Before AI can deliver value at scale, public-sector organizations need fit-for-purpose data, clear ownership and an operating model that allows teams to innovate within shared guardrails.
Fartash Haghani, City of Toronto’s Director of Enterprise Data and AI emphasizes that AI success should be measured through resident outcomes—not model accuracy alone. Residents care whether they receive an understandable answer, wait less time and get their problem resolved. If a more accurate model does not improve that experience, it has not created meaningful value.
Trusted data also requires more than accuracy. Governments must understand why information was collected, who owns it, how it is used and where it originated. Rather than slowing innovation, governance by design can help organizations move faster by establishing reusable controls based on risk. These enterprise foundations make more proactive, connected and resident-centred services possible.
Help your peers
Share what you've learned with fellow public servants