Skip to main content

Prompt examples

Examples of how you can prompt the AI Assistant and the kinds of things you can ask it.

Written by Veronica Fletcher

1. Data-led

Built around an original dataset, survey, or analysis you're releasing - the data is the angle.

You: "Build a list for a study showing UK commuters spend an average of 11 days a year stuck in traffic, broken down by city. I want reporters who write up new data and transport trends."
​

Refine: "Drop anyone who only covers national politics - I want transport and motoring specialists."
​

Refine: "Now create a separate regional list: reporters at local outlets in the worst-affected cities, so we can localise the stat."
​

2. Reactive newsjacking

Pitches that ride on breaking news. The assistant leans on near-real-time news search to catch reporters already on the story.

You: "There's a major supermarket payment-system outage happening today. Build a list of reporters already covering it, so we can offer a retail-tech expert for comment."

Refine: "Focus on the ones who've posted in the last few hours, drop anyone whose coverage is older."

Refine: "Add consumer affairs reporters too, not just tech. This is a 'what shoppers should do' angle as well."

3. Expert commentary

A long-lived beat list around an expert or spokesperson - "here's someone who can comment on X."

You: "Build a list of journalists who cover personal finance and mortgages from an expert led perspective, for a spokesperson who's a chartered financial adviser. This is for ongoing commentary, not one story."

Refine: "Lean towards reporters who quote named experts in their pieces, rather than ones who only run data or company lines."

Refine: "Split this into broadsheet/national finance desks and consumer money sites, so I can tailor how I approach each."

Refine: "tag anyone who cover first-time buyers specifically. That's our spokesperson's strongest topic."

4. Local and regional

A story anchored to a place. Geography is the main filter.

You: "Build a list for the opening of a new community sports centre in Leeds. I want Leeds-local and West Yorkshire reporters who cover community and city news."

Refine: "Keep it strictly local - remove anyone at a national desk, even if they covered a Leeds story once."

Refine: "Add the local BBC and regional radio contacts, and any 'what's on' or community editors."

5. Product launch

A new product, funding round, partnership, or feature.

You: "Build a launch list for a B2B scheduling app aimed at hospitality businesses. I need hospitality-trade press."

Refine: "Add business reporters who cover funding rounds; we're announcing a seed raise alongside the launch. Add these to a separate segment."

Refine: "Drop anyone who mainly covers consumer apps - this is strictly B2B."

Other things you could ask for:

Beyond building and trimming, the assistant can do a few things people don't always think to ask:

  • Explain a match. "Why is this person on the list?" - it'll tell you what they cover and the article that matched. Good for sanity-checking before you pitch.

  • Find lookalikes. "Find me more reporters like the top five here." - seeds a search on the journalists you already rate.

  • De-risk a list. "Flag anyone who looks like a weak or borderline match." - surfaces the contacts most likely to be a stretch, so you can review them before sending.

  • Sense-check coverage. "Is there an obvious angle or type of outlet I'm missing here?" - useful when a list feels lopsided.

  • Tier for sending order. "Group these into a priority tier to pitch first and a wider tier for later." - turns one list into a sequenced rollout.

  • Pull context for the pitch. "What's the most recent relevant thing this reporter wrote?" - handy for a personalised opening line.

  • Spot duplicates and near-misses. "Are there two people here from the same desk I should pick between?" - avoids pitching the same newsroom twice.

Did this answer your question?