Insurance 2026: AI and Third-Party Risk Share One Foundation - Trusted Data
27 January 2026
The UK insurance sector heads into 2026 with a simple brief: move faster on innovation and get tougher on external exposure. New Dun & Bradstreet research shows AI adoption and third-party risk reduction sit at the top of the list, and both rely on the same thing - clean, governed, connected data. Without that, AI misses the mark and risk strategies lose precision.
AI and Third-Party Risk: The Twin Priorities
AI remains a headline focus, with 37% of insurance firms prioritising internal adoption in 2026. Third-party risk reduction is just as prominent at 37%. Market expansion (31%) and embedded finance with better decisioning (26%) follow behind.
Intent isn't the issue. Execution is. Over half of firms (55%) report AI projects failing due to poor data quality, and only 17% say they run unified data platforms. If AI is to deliver real value, insurers must raise the bar on data accuracy, governance, and integration.
Third-Party Risk Is Broader - and Closer to the Core
Fraud, cyberattacks, and supply chain weaknesses are intensifying, made worse by fragmented data and legacy systems. Regulation is adding fresh pressure to prove governance and control, with the EU AI Act and DORA setting firm expectations for oversight and resilience.
It fits the mood: 76% of respondents cite operational risk - including third-party and supply chain exposure - as a top concern. The advantage now comes from the ability to integrate, verify, and act on near real-time data.
This approach also supports practical AI - see the AI for Operations.
Industry View
Zulf Raja, Head of Insurance at Dun & Bradstreet, notes that the market is balancing innovation and risk management. Firms are advancing AI and digital change with a sharper focus on third-party risk, data integrity, and economic uncertainty. The winners will connect growth plans with resilience.
What To Do Next
- Define a single data model and shared glossary across underwriting, claims, finance, and risk.
- Stand up a unified data platform or fabric for core sources; retire duplicate feeds and manual workarounds.
- Install a data quality programme: profiling, lineage, issue management, and controls tied to SLAs and business KPIs.
- Operationalise AI governance: model inventory, documentation, testing, monitoring, and clear accountability mapped to AI risk classes.
- Strengthen third-party risk management: tier suppliers, screen at onboarding (KYB, sanctions, adverse media), and monitor continuously.
- Contract for external data that matters: fraud signals, cyber posture, and supply chain dependencies with defined refresh rates.
- Deploy fraud and cyber analytics with labelled training data, drift checks, and human-in-the-loop review for high-impact decisions.
- Exercise contingency plans under DORA: scenario tests, supplier failure simulations, and communication playbooks.
- Formalise stewardship and RACI across data domains; align incentives so teams fix data at the source.
- Upskill teams in data literacy, AI operations, and TPRM to close execution gaps.
For regulatory-focused AI governance and oversight, see the AI Learning Path for Regulatory Affairs Specialists.
Build Capability Faster
If you're developing AI, data, or risk skills across underwriting, claims, or enterprise risk, explore practical programmes here: AI courses by job and AI certification for data analysis.
For governance leadership in AI adoption, see the AI Learning Path for CIOs.
Explore the full findings in the Dun & Bradstreet report.
Image credit: Canva
Your membership also unlocks:
