Search across all content
Recent pilots show that fewer than half of public servants who report using AI apply it in their day-to-day work.1 We call this the “application gap”.
This AI Best Practices Framework is a guide to the approaches and practical steps designed to help close that gap through the "3 Es of AI": demonstrating how to explore AI, run experiments and embed what works, based on real examples from across the Apolitical network.
Click on each best practice below to read a case study, hear from experts and share your experience in the poll.
The Government AI Campus brings together tools, courses and communities to support your AI journey:
Take the AI Readiness Check for a personalised assessment of your AI archetype.
Take a course on AI fundamentals for public servants to understand how AI works, where it helps, the risks to watch for and how human-centred approaches support safer, more effective use.
Follow selected experts from Apolitical’s AI Government 100 list to stay up to date with how governments are exploring AI in practice.
Join the Apolitical AI in Government community to discover how public servants around the world are using AI and what they’re learning.
Check your organisation’s guidance on generative AI use. Resources like Canada’s Guide on the Use of Generative AI, the UK Government's AI Playbook and the United States' Generative AI Playbook show the kinds of principles, safeguards and decision steps to look for in your own context.
Run a quick team survey: “What’s one task that feels clunky, slow or frustrating?” Capture the responses in a shared ‘problem backlog’ where anyone can log recurring issues or bottlenecks.
Build a task-specific AI assistant so you can offload repetitive tasks, like rewriting jargon-heavy emails into plain English, saving time while keeping outputs consistent and tailored to your context.
Take a course to learn the basics of prompt engineering so you can guide AI to produce useful, accurate and responsible outputs.
Use Apolitical’s AI Prompts for Public Servants cheat sheet to practise better prompting on real tasks.
When measuring impact, use guidance like the UK Government’s guidance on the impact evaluation of AI interventions to set objectives, build a theory of change, establish a baseline and evaluate flexibly as tools evolve.
Take a course to become a data confident public servant, so you can collect, interpret and use data effectively with AI.
Join the Apolitical Data Skills Network to explore how other public servants safely use data and AI.
Return to the explore phase by asking a question in the Apolitical AI in Government community about how other public servants are scaling what works with AI in their workflows.
Government Digital Service (GDS), Microsoft 365 Copilot Experiment: Cross-Government Findings Report (London: GOV.UK, June 2, 2025), https://www.gov.uk/government/publications/microsoft-365-copilot-experiment-cross-government-findings-report.
Meritalk, “Civilian Agencies See AI Productivity Gains, Still Struggle With Scaling Pilots,” Meritalk, September 16, 2025, https://www.meritalk.com/articles/civilian-agencies-see-ai-productivity-gains-still-struggle-with-scaling-pilots/.
Skip Descant, “ San Jose, California, Develops AI Upskilling Training for Staff”, Government Technology, July 17, 2025, https://www.govtech.com/artificial-intelligence/san-jose-calif-develops-ai-upskilling-training-for-staff.
Skip Descant, “ San Jose, California, Develops AI Upskilling Training for Staff”, Government Technology, July 17, 2025, https://www.govtech.com/artificial-intelligence/san-jose-calif-develops-ai-upskilling-training-for-staff.
Skip Descant, “ San Jose, California, Develops AI Upskilling Training for Staff”, Government Technology, July 17, 2025, https://www.govtech.com/artificial-intelligence/san-jose-calif-develops-ai-upskilling-training-for-staff.
Brian Simms, “Government Employees Need AI Education Now More Than Ever,” Graduate School USA, 2025, https://www.graduateschool.edu/blog/government-employees-need-ai-education-now-more-than-ever.
Erik Brynjolfsson, Professor at Stanford Institute for Human-Centred AI, US, Apolitical interview, 2025.
UK Local Government Association, “Artificial Intelligence Case Study Bank,” Local Government Association, Accessed November 14, 2025, .
UK Local Government Association, “Artificial Intelligence Case Study Bank,” Local Government Association, Accessed November 14, 2025, https://www.local.gov.uk/our-support/cyber-digital-and-technology/artificial-intelligence-hub/artificial-intelligence-case.
Ott Velsberg, Government Chief Data Officer of Estonia, Apolitical interview, 2025.
Maggie Johnson and Robyn Scott, “Building Skilled Governments: Unlocking AI's Promise for People Worldwide,” Google Public Policy, July 24, 2025, https://publicpolicy.google/article/apolitical-building-skilled-governments/.
James Lowe, “Redbox Reflections: 5 Key Lessons from Building (and Sunsetting) Our Government AI Chatbot,” AI.gov.uk, 23 October 2025, https://ai.gov.uk/blogs/redbox-reflections-5-key-lessons-from-building-and-sunsetting-our-government-ai-chatbot/.
