Meta Ads After Andromeda: What Compliance Looks Like Now (2026)

Andromeda is Meta's AI retrieval engine — the system that decides which ads from tens of millions of candidates are even considered for a person, built on a deep neural network with 10,000x the model capacity of what it replaced. GEM, its LLM-scale ranking foundation model, sits above it. Here is the part most explainers skip: neither is an enforcement system. Meta has not announced an "Andromeda policy." The rules that reject ads are still the Advertising Standards, documented separately — along with re-review of live ads and proactive enforcement at scale (Meta reported removing over 159 million scam ads in 2025, 92% before anyone reported them). What changed is the environment: automation multiplies your creative variants, retrieval models learn from people and ad content together, and every variant competes under the same rules. Compliance is the same discipline, applied to more surface, faster.

Key takeaways

Andromeda = retrieval (which ads get considered). GEM = ranking (which considered ad wins the impression). Neither writes or enforces policy.
The rejection rules are the Advertising Standards — unchanged by Andromeda, updated on their own schedule (Meta's Health and Wellness standard shows a July 23, 2026 change-log entry).
"Andromeda made rejections worse" is advertiser observation, not a Meta statistic — Meta publishes no rejection rates.
What is documented: ads can be re-reviewed at any time, and enforcement is increasingly proactive at large scale.
The practical shift: creative volume exploded, so per-ad checks became per-variant checks.

What Andromeda actually is

VERIFIED FACT. Meta's engineering post (December 2, 2024) describes Andromeda as a personalized ads retrieval engine: the first stage of the recommendation pipeline, selecting from tens of millions of ad candidates down to a few thousand per person. Key documented points (Meta Engineering: Meta Andromeda):

A custom deep neural network co-designed with NVIDIA Grace Hopper Superchips and Meta's MTIA accelerator hardware.
A 10,000x increase in retrieval model capacity, replacing rule-based heuristics and isolated model stages.
Meta reports +6% recall improvement in the retrieval system and +8% ad quality improvement on selected segments.
Built to absorb the creative explosion from Advantage+ automation and generative AI tools — over a million advertisers used Meta's GenAI tools to create more than 15 million ads in a single month.

What GEM adds

VERIFIED FACT. Meta's engineering post (November 10, 2025) describes GEM — the Generative Ads Recommendation Model — as Meta's most advanced ads foundation model: LLM-scale, trained across thousands of GPUs, and the largest recommendation foundation model Meta says the industry has deployed. Documented results: a 5% increase in ad conversions on Instagram and 3% on Facebook Feed in Q2 2025, and a Q3 architecture change that Meta says doubled the performance benefit gained per unit of data and compute (Meta Engineering: GEM). GEM's knowledge is transferred into the hundreds of smaller models that rank ads across Meta surfaces; Meta describes a future multimodal foundation spanning text, image, audio, and video across organic and ads behavior.

The enforcement layer is a different system — and it did move

VERIFIED FACT. Policy enforcement is documented outside the Andromeda/GEM posts, and those are the changes that affect whether an ad runs:

Ads are reviewed continuously: Meta states an ad may not be checked against all policies before it starts delivering, and all ads are subject to re-review at any time (About Ads In Review).
Previously approved ads get pulled back into review after negative feedback — people hiding, blocking, or reporting — or at random for accuracy checks (Why some ads are approved, then rejected).
Proactive enforcement operates at large scale: Meta's Integrity Reports state that in 2025 it removed over 159 million scam ads, 92% of them before anyone reported them (Meta Integrity Reports, H1 2026).
Policy standards keep updating on their own cadence — Meta's Health and Wellness advertising standard shows change-log entries including December 27, 2024 and July 23, 2026 (Health and Wellness standard).
Severe or repeated violations escalate from ad-level to account-level enforcement across the business portfolio, ad account, Page, or user account (Common examples of Meta advertising policy violations).

EVIDENCE-BACKED. Advertisers in regulated verticals — supplements, health, finance — report higher rejection rates in 2025–2026, and third-party trackers attribute it to better automated detection. Meta does not publish rejection-rate figures, so treat any specific percentage you see online as anecdotal. What you can verify yourself: the standards text, its change log, and your own account's ad-review outcomes in Business Support Home.

Post-Andromeda compliance workflow

EVIDENCE-BACKED RECOMMENDATION — built on Meta's documented review behavior, not on Andromeda speculation

Keep a claim inventory: every claim you make anywhere, with the evidence file behind it. Variants inherit claims — control them at the source.
Write variants from the compliant claim set, not around it. Automation multiplies whatever wording you feed it.
Scan each variant's copy against the Meta profile before submission — the same 65 detectors, per variant.
Watch for re-review signals you control: hide/block/report rates on live ads pull ads back into review; creative fatigue raises them.
Re-check the standards' change log before each campaign flight — the Health and Wellness standard changed as recently as July 23, 2026.
Treat a clean approval as temporary. It was never a permanent license, and re-review is documented behavior.

