Policy interview prep when AI summarizes the evidence

Check the evidence, compare who benefits and show how your recommendation would work.

What companies say they’re looking for

01 Check what the summary counted

“When the themes were finalised, Consult counted the number of responses that mentioned each theme and reported the results for officials to analyse.” (opens the source)

Department for Science, Innovation and Technology, UK consultation methodology · updated 19 August 2026 · GOV.UK, Jul 15, 2026

In this UK consultation, officials reviewed the themes before using the AI counts. Check which submissions and channels the draft actually covers.

Local policy analyst interviewWhat can the library tell its board about support?Example

02 Compare who benefits

“Assess the economic, competition, social, environmental or other impacts and how those impacts are likely to be distributed.” (opens the source)

Office of Impact Analysis, Australian Government impact-analysis guidance · Office of Impact Analysis, Jul 1, 2026

Australia’s guidance asks analysts to show who each option affects. Compare overall reach with the goal for households in the limited-service area.

Policy adviser interviewWhich transport option reaches the intended households?Example

03 Check what the teams can deliver

“The paper sets out key activities and timeframes for implementation and delivery” (opens the source)

The Policy Project, Policy Quality Framework · updated 5 July 2026 · New Zealand Department of the Prime Minister and Cabinet, Jul 30, 2025

New Zealand’s policy advice framework asks for delivery activities and timeframes. Check what both teams can complete and who owns the remaining work.

Policy implementation adviser interviewWhat first phase can the teams commit to?Example

04 Test the comparison

“Several assumptions need to hold for this comparison to credibly estimate program impact.” (opens the source)

Michala Riis-Vestergaard, Research and Training Lead, J-PAL Global · J-PAL, Jun 17, 2026

Riis-Vestergaard explains why evaluation depends on its assumptions. Check the results missing from the comparison group before advising on expansion.

Evaluation adviser interviewCan the reading pilot support expansion?Example

05 Check the cost per result

“Leaders must weigh costs against impact often under pressure, with imperfect data.” (opens the source)

Kayla Hoyer, Economist II and Cost-Effectiveness Unit Lead, IDinsight · IDinsight, Mar 27, 2026

Hoyer asks decision makers to weigh cost against impact. In the case, check the added cost per forecast renewal after the full support cost comes in.

Policy analyst interviewDoes either option still meet the cost target?Example

Sources

Department for Science, Innovation and Technology, UK consultation methodology · updated 19 August 2026, at the time. GOV.UK, “Summary of evidence, methodology, and organisations who responded to the consultation: July 2026”, Jul 15, 2026. Open the source. Checked against the published text.
Office of Impact Analysis, Australian Government impact-analysis guidance, at the time. Office of Impact Analysis, “The 7 Impact Analysis questions”, Jul 1, 2026. Open the source. Checked against the published text.
The Policy Project, Policy Quality Framework · updated 5 July 2026, at the time. New Zealand Department of the Prime Minister and Cabinet, “Policy Quality Framework”, Jul 30, 2025. Open the source. Checked against the published text.
Michala Riis-Vestergaard, Research and Training Lead, J-PAL Global, at the time. J-PAL, “Test, learn, adapt: Impact evaluations are only as strong as their assumptions”, Jun 17, 2026. Open the source. Checked against the published text.
Kayla Hoyer, Economist II and Cost-Effectiveness Unit Lead, IDinsight, at the time. IDinsight, “Turning cost data into a tool for impact”, Mar 27, 2026. Open the source. Checked against the published text.

Last reviewed Nov 11, 2026