James Lowe, “Redbox Reflections: 5 Key Lessons from Building (and Sunsetting) Our Government AI Chatbot,” AI.gov.uk, 23 October 2025, https://ai.gov.uk/blogs/redbox-reflections-5-key-lessons-from-building-and-sunsetting-our-government-ai-chatbot/.
James Lowe, “Redbox Reflections: 5 Key Lessons from Building (and Sunsetting) Our Government AI Chatbot,” AI.gov.uk, 23 October 2025, https://ai.gov.uk/blogs/redbox-reflections-5-key-lessons-from-building-and-sunsetting-our-government-ai-chatbot/.
Álvaro Vargas Franzani, Change Management Manager, Department of Justice, Canada, “AI and Public Service: Evolving Together”, Apolitical, December 2, 2024, https://apolitical.co/solution-articles/en/ai-and-public-service-evolving-together-165.
Robyn Scott, Co-founder and CEO, Apolitical, Apolitical interview, 2025.
Australian Government, Department of Finance, “GovAI: Enabling AI Use Across the APS,” accessed 2025, https://www.govai.gov.au/about.
Francesco Petracca, Lecturer of AI and Digital Innovation, SDA Bocconi School of Management, Apolitical interview, 2025.
Iwin Nayak, Senior Manager, Cabinet Office, Government of Canada, “How to Experiment with AI in Government”, Apolitical event, August 27, 2025, https://apolitical.co/events/aug-2025-how-to-experiment-safely-with-ai-in-government.
Waldo Jaquith, Government Delivery Manager, U.S. Digital Response, describing a colleague’s experience, Apolitical interview, 2025.
Ott Velsberg, Government Chief Data Officer of Estonia, Apolitical interview, 2025.
Peter Clutton-Brock, Executive Director, Centre for AI and Climate, UK, interview for Apolitical, 2025.
Global Government Forum, “Australian government workers save an hour a day in large generative AI trial,” Global Government Forum, November 20 2024, https://www.globalgovernmentforum.com/australian-government-workers-save-an-hour-a-day-in-large-generative-ai-trial/.
Global Government Forum, “Australian government workers save an hour a day in large generative AI trial,” Global Government Forum, November 20 2024, https://www.globalgovernmentforum.com/australian-government-workers-save-an-hour-a-day-in-large-generative-ai-trial/.
Office of Information Technology, State of Colorado, “Case Study: Google Gemini Pilot”, 2024, https://oit.colorado.gov/standards-policies-guides/guide-to-artificial-intelligence/case-study-google-gemini-pilot.
Dr T. Alexander Puutio, Stanford Professor, AI & Leadership Expert, Apolitical interview, 2025.
David Edinger, “NASCIO 2025: Q&A: Colorado CIO David Edinger Strives for the Right Mix of Solutions,” interview by Mickey McCarter, StateTech Magazine, October 13, 2025, https://statetechmagazine.com/article/2025/10/nascio-2025-qa-colorado-cio-david-edinger-strives-right-mix-solutions.
Mike Gifford, Senior Strategist, CivicActions, Canada, Apolitical community post, 2025, https://apolitical.co/en/communities/ai-in-government/eebe2450-97c7-11f0-8080-800160ff5c18.
Oxford Institute of Technology and Justice, “Brazil | Global Leader in AI Adoption in Criminal Justice,” Tech & Justice, Blavatnik School of Government, University of Oxford, Accessed November 15 2025, https://www.techandjustice.bsg.ox.ac.uk/research/brazil.
Oxford Institute of Technology and Justice, “Brazil | Global Leader in AI Adoption in Criminal Justice,” Tech & Justice, Blavatnik School of Government, University of Oxford, Accessed November 15 2025, https://www.techandjustice.bsg.ox.ac.uk/research/brazil.
Alex Uchoas, Deputy Superintendent, National Statistics Office, “Public Service and AI,” Apolitical, February 7, 2025, https://apolitical.co/solution-articles/en/public-service-and-ai-579.
Taki Sarantakis, President, National School of Public Service, Canada, Apolitical interview, 2025.
TechUK, “Case Study: Responsible AI for Government Decision Making,” TechUK, May 2023, https://www.techuk.org/resource/case-study-responsible-ai-for-government-decision-making.html.
TechUK, “Case Study: Responsible AI for Government Decision Making,” TechUK, May 2023, https://www.techuk.org/resource/case-study-responsible-ai-for-government-decision-making.html.
TechUK, “Case Study: Responsible AI for Government Decision Making,” TechUK, May 2023, https://www.techuk.org/resource/case-study-responsible-ai-for-government-decision-making.html.
Statistics Canada, “Responsible Use of Automated Decision Systems in the Federal Government,” Government of Canada, December 2021, https://www.statcan.gc.ca/en/data-science/network/automated-systems.
Taki Sarantakis, President, National School of Public Service, Canada, Apolitical interview, 2025.
Himansu Pandey, Innovation Lead Capacity Building Commission, India, Apolitical interview, 2025.
