Almost every government talks about AI. (Good!)
But hardly any measures its success. (And that’s not good!)

Progress is always about making things better. In digital government, that means: simpler, faster, more accessible, more cost-efficient. So why do hardly any governments measure what AI actually delivers?

According to the new OECD Digital Government Outlook 2026, almost all OECD states now use AI in at least one area of government. But only 28 percent systematically measure the financial or non-financial impact of these AI applications. That’s remarkable: nowhere in the economy would anyone make gigantic investments without evaluating their success. Change always needs KPIs — especially for AI applications too:

Is an application processed faster?
Does the number of follow-up queries drop?
Does the quality of decisions improve?
Or does it, in the end, just bolt an extra layer of technology onto existing bureaucracy?


Image: This image was created with AI.

Source:
OECD, Digital Government Outlook 2026, chapter “Adopting and governing AI in government”: only 28 percent of OECD countries systematically measure the financial or non-financial impact of their AI use in government.

You can find this post on LinkedIn here: https://www.linkedin.com/posts/janwegener_digitalisierung-verwaltung-oecd-activity-7477639586693861376-OHiK?utm_source=share&utm_medium=member_desktop&rcm=ACoAAF-yDsQBgbICDl3wlRGoCQSZHiHxHemNJR8