← Back to blog
AI in Marketing

Meta Just Made Its AI Pick Your Ad Type at Cannes 2026: What Live Shopping Ads + Auto-Ad-Assembly Mean for DTC Advertisers

June 24, 2026 By Alex Neiman

Creator broadcasting a live shopping stream with a microphone, illustrating Meta's new live video ads and AI-assembled ad formats announced at Cannes 2026

Updated June 24, 2026.

TL;DR: At Cannes Lions 2026 (June 17), Meta confirmed that product data and creative assets — not campaign structure — are now the levers that matter: you upload both, and Meta’s AI “assembles the best-performing ad for each viewer in real time.” Combined with global Live Video Ads and in-stream shopping, the format itself is now an algorithm decision. For DTC advertisers, the move that actually pays off is feeding the system a clean product catalog and a deep, well-labeled creative library — then re-reading your creative tests, because format is no longer your variable to hold constant.

What did Meta announce at Cannes 2026?

On June 17, 2026, at Cannes Lions, Meta published “New Ways to Turn Discovery Into Purchase” — a commerce-focused release that quietly changes how DTC ads get built. The headline features are about live shopping and frictionless checkout, but the line that should stop every media buyer is buried in the campaign mechanics: Meta’s system now assembles the ad for you.

The key facts, straight from Meta for Business:

“Product data is becoming a foundational input across all Sales campaigns, so Meta’s AI can automatically match the right products to the right people in the most effective format.” — Meta for Business, June 17, 2026

One note before we go further: this is the Cannes commerce release. Meta ran a separate Cannes 2026 announcement covering AI creative and creator tools — don’t conflate the two. And a smaller change spotted by Jon Loomer in mid-June — an AI-disclosure checkbox in Ads Manager — is real, while his report that the Creative Testing Tool now supports seven ads is still unofficial (Meta’s docs say five). Treat that one as rolling out, not confirmed.

Why does “Meta picks the ad type” matter so much?

It matters because the unit of control just moved. For years, the media buyer’s job was to decide the format — single image, carousel, video, collection — and then test creatives within that decision. Meta’s Andromeda system already chooses placement and which creative to serve. Now it chooses the ad type itself, per viewer, in real time. The format is no longer something you hold constant in a test. It’s another thing the algorithm optimizes.

This isn’t a bolt-on feature. Meta has been pushing toward end-to-end automation for years, and the momentum is real: on its Q3 2025 earnings call, CFO Susan Li said revenue running through Meta’s “end-to-end automated solutions” had reached a $60 billion annual run rate, advertisers using Advantage+ lead campaigns saw 14% lower cost per lead on average, and the number of advertisers using at least one video-generation feature was up 20% quarter over quarter (Meta Q3 2025 earnings call). Auto-assembly is the logical endpoint of that curve.

The commerce surface Meta is automating into is enormous, which is why the company is moving fast:

The US commerce surface Meta is automating ($B) Social commerce, 2026 $100B+ Livestream shopping, 2026 ~$68B Livestream shopping, 2024 ~$50B Sources: eMarketer (US social commerce >$100B in 2026); Statista (US livestream shopping). Categories differ in definition.
US social commerce is forecast to surpass $100 billion in 2026 (eMarketer), with livestream shopping climbing from roughly $50B in 2024 toward ~$68B by 2026 (Statista).

When a platform sees a six-figure-billion commerce opportunity, it doesn’t leave format selection to media buyers. It builds a system to make that decision a million times a second. That’s what auto-assembly is — and it’s why this announcement is bigger than “Meta added live shopping.”

What does auto-assembly change for DTC advertisers?

For DTC brands, the immediate impact is that your inputs — catalog and creative library — now matter more than your structure. Here’s where the work shifts.

Your product feed is now ad infrastructure, not a catalog chore

Meta said product data is “a foundational input across all Sales campaigns.” That means a sloppy feed — missing fields, stale prices, weak titles, low-res images, bad product groupings — now degrades every campaign, not just Advantage+ Shopping. If you’ve treated the catalog as a Shopify export you set and forgot, that’s the first thing to fix. The brands that win auto-assembly are the ones whose feed is clean enough that the AI never has to choose from a bad option. This is the same discipline that powers Advantage+ Shopping campaigns — auto-assembly just raises the stakes.

Format is now an algorithm variable, which breaks naive creative tests

If Meta picks the ad type per viewer, you can no longer cleanly attribute a win to “the video creative” versus “the carousel.” The format and the creative are entangled. That doesn’t kill creative testing — it changes what you’re testing. You’re no longer testing formats against each other; you’re testing concepts, hooks, and angles and letting Meta express each one in whatever format performs. I dug into how this reshapes the testing read in my guide on using AI for creative testing in Meta Ads, and it’s worth re-reading now that format is off your control panel.

