The OECD.AI Policy Navigator

Our policy navigator is a living repository from more than 80 jurisdictions and organisations. Use the filters to browse initiatives and find what you are looking for.

Public Service Commission (PSC) AI Agent for Travel


Added by:   OECD analyst
Added on:   23 Sep 2026
Updated by:   OECD analyst
Updated on:   23 Sep 2026

The AI Agent for Travel is an artificial intelligence assistant that gives public service employees quick, clear answers about travel rules and expense claims. It was created by government staff to reduce the large number of emails, delays, and confusion caused by complex travel policies. The assistant provides answers in seconds, saves staff time, and lowers administrative costs.

Initiative overview

The AI Agent for Travel was launched to address a persistent operational challenge across the public service: navigating federal travel and expense policies is complex, time-consuming, and heavily reliant on manual email support. Employees often wait days for answers to routine questions, leading to delayed approvals, last-minute bookings, inconsistent guidance, and avoidable costs. At the same time, departments face increasing pressure to modernise internal services, improve employee experience, and strengthen financial stewardship.

This initiative applies artificial intelligence in a practical and responsible way to reduce administrative burden while improving compliance and service quality. The objective of the AI Agent is to provide employees, managers, and finance teams with clear, consistent, and real-time guidance on government travel rules and expense claims.

Developed entirely in-house using existing Microsoft 365 tools within a secure Government of Canada environment, the solution relies exclusively on validated policy documents and approved procedures. It guides users step by step, reducing reliance on email inquiries and manual searches.

Since implementation, travel-related email volume has decreased by approximately 90%, response times have improved from several days to seconds, and the organisation is realising an estimated $120,000 in annual savings. Additional indirect savings are achieved through earlier travel planning and reduced last-minute booking costs.

Beyond operational efficiencies, the initiative strengthens financial oversight and supports broader transformation objectives. Financial officers are now able to focus on higher-value analysis and risk management rather than responding to repetitive policy questions. The solution operates with clear governance, documentation, and human oversight to ensure responsible and ethical AI use, with all data remaining within secure government systems.

By developing the capability internally, the organisation has strengthened its digital expertise while avoiding external consulting costs. The initiative has evolved from a targeted travel support solution into a scalable model for responsible AI adoption. Interest from over 40 federal departments and agencies demonstrates its transferability across government, and additional corporate service areas, including human resources and procurement, are beginning to adopt similar agents using the same framework.

As governance, training, and knowledge management processes continue to mature, the AI Agent provides a sustainable and replicable pathway for modernising internal government services at scale. The initiative has also received international recognition, including coverage by Global Government Finance in the article “Canada government agency AI tool tackles ‘bottleneck’ of travel and expense inquiries,” extending its influence beyond the Canadian public service.

Other relevant details

Results, outcomes and impacts: Since deployment, the AI Agent for Travel has reduced travel-related email inquiries by approximately 90%, cut response times from several days to seconds, and generated an estimated $120,000 in annual savings by avoiding additional staffing and enabling earlier bookings. These results were measured through email volume tracking, response time comparisons, and workload analysis. Financial officers now focus on higher-value tasks, while employees receive faster, consistent guidance. Adoption across more than 40 departments and additional service areas is expected to further increase efficiency, cost savings, and compliance. Challenges and lessons learned: Key challenges included ensuring guidance was fully policy-aligned, maintaining data security within the Government of Canada environment, and balancing automation with human oversight. Risks such as inconsistent outputs, user mistrust, and resistance to change were mitigated through rigorous testing, using only validated policy sources, and providing clear governance and support. Lessons learned include the importance of in-house development to build internal expertise, iterative user testing to improve adoption, and strong leadership support to drive cultural change. Success requires secure technology, clear policies, stakeholder engagement, and ongoing training to sustain responsible AI use.

About the policy initiative