As of 13 August 2026, AI can only partly choose audience targeting for your paid ads.

What to actually do

1. Use a tool built for this

   The route this page recommends

2. Do it yourself

   Second choice

   A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open the advertising platform's campaign and audience documentation, then note which targeting controls it currently supports for your chosen campaign objective.
   2. Gather your offer, landing page, UK locations, customer description, exclusions, budget, campaign dates and anonymised evidence from previous campaigns or customer records.
   3. Paste that information into the prompt, including the exact advertising platform and objective, and ask for ranked audience hypotheses rather than one untested answer.
   4. Compare every proposed targeting control with the platform documentation and remove settings the platform does not offer or that conflict with its policies.
   5. Build the recommended broad and narrow audiences in the platform without granting the chatbot account access, then compare the platform's estimated reach and warnings with the drafted assumptions.
   6. Launch only the agreed test, record the audience settings and budget, and compare results against the stated objective before changing the targeting.

   Prompt

   Act as a paid advertising strategist, not as the person who launches the campaign. Based only on the information below, propose audience targeting for a UK paid advertising campaign.
   
   Business and offer: [describe the business, product or service]
   Campaign objective: [sales, leads, app installs, awareness or other]
   Landing page or destination: [paste the URL or describe it]
   Ideal customers: [describe who buys, including relevant locations, needs, buying triggers and exclusions]
   Existing customer evidence: [paste anonymised customer segments, order data, CRM summaries or campaign results]
   Advertising platform: [Meta, Google Ads, LinkedIn, TikTok or other]
   Budget and dates: [state the budget and campaign period, without sharing account credentials]
   Legal, brand or platform constraints: [state any restrictions]
   
   Return:
   1. Three audience hypotheses, ranked by the evidence supporting them.
   2. For each hypothesis, give the exact audience signals, interests, keywords, job attributes, locations, age or other settings that the named platform supports, and say when a setting is unavailable rather than inventing one.
   3. Recommended exclusions and the reason for each.
   4. A broad-audience option and a narrower option, with the risk of each.
   5. A simple test plan showing what should be held constant and what should change.
   6. A list of claims or assumptions I must verify in the advertising platform before launch.
   
   Do not claim that an audience will perform well. Do not use sensitive personal data or suggest targeting people based on protected characteristics, health, financial distress or other sensitive circumstances. Flag any recommendation that needs specialist legal, privacy or platform-policy review. Separate evidence from assumptions and ask concise questions where the information is insufficient.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

3. Hand it to a person

   The distant third

   A person who owns the outcome does this end to end, worth it when the failure is dear.