Home · Business · Marketing & Content · Paid ads
As of 13 August 2026, AI can only partly build a Meta lookalike audience for your business.
This still needs a person who signs their name to it.
Can you do it?
5 minutesto a draft.
30 minutesto something you’d act on.
Cost, all in£0
Skill neededpower-user
Who has to check ityou
What the alternative costsThe available tool data does not provide a price for a human or software alternative that builds a Meta lookalike audience.
If this goes wrong, you may target unsuitable people, breach data-use requirements or spend advertising money on a poor audience.
What to actually do
Hand it to a person
The route this page recommends
A person who owns the outcome does this end to end, worth it when the failure is dear.
Use a tool built for this
Second choiceDo it yourself
The distant thirdA chat interface, power-user skill, and roughly 30 minutes until you can act on the result.
How to actually do it
- Open Meta Ads Manager and record the business account, ad account, campaign objective, target UK location and any existing pixel or dataset setup.
- Choose the permitted source audience, such as customers or website visitors, and confirm its origin, date range, exclusions and lawful basis for using it for advertising.
- Remove raw personal data from the chatbot conversation and paste only the source-audience description and campaign details into the prompt.
- Ask the chatbot for the Meta Ads Manager route, settings, exclusions and pre-launch checks, and separate its instructions from anything it says you must confirm in Meta.
- Prepare the source audience in Meta Ads Manager, create the lookalike audience and compare the selected source, location, exclusions and other settings against the chatbot checklist before saving.
- Use the new audience in a controlled campaign, then compare delivery and results with a relevant audience while checking spend, reach and targeting against your campaign plan.
Prompt
Help me build a Meta lookalike audience for my UK business without accessing my advertising account or handling raw personal data. Use these details: business type: [BUSINESS TYPE]; campaign objective: [CAMPAIGN OBJECTIVE]; target country or region: [LOCATION]; source audience available: [SOURCE AUDIENCE, SUCH AS CUSTOMERS OR WEBSITE VISITORS]; source audience size and date range: [DETAILS]; exclusions: [EXCLUSIONS]; customer consent and lawful data-use position: [DETAILS]; existing Meta setup: [PIXEL, DATASET OR OTHER SETUP]. Give me: 1) the exact source audience and exclusions to prepare, 2) a step-by-step route through the current Meta Ads Manager interface, 3) the settings to choose, 4) a pre-launch checklist, and 5) a testing and monitoring plan. Do not invent interface labels, audience sizes, legal permissions or performance estimates. Clearly mark anything I must confirm in Meta. Flag any data-protection or consent issue and say when I should ask a UK data-protection professional. Keep raw customer data out of the response.
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
What it gets wrong
- AI cannot access your Meta account, upload the source audience or click the final creation controls for you.
- It cannot establish whether your customer data has the required permission or lawful basis for this advertising use.
- It cannot know whether your source audience is representative of the customers you actually want to acquire.
- It cannot verify audience quality before the campaign produces real delivery and performance data.
- It does not carry responsibility for privacy complaints, unsuitable targeting or wasted advertising spend.
What caps this at PARTLY: private data access, consent and privacy and verification cost.
How we scored this
Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.
| Axis | Score (0–2) |
|---|---|
| Output | 1 |
| Inputs | 1 |
| Verification | 1 |
| Liability | 1 |
| Effort delta | 1 |
| Total | 5 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can ChatGPT create a lookalike audience on Meta?
- No, it cannot create one inside your Meta account without an appropriate account connection and permission. It can prepare the source-audience plan, explain the settings and give you a checklist for creating it in Meta Ads Manager.
- What do I need to create a Meta lookalike audience?
- You need an eligible source audience, a target location, clear exclusions and the relevant Meta business and advertising account access. You also need to confirm that you can lawfully use the source data for this advertising purpose.
- Is it safe to use AI to build a Meta lookalike audience?
- It is safer to use AI for planning than to paste customer records into a chatbot. You remain responsible for consent, data protection, targeting choices and checking every setting in Meta before spending money.
- Can AI choose the best source audience for my Meta lookalike?
- AI can compare the source-audience descriptions and campaign objective that you provide. It cannot know whether the data represents your best customers, whether it is legally usable or whether the resulting audience will perform well without campaign evidence.
Nearby answers
- Can AI choose audience targeting for my paid ads?PARTLY
- Can AI create a retargeting campaign for my business?PARTLY
- Can AI create images for my social media ads?YES
- Can AI run A/B tests for my paid ads?PARTLY
- Can AI track the return on investment from my paid ads?PARTLY
- Can AI write landing page copy for my paid ads?YES
Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.
The newsletter
AI news, new answers and product picks, straight to your inbox.