Insights / Tools
The "ask the assistants" prompt pack
15 copy-paste prompts to run through ChatGPT, Gemini and Perplexity. Replace the [brackets] with your details, run them monthly, keep notes. What gets measured gets fixed.
Recommendation tests
The bottom line: when a stranger asks for what you sell, are you the answer?
Who are the best [your service] providers in [your city]?
I need help with [the problem you solve]. Who should I contact in [your city]?
Recommend a [your service] business for a small company in [your city]. Explain why.
What should I look for when choosing a [your service] provider, and who does it well in [your region]?
Accuracy tests
When AI talks about you directly, does it get the facts right?
What is [your business name]?
What services does [your business name] offer?
Where is [your business name] located and how do I contact them?
Is [your business name] reputable? What do reviews say?
Competitor comparisons
Learn why the model prefers whoever it prefers — that's your roadmap.
Compare [your business name] and [competitor] for [service]. Which would you recommend and why?
What are the strengths and weaknesses of [competitor]?
Who are the main competitors of [your business name], and how do they differ?
Source-tracing tests
Find out WHERE the model gets its information — those sources are your citation targets. Perplexity shows sources by default; ask the others to cite.
Who are the best [your service] providers in [your city]? Cite your sources.
What sources would you check to evaluate a [your service] business in [your city]?
Summarise what the web says about [your business name]. List the sites you drew from.
What questions do people commonly ask about [your service] in [your city]?
Reading the results
Score each prompt: named first (3), named at all (2), described accurately when asked directly (1), absent or wrong (0). Under 25 total across the pack? Read the AEO guide — or let us run the full audit.