Meta just dropped a MAJOR update that’s about to change performance marketing on Facebook & Instagram

From 16 December 2025, Meta will start using your conversations with Meta AI (text + voice) to decide:

• What ads you see
• What content shows up in your feed
In short: your chats become targeting signals

What’s changing?

• Every time someone asks Meta AI a question (“best running shoes under ₹5K”, “how to start a podcast”, “CRM tools for small business”), that intent can now be used to power ad delivery

• This goes beyond interests & behaviour. It’s real-time, explicit intent – what people say they want right now

• No separate “opt-in” for ads. If you use Meta AI, your chats can fuel targeting.
This is the closest Meta has come to search-like intent inside a social product

Why this is HUGE for marketers

Here’s what it means if you run ads:

• Higher-intent audiences:

You’re no longer only chasing people who clicked on a Reel or visited a website. You’re tapping into people actively asking for solutions

• Smarter delivery for performance ads:

Meta’s AI can better predict who is likely to buy / sign up / install – which should improve ROAS, CVR, and lead quality over time

• Faster learning loops:

As AI conversations scale, signals grow richer. Performance campaigns (Advantage+, broad targeting, value optimisation) will likely outperform over-segmented manual setups

• Creative relevance becomes make-or-break

If the system is matching ads to intent, lazy, generic creatives will die fast. Context-rich hooks that mirror user questions will win

How brands & ad teams should respond (practically)

If you’re a marketer, media buyer, founder or creator-led brand, here’s what to start doing:

1) Rewrite your hooks for intent

• Old: “Premium wireless earbuds for everyone”
• New: “Looking for wireless earbuds under ₹5K that don’t drop from your ears?”

2) Lean into broad + Advantage setups
Let Meta’s AI do the heavy lifting. Give it:

• Clean conversion data (CAPI, proper events)
• Enough budget & time to learn
• Multiple creatives to test, not 1–2 static ads

3) Build modular creative libraries:

• Multiple angles: problem, comparison, testimonial, offer
• Multiple formats: short UGC, explainers, carousels, product demos
You’re optimising a creative portfolio, not a single “hero” ad

4) Double down on first-party data:

• Pass back accurate purchase / lead quality data
• Use this with Meta’s AI to train it on your best customers, not just generic buyers

5) Watch performance, not just CPM
CPM might go up in the short term as the system tests. The real question:

• Is ROAS improving?
• Are CPAs stable or dropping?
• Is lead / customer quality better?

The mindset shift

For years, performance marketers learned to:

• Hack interests
• Stack lookalikes
• Over-engineer ad sets