Treasury Chief: ‘AI Can Relieve Staff of Lower-Value Work’

Why It Matters

AI adoption promises to boost Treasury efficiency, reduce staffing pressures and foster more transparent, data‑driven collaboration across government, reshaping public‑sector productivity.

Key Takeaways

• Treasury testing AI tools with 25% staff.
• First UK department with HR chatbot.
• AI handles correspondence and sanctions casework.
• Project Reset seeks real‑time inter‑department data.
• AI could sustain output with smaller workforce.

Pulse Analysis

The UK Treasury is accelerating its digital transformation by embedding artificial intelligence across both operational and policy functions. Senior civil servant Beth Russell disclosed that the department has already deployed an internal HR chatbot and is piloting Microsoft Copilot with a quarter of its staff, alongside a bespoke HMT‑GPT model. Early applications focus on routine tasks such as drafting correspondence, processing financial‑sanctions casework and managing internal queries, freeing civil servants to concentrate on higher‑value analysis. This rollout positions Treasury as one of the first central government bodies to operationalise generative AI at scale.

By automating low‑value work, the Treasury aims to sustain policy output while trimming headcount, a claim Russell linked to a “slightly smaller workforce” delivering the same quality of advice. AI‑driven drafting and data‑analysis tools can accelerate research, synthesize consultation responses and generate briefing notes, potentially shortening the policy cycle and improving evidence‑based decision‑making. The initiative also promises cost savings through reduced manual processing and faster turnaround on financial‑sanctions enforcement, aligning with broader public‑sector efficiency targets. For private‑sector stakeholders, the move signals a shift toward more data‑rich, transparent fiscal policy formulation.

Deploying generative AI in a government setting raises governance, security and bias concerns that the Treasury must address through robust oversight frameworks. Ensuring data confidentiality, especially when handling sensitive sanctions information, requires strict access controls and audit trails. Moreover, the “Project Reset” agenda to obtain real‑time information from other departments will depend on interoperable IT systems and clear incentive structures to avoid new bottlenecks. If managed effectively, Treasury’s AI experiment could become a blueprint for other ministries, accelerating the UK’s broader digital‑government agenda and reshaping public‑sector productivity.