Good evidence does not automatically inform policy. Drawing on practitioner experience, Frank Tu Ngo offers three hands-on tactics to help researchers make their work count: seize policy windows, engage policymakers early, and pressure-test messages with artificial intelligence
Let’s begin with a story.
Meet Sam: a motivated policy researcher who has spent years studying how governments can govern artificial intelligence while encouraging innovation. Sam has analysed regulations, investment data and evidence at the global, national and local levels. He has also considered questions of data security, equity and environmental impact. After many long days reviewing the literature and discussing his findings, Sam publishes an 8,000-word article in a leading policy journal.
Now meet Sandra: a mid-career policy lead in a Ministry for Digital Transformation. Her four-person team covers a rapidly changing AI portfolio. Every day, they search for data, draft policies, track KPIs and coordinate with politicians, researchers and technology companies. Their inboxes are filled with urgent requests and lengthy reports. Sandra has a public policy background, but is not a specialist in every aspect of AI. In a few years, she may rotate to another portfolio.
Sam, the researcher, and Sandra, the policymaker, could collaborate with each other. Yet researchers are trained to prioritise strong theory, sound methodology, and deep expertise. Policymakers, meanwhile, often need quick, decision-relevant insights across several subjects. The challenge is how to bring them together to support evidence-based decision-making.
Many excellent resources already explain the core principles of effective policy communication and engagement, e.g., know your audience, write clearly, and put your main argument first. This article builds on those principles, with three hands-on tactics drawn from practice.
'Never waste a good crisis,' one policymaker – whom I will call Adam – once told me with a laugh.
A crisis can create a narrow window in which decision-makers urgently need credible evidence. The aim is not to capitalise on suffering. It is to ensure that reliable evidence is ready when attention and demand suddenly increase. When an OpenAI agent gained unauthorised access to an Australian government website, for example, researchers with expertise in cybersecurity and public-sector data governance had an opportunity to help policymakers understand what had happened, and what safeguards might prevent it from happening again.
A crisis may be an extreme example. Policy windows also open around budget negotiations, elections, political scandals, congressional hearings, and major international events. The viral ‘Senator, I’m Singaporean’ exchange during TikTok CEO Shou Zi Chew’s 2024 Senate testimony, or the UN’s Digital Cooperation Day, can create opportunities to connect established research on platform governance, digital sovereignty or AI regulation with live policy discussions.
Practitioner’s tip
Researchers should map likely moments of demand in advance. Ask yourself: if this issue entered tomorrow’s headlines, how would you convincingly communicate your strongest finding within 24 hours?
Researchers often approach policymakers only after the final report is ready. Instead, consider sharing preliminary findings much earlier. Invite policymakers to a research seminar, working group or short interview. If you are teaching, invite them to share their experience with students. Many are generous and willing to help future generations. Ask for feedback on your preliminary findings by email, LinkedIn, or during an informal conversation.
These early forms of engagement are a way of sharing evidence. Even if a policymaker never reads the final report, they have already encountered its findings, questioned its assumptions and perhaps shared it with colleagues.
During my research in Japan, I interviewed a local politician and informally shared some observations and preliminary findings. The politician – whose policy agenda included promoting international exchange and digitalisation – then asked his secretary to share a photograph of us and several points from our discussion on his social media channels. This gave the research additional visibility and allowed the conversation to continue beyond the interview itself.
Practitioner’s tip
If policymakers participate in your events or interviews, ask whether they can connect you with their communications team. This can extend the reach of your work, enable their institution to reuse your materials, and draw on your expertise.
Here we go again – you may think. You probably do not need another article telling you that AI is changing our lives. But it can definitely help with the everyday work of connecting evidence and policy.
Researchers are close to their data. Deep expertise can sometimes make it difficult to recognise jargon or hidden assumptions. AI can offer a fresh perspective, helping researchers identify technical language and explore clearer ways of presenting their findings. I sometimes use it to test different visualisations of the same dataset; for example, to see which type of chart communicates an important comparison or trend most clearly. It can also turn the same findings into ready-to-use products, such as a one-pager, social media post or short-video script.
Three of my favourite AI prompts
1. Identify relevant policy windows
Identify forthcoming policy events, consultations and international observance days relevant to [my field] over the next 3–6 months. For each one, explain its relevance and provide an official source.
2. Test clarity
Revise this text for a time-pressed policymaker without specialist knowledge. Identify and explain technical terms. Do not change the meaning or add facts.
3. Create a professional communication product
I really like the style of this [one-pager or chart]. Using my evidence and analysis, create a [product] with a similar visual style. Make it clean and professional. Do not copy its branding or change my analysis.
Now, let’s return to the story we began with.
At 4pm, an issue lands on Sandra’s desk: she must brief the digital minister the next morning. This time, Sam is ready: a clear finding, a one-page visual and a short explanation her team can use. Sandra already knows Sam from an earlier exchange. She also remembers seeing his work during her brief daily scan of LinkedIn. So she calls him. By morning, his evidence is in the briefing. It may not determine the outcome. But it has made it into the room.