Asset depth beats asset polish

Auto-assembly needs raw material. A brand that uploads one hero video and three static images is handing the AI a thin menu. A brand that uploads varied hooks, multiple aspect ratios, product-demo clips, UGC, and lifestyle stills gives the system room to assemble something that fits each viewer. The constraint is no longer “make the perfect ad” — it’s “give the machine enough good parts.” This is exactly what Andromeda rewards; I broke down how Meta’s Andromeda algorithm reads creative if you want the mechanics.

What most of the Cannes coverage is missing: ceding format selection only works if you’ve also decided how much creative control to give up. Auto-assembly and Advantage+ creative tweaks are related but separate decisions, and they deserve a deliberate answer — not a default opt-in. I walk through that call in my Advantage+ creative opt-out framework.

Are live shopping ads worth it for DTC?

For most DTC brands, live shopping ads are worth testing only if you already have a live-content habit or a product that genuinely benefits from demonstration — think beauty application, apparel fit, supplement routines, or gadgets with a “wait, it does that?” moment. The virtual-card checkout removes a real friction point, and in-stream purchasing compresses the discovery-to-buy gap that usually leaks conversions.

But live shopping is a content commitment, not a campaign setting. You need someone on camera, a reason to tune in, and a cadence. If you don’t have that engine, a live shopping ad is just an expensive way to broadcast to an empty room. My honest take from running accounts: pressure-test whether your product is shown better than it’s described. If yes, live shopping deserves a budget line. If your product is a commodity refill people buy on autopilot, put that money into feed hygiene and creative depth instead — that’s where auto-assembly will reward you regardless.

What should DTC advertisers do now?

Here are the moves to make, ordered by urgency.

  1. This week — audit your product feed. Check titles, prices, image quality, product groupings, and field completeness. Auto-assembly is only as good as the catalog it draws from, and the feed now touches every Sales campaign.
  2. This week — inventory your creative library by concept, not by format. Tag assets by hook, angle, and message so you can see where you’re thin. The AI needs variety, not one polished hero.
  3. This month — rewrite your creative test design. Stop testing formats head-to-head. Test concepts and let Meta assemble the format. Re-baseline what “a winning ad” means now.
  4. This month — make a deliberate live shopping decision. Either commit to a content cadence and test it properly, or skip it and reinvest in feed and creative depth.
  5. This quarter — decide your automation posture on purpose. Map which controls you’re ceding (format, placement, creative tweaks) and which you’re keeping. Don’t sleepwalk into full automation; choose it.

Do NOT:

The bigger picture

Auto-assembly is part of a clear trajectory: Meta keeps absorbing the decisions media buyers used to make, one layer at a time — bidding, then placement, then creative selection, and now ad type. The job that survives isn’t “operate the levers.” It’s “feed the system better than your competitors and judge its output well.” The buyers who treat their catalog and creative library as strategic assets — not afterthoughts — are the ones who’ll compound while everyone else argues about whether the robots are taking over. For the full operating model, see my DTC Meta ads strategy for 2026, and for where this fits in Meta’s broader AI rollout, my breakdown of Meta’s generative AI ad features and how the Andromeda algorithm actually works.

Related: Meta “Brand Memory” AI Creative in 2026: Will It Keep Your DTC Brand On-Voice — or Flatten It?.

Frequently Asked Questions

What exactly did Meta announce at Cannes 2026?

On June 17, 2026, Meta announced Live Video Ads on Instagram and globally on Facebook, in-stream Live Shopping Tools, virtual one-time-card checkout, expanded affiliate partners, and — most significantly — that advertisers provide product data plus creative assets and Meta’s AI assembles the best-performing ad for each viewer in real time.

Does “Meta picks the ad type” mean I lose control of my ads?

You lose control of format selection, not of inputs. You still decide the product catalog, the creative assets, the budget, and the targeting guardrails. The shift is that Meta now chooses which format to show each viewer, so your leverage moves to feed quality and creative-library depth rather than picking the format yourself.

Are Meta live shopping ads worth it for DTC brands?

They’re worth testing if your product benefits from demonstration and you can sustain a live-content cadence. The virtual-card checkout reduces friction, but live shopping is a content commitment, not a campaign toggle. Commodity, autopilot-purchase products usually see a better return from feed hygiene and creative depth.

Sources