Security and resilience bolted on, not engineered in. DORA’s enforcement tolerance for “we’ll remediate in the next release” is zero. Our benchmarks show only best-in-class carriers automate security across >60% of CI/CD pipelines; the mid-tier median is below 30%.
5. Distribution economics quietly collapsing. 40–65% of mid-tier premium still flows through intermediated channels, with digital quote-to-bind conversion under 20%. Embedded insurance and aggregators will not wait for carriers to catch up.
The 12-to-24-month pivot to the AI-insurer
Carriers do not need a five-year transformation programme. They need three overlapping horizons, executed in parallel and governed ruthlessly:
- Horizon 1 (0–6 months) — Foundations. Consolidate to a single cloud landing zone with FinOps guardrails. Stand up a governed data product layer for two or three insurance domains (policy, claims, customer). Roll out Copilot-class tooling at scale for knowledge work. Form two or three platform teams (data, integration, engineering productivity). Kill 30% of active projects to free funding and management attention.
- Horizon 2 (6–12 months) — Targeted AI in the core. Deploy AI / GenAI in four high-ROI use cases: claims triage and FNOL intake, SME underwriting triage, fraud and subrogation detection, and servicing / contact-centre copilots. Hold the organisation to measurable outcomes — 20–30% reduction in claims cycle time, 15% lift in underwriter throughput, 10–15 points of STP uplift on priority journeys.
- Horizon 3 (12–24 months) — Agentic and straight-through. Move from co-pilots to agentic workflows across the lowest-complexity 40–50% of claims and new business. Rewire the product operating model so AI capability ownership sits with product teams, not a central CoE. Publish AI-native journey KPIs alongside the combined ratio — or the transformation is not real.
The ambition is not to become a tech company. It is to become an insurance company that decides, prices, pays, and serves at machine speed — with human judgement reserved for where it actually moves the needle.
What this means for each of you
| CUO | Own the AI underwriting operating model. Decide which decisions AI makes, which it recommends, which it never touches. Treat this as a risk-appetite question, not an IT question. |
| CCO | Defend distribution economics. Fund digital quote-to-bind and embedded channels at parity with intermediated ones; stop subsidising structurally declining channels. |
| CIO | Stop delivering projects, start delivering platforms. Rebalance the book to at least 40% build within 12 months. Put DORA and AI Act compliance inside engineering standards, not beside them. |
| COO | Make STP a line-of-sight KPI. Every manual hand-off in claims and servicing is a future AI automation opportunity and a present cost centre — instrument and publish them. |
The stakes
The carriers that close this gap in the next 24 months will own the economics of the next decade. Those that do not will be acquired, partitioned, or quietly run off. The pivot is not optional; the timing is what remains to be decided.