Summary

• Middle East conflict produced unprecedented use of the FCDO’s public-facing digital services.
• The department reports regular Copilot Chat use among 16,000 staff and is developing internal policy agents.
• Crisis resilience, classification, data quality, and accountable human judgement remain central to the deployment.

Conflict in the Middle East has subjected the Foreign Office’s digital services to unusually heavy demand while the department expands generative AI and automation across diplomatic, policy, and consular work.

The Foreign, Commonwealth and Development Office says renewed regional conflict during the spring produced unprecedented use of its public-facing services for travel information, crisis guidance, and consular support.

Its annual report also says 16,000 employees regularly use Microsoft Copilot Chat, supported by instructor-led training, practical assistance, an AI summer school, and department-wide learning. The FCDO describes productivity gains, although it has not published a detailed independent evaluation of time saved, rework, or output quality.

An internal AI Catalyst team is developing First Drafts, which produces analytical policy material aligned with departmental standards, and Submit.ai, which checks ministerial submissions for clarity, completeness, and compliance.

Crisis traffic exposes the whole service

Public services are often assessed against normal demand even though their most consequential use may arrive during a sudden surge. Conflict, evacuation, natural disaster, or border closure can send large numbers of people to the same guidance and contact channels within hours.

During those periods, stale information, slow publishing, inaccessible pages, or disconnected case management can produce direct harm. Citizens need reliable guidance about government advice, available routes, registration, and emergency support.

The FCDO says it has improved public-facing services and its internal case management platform while updating guidance on GOV.UK. Its 2025–26 annual report outlines plans for better direct communication, simpler transactions, wider government data sharing, and further investment in AI and automation.

Resilience extends beyond page views because systems must support secure identity, case management, multilingual content, contact centre integration, and publishing when overseas posts, networks, or staff are disrupted.

Operational teams also need a common view of rapidly changing information. Advice published online, guidance used by call centre staff, and instructions followed by overseas posts must remain aligned as diplomatic and security conditions change.

AI is moving closer to policy production

Copilot Chat provides broad access across the workforce, while First Drafts and Submit.ai move automation nearer to formal policy and ministerial processes. Drafting and checking can reduce repetitive work, but those systems operate where nuance, source reliability, classification, and diplomatic consequence carry unusual weight.

A tool producing a first draft can influence which evidence is included, how uncertainty is presented, and which options appear plausible. Human review must therefore cover sources and reasoning rather than concentrating on grammar and formatting.

Submit.ai presents a different model by checking work against expected standards. Such a system may identify missing sections or inconsistent presentation, yet compliance with a template does not establish that the policy is well supported or operationally sound.

Security boundaries are equally important because diplomatic work ranges from public material to highly classified intelligence. General AI tools, internal agents, and the department’s Official-level IT environment need clear rules governing which data can be processed, where logs are held, who can access outputs, and how incidents are investigated.

Core systems carry the AI layer

The annual report describes a product-centred operating model built around clearer prioritisation, faster delivery cycles, and closer work between digital, policy, operational, and corporate teams. Continuous improvement suits consular technology because demand and geopolitical conditions do not wait for periodic transformation programmes.

Persistent services also require stable multidisciplinary teams, accountable ownership, operational metrics, and funding for maintenance after launch. A succession of short projects can leave public systems with features but without the capacity to keep them dependable.

The worldwide deployment of Osprey, the department’s standard Official-level platform, and the introduction of Nexus for security management provide the less visible foundation beneath the AI programme. Agents remain dependent on identity, access, networks, data, and core workflows operating reliably.

Public reporting will need more detail as automated tools enter consequential work. Adoption figures indicate reach but not accuracy, time saved, correction, user confidence, or the proportion of AI suggestions that staff reject.

Evidence of productivity should also include the cost of review. A quick draft creates little value if experienced officials spend longer tracing claims or correcting context, particularly when policy material can influence ministerial decisions and international relationships.

The surge in consular demand shows how digital infrastructure has become part of a foreign ministry’s operational capacity rather than an administrative support function. AI is being introduced into the same environment, where resilience, security, and accountable judgement will determine whether the systems strengthen diplomatic and public service delivery.