What Optimus Pass checks (and what it does not)

OPTIMUS HEURISTIC — detector families, not Meta rules; scores are directional

Optimus Pass scans ad copy against 65 risk detectors across 10 platform profiles. For Meta campaigns in the post-Andromeda environment, the relevant detector families are:

Prohibited and restricted claim language — disease, treatment, or clinical-sounding wording.
Overpromise framing — "guaranteed", "instantly", "100 percent", and similar absolute outcomes.
Personal-attribute wording — language that implies the reader's health, financial, or other sensitive status.
Restricted-vertical patterns — supplements, weight loss, financial services, and similar categories.
Implied transformation narratives — before/after and decline-and-restoration stories told without claim words.

It does not read Andromeda's or GEM's decisions, cannot predict delivery, and does not access ad accounts. Delivery performance and policy enforcement are Meta's systems; Optimus Pass only reviews the copy you paste in.

Scan every variant before you submit

Free, no signup for an anonymous scan: paste the ad copy, select Meta, and read the risk flags. A clean scan does not guarantee approval — the decision is Meta's, at any time, on re-review too.

Run the free Ad Compliance scan →

Scaling creative volume?

The Agency plan covers 250 checks a month with shareable client reports — built for the reality that post-Andromeda accounts run dozens of variants, not one hero ad.

Compare plans →

Frequently asked questions

What is Meta Andromeda?+

Per Meta's engineering blog, Andromeda is Meta's machine-learning system for ad retrieval — the first stage of the ads recommendation pipeline, which narrows tens of millions of candidate ads down to a few thousand relevant ones per person. It runs on a custom deep neural network co-designed with NVIDIA Grace Hopper Superchips and Meta's own MTIA hardware, enabling a 10,000x increase in model capacity, with Meta reporting a 6% recall improvement and 8% ad quality improvement on selected segments. It launched in late 2024.

Did Andromeda change Meta's ad policies?+

No — and this is the distinction most explainers get wrong. Andromeda is a retrieval system and GEM is a ranking model; Meta's engineering posts describe them as delivery-performance systems, not enforcement systems. The rules that reject ads are the Advertising Standards, and Meta documents enforcement separately — re-review of live ads, the Health and Wellness standard updates, and proactive removal statistics in its Integrity Reports. What Andromeda changes is the environment your compliant ads compete in, not what is allowed.

What is GEM in Meta ads?+

GEM (Generative Ads Recommendation Model) is Meta's ads foundation model, described by Meta in November 2025 as the largest recommendation foundation model in the industry, trained at LLM scale across thousands of GPUs. Meta reported it delivered a 5% increase in ad conversions on Instagram and a 3% increase on Facebook Feed in Q2 2025, and that a Q3 architecture change doubled the performance benefit per unit of data and compute. GEM transfers learnings into the hundreds of smaller models that rank ads across Meta surfaces.

Why do advertisers say rejections got stricter after Andromeda?+

Advertiser reports — especially in regulated verticals — describe more rejections and faster enforcement in 2025–2026. That is observation, not a published Meta statistic: Meta does not publish rejection-rate figures. The documented changes that plausibly drive the experience are automation everywhere (retrieval, ranking, creative generation), Meta's statement that ads can be re-reviewed at any time, and its proactive enforcement scale — Meta reported removing over 159 million scam ads in 2025, 92% before anyone reported them.

Does Andromeda mean creative quality affects compliance?+

Indirectly, yes. Andromeda's retrieval models learn higher-order interactions between people and ad content rather than relying only on advertiser-set targeting, and GEM is trained on ad content plus engagement data from both ads and organic behavior. That means more of your ads compete in more auctions, each with its own wording and imagery. Every variant still has to pass the same Advertising Standards, so a scaled creative library multiplies your compliance surface even though the rules did not change.

How should compliance work change in the post-Andromeda era?+

Shift from checking one hero ad to checking the whole creative set: keep a claim inventory with evidence behind each claim, hold every variant to the same standard, scan copy against the Meta profile before submission, and treat approval as temporary — Meta documents re-review of live ads after negative feedback or at random. The discipline is the same as before Andromeda; the volume it has to cover is not.

Can Optimus Pass keep my ads approved under Andromeda?+

No tool can promise that — Meta owns the decision. Optimus Pass scans ad copy against 65 risk detectors across 10 platform profiles, including the claim-language and personal-attribute patterns that drive most regulated-vertical rejections, and suggests lower-risk wording. Use it per variant before submission. It does not access your account, contact Meta, or influence delivery or enforcement outcomes.

References

Claim classes used on this page: VERIFIED FACT (documented directly by Meta), EVIDENCE-BACKED (grounded in those sources or in clearly-labeled advertiser observation), OPTIMUS HEURISTIC (Optimus Pass scoring conventions).

Meta Ads After Andromeda: What Changed, What Did Not