Formal AI oversight at board level, but little evidence on rules and testing of AI products
Overall rating “limited”. No deduction for controversies. The superscript numbers link directly to the source.
1 · Responsibility
WTW documents formal AI oversight at board level and a new Chief AI Officer, but publishes neither its AI guidelines nor an approval procedure.
What is documented
- According to the 2026 proxy, the Risk and Operational Oversight Committee is explicitly responsible for AI risks and met four times in 20251
- Item 1C confirms the committee's responsibility and regular management reports on cybersecurity and AI2
- Since April 2026 there has been a Chief AI Officer and Head of AI Acceleration reporting to the CEO3
- The code of conduct requires compliance with internal AI guidelines; a hotline is also open to third parties4
- The 10-K names risks of inaccurate or biased AI outputs and the need for human oversight5
The case against
Oversight without documented management
- The AI guidelines are not public; risk classification, inventory and approval procedures for in-house and purchased AI are missing4
- No governance or risk mandate is described for the Chief AI Officer3
- AI training is mentioned without figures; external audits and AI complaints channels for those affected are missing7
Proceedings
- ERISA lawsuits over voluntary benefits name large benefits brokers, including WTW; WTW's role is unclear8
Sources for Responsibility (8)
- Willis Towers Watson Public Limited Company / U.S. SEC: Definitive Proxy Statement (DEF 14A) for the annual general meeting of 2026-05-20 · 20 May 2026 · primary source
- Willis Towers Watson Public Limited Company / U.S. SEC: 10-K 2025, Cybersecurity Risk Management, Strategy, and Governance (XBRL R11, Item 1C) · 25 February 2026 · primary source
- WTW: WTW announces new leadership roles to accelerate enterprise AI strategy and adoption · 27 April 2026 · secondary source
- WTW: WTW Code of Conduct «Transforming tomorrows with our values» (July 2025) · 1 July 2025 · secondary source
- Willis Towers Watson Public Limited Company / U.S. SEC: Form 10-K for the fiscal year ended 2025-12-31 · 25 February 2026 · primary source
- WTW Careers: Manager – Compliance AI & Automation (job advertisement, Thane, India) · 3 November 2026 · secondary source
- WTW: 2025 WTW Sustainability Report "Taking action to create sustainable value" · 31 December 2025 · secondary source
- JD Supra: Voluntary Benefits Are Not «Low-Risk»: A New Wave of ERISA Fiduciary Litigation · 3 November 2026 · secondary source
2 · Safety
In Item 1C WTW documents an incident response plan and security standards, but its ISO 27001 certification is only a company claim and testing of Violet, Radar 5 or Rewards AI is missing.
What is documented
- Item 1C describes a group-wide incident response plan with roles, severity levels and escalation; no material incidents are reported6
- Security standards cover access, encryption, network, secure development and third-party providers; annual mandatory training is in place6
- The CISO organisation reports to the Risk and Operational Oversight Committee6 11
- According to product information, the Embark assistant is limited to approved content and uses role-based access controls2
The case against
Missing evidence of testing
- Testing, validation and monitoring procedures for Violet, Radar 5, Rewards AI and Expert are not published
- The ISO 27001 certification is only claimed in the sustainability report; certificate and scope are missing7
- AI-specific security controls and a public vulnerability reporting channel were not found
Sources for Safety (11)
- Willis Towers Watson Public Limited Company / U.S. SEC: Form 10-K for the fiscal year ended 2025-12-31 · 25 February 2026 · primary source
- WTW: WTW enhances its EX software, Embark, to strengthen employee connection and drive decisions · 4 March 2026 · secondary source
- WTW: The power of AI in action (campaign page on Violet) · 3 November 2026 · secondary source
- WTW: WTW launch of Radar 5 with Gen AI capability marks a major milestone in insurance pricing and underwriting technology · 14 October 2025 · secondary source
- WTW: WTW adds real-time Python deployment and governance to Radar · 17 September 2024 · secondary source
- Willis Towers Watson Public Limited Company / U.S. SEC: 10-K 2025, Cybersecurity Risk Management, Strategy, and Governance (XBRL R11, Item 1C) · 25 February 2026 · primary source
- WTW: 2025 WTW Sustainability Report "Taking action to create sustainable value" · 31 December 2025 · secondary source
- WTW: Privacy Notice for Clients – Thailand · 27 April 2020 · secondary source
- board-cybersecurity.com: Willis Towers Watson PLC Cybersecurity Tracker · 3 November 2026 · secondary source
- Beyond Machines: Willis Towers Watson reports data breach via Pension Benefits MOVEit incident · 28 July 2023 · secondary source
- Willis Towers Watson Public Limited Company / U.S. SEC: Definitive Proxy Statement (DEF 14A) for the annual general meeting of 2026-05-20 · 20 May 2026 · primary source
3 · Transparency
WTW names AI use and risks in the 10-K only in general terms; whether plan participants learn about the AI and which models and data are used is not documented.
What is documented
The case against
What is not disclosed
- Whether plan participants using Violet learn that they are interacting with AI is not documented
- The privacy notice read contains no statements on AI or automated decisions6
- No model or system information, limits, error rates or training data for the AI products were found
Sources for Transparency (8)
- Willis Towers Watson Public Limited Company / U.S. SEC: Form 10-K for the fiscal year ended 2025-12-31 · 25 February 2026 · primary source
- WTW: WTW launch of Radar 5 with Gen AI capability marks a major milestone in insurance pricing and underwriting technology · 14 October 2025 · secondary source
- WTW via GlobeNewswire (Researchpool): WTW launches HR AI assistant, Expert · 11 February 2025 · secondary source
- WTW: WTW launches Rewards AI, providing compensation intelligence with generative AI · 2 February 2026 · secondary source
- WTW: Violet – AI suite for benefits (campaign page) · 3 November 2026 · secondary source
- WTW: Privacy Notice for Clients – Thailand · 27 April 2020 · secondary source
- WTW: Global Website Privacy Notice · 3 November 2026 · secondary source
- WTW: WTW enhances its EX software, Embark, to strengthen employee connection and drive decisions · 4 March 2026 · secondary source
4 · Why the harm rule does not apply
The harm rule does not apply: during the period no judgment, regulatory decision, fine or pattern of serious incidents involving its own AI systems was found.
5 · Controversies (0)
No AI controversy found.
6 · What remains open
- Content of the internal AI guidelines and the Chief AI Officer's mandate
- Testing, validation and monitoring procedures of the AI products for plan participants and insurers
- Labelling of AI interactions in Violet and Embark
- Full text of Items 1A and 3 of the 2025 10-K, which was truncated on retrieval
7 · How it was assessed
Cut-off date: 3 November 2026 · 40 searches, 31 fetches.
Primary documents read (9)
- Form 10-K Fiscal Year 2025 (Cover Page, Item 1, Part of Item 1A; retrieved in two parts)
- 10-K 2025, Item 1C Cybersecurity (XBRL R11, two parts)
- 10-K 2025, Note Commitments and Contingencies (XBRL R26)
- Form 10-Q as at 2026-09-30, Note Commitments and Contingencies (XBRL R23)
- Definitive Proxy Statement 2026 (two parts)
- Form 8-K relating to the AGMs 2024 and 2025
- 2025 WTW Sustainability Report (two parts)
- WTW Code of Conduct (July 2025)
- Privacy Notice for Clients – Thailand
Research of 3 November 2026 under Globalance AI Footprint rulebook version 0