Help us prepare for our appointment
Meta is taking a step forward in AI-powered ad targeting. Starting on 16 December 2025, Meta will use text, voice, and visual exchanges with Meta AI to refine the personalisation of ads and content on Facebook, Instagram, and the group’s other platforms.
Until now, targeting has been based on explicit signals — clicks, likes, subscriptions — or implicit signals, such as time spent and browsing history. Now, every conversation with Meta AI becomes behavioural data. This development illustrates the group’s desire to move towards more predictive and contextual marketing.
When a user asks Meta AI for “travel ideas” or a “vegetarian recipe”, these exchanges become signals of interest. Meta will merge this information with existing engagement data, such as likes or subscriptions. This will enrich advertising profiles and enable better anticipation of needs.
This approach goes beyond advertising. Conversations will also influence suggested Reels, recommended groups, and the news feed. For example, a search for curry could generate cooking videos, recipes, or advertisements for utensils. Thus, personalisation extends to the entire user experience.
In reality, Meta pushes behavioural targeting to its limits. Expressed intent becomes an exploitable marketing resource, rather than just an indirect signal.
Meta says it will not exploit conversations on sensitive topics such as health, religion, sexual orientation, political opinions or ethnicity. For the time being, this development will not affect Europe, the United Kingdom or South Korea, due to local privacy regulations.
Only users who interact with Meta AI will be affected. However, there is no option to disable it. Those who use the assistant will therefore have to accept that their exchanges will be used to personalise adverts and content.
For advertisers, this change opens up new opportunities, but also new challenges. Targeting becomes more precise, based on immediate signals of intent. This can improve conversions and enhance the relevance of campaigns.
However, this power comes with increased dependence on Meta’s algorithms. Manual control over audiences is diminishing, and transparency is declining. Furthermore, overly personalised advertising could be perceived as intrusive, especially in a context of mistrust towards data collection.
To anticipate this turning point, several actions are recommended:
Meta’s use of AI conversations heralds a new era in conversational marketing. Advertisers will see this as an opportunity to enrich their customer knowledge, provided they understand the ethical and regulatory limitations.
Please do not hesitate to contact our Social Ads agency to anticipate these developments, activate these levers as soon as they are launched, and maximise the impact of your campaigns on the Meta ecosystem.
Julien Soumahoro