Meta AI ads manager
Meta Ads Manager is the tool for building, running and measuring ads across Facebook, Instagram, Messenger, Threads and the Audience Network, and its AI layer is branded Advantage+. That layer now picks your audience, your placements, your budget split and which creative combination each person sees. It does not decide what the ad should say, and that is the one input still fully yours. Selzee is the layer that answers it: it reads your reviews, comments and campaign performance and returns the hooks, angles and briefs worth testing, so your team stops guessing and starts launching ads grounded in what customers actually said.
Meta Ads Manager is Meta's own tool for creating ads, managing when and where they run, and tracking how they perform across Facebook, Instagram, Messenger, Threads and the Audience Network. It is the single place where money is committed and results are reported. Every other tool in a paid social stack, this one included, sits either side of it.
Building a new ad starts with an objective. Sales, leads, app promotion, engagement, awareness and traffic each load a different optimisation model, and that choice does more to determine your outcome than any setting after it. An ecommerce campaign optimised for engagement will reliably buy engagement. From there you attach creative, whether that is a single image, a video, a carousel or a collection, plus the headline, primary text and destination.
Audience definition is the part that has changed most. Detailed interest targeting used to be the craft of the job. Today you supply a suggestion and Meta looks beyond it, with custom audiences from your customer list or site traffic acting as a strong signal rather than a boundary. Budget works the same way: you set a daily or lifetime amount, and campaign budget optimization moves it toward whichever ad set is converting, with cost and bid caps available when unit economics matter more than volume.
Delivery spans every Meta surface from one ad set, so a single upload can serve a Reel, a Story, a feed placement and an in-stream slot, each cropped differently. Live campaigns stay editable: you can pause an ad, shift budget or swap creative mid-flight, though large edits restart the learning phase, which is why experienced buyers pause at the ad level and leave the ad set alone. Reporting updates in near real time, with cost per result, frequency, and hook and hold rates on video available while spend is still running.
One point of confusion worth clearing up: Ads Manager is not Meta Business Suite. Business Suite is the wrapper that holds your assets, pages, Instagram accounts, pixels, product catalogues and the people with permission to touch them. Ads Manager is the advertising surface inside it. Any brand spending real money should run ads through a Business Manager rather than a personal profile, because that is what separates the ad account from one individual's login and lets you keep two admins on it.
What Ads Manager does not do is tell you why. It will show that one video held attention twice as long as another and stop there. It has no view on which customer objection the winner happened to answer, and no way to know, because it never reads your reviews. That gap is the reason this page exists, and it is what a Meta ads tool built on customer data is for.
Searching for a "Meta AI ads manager" turns up no such product, because the AI is not a separate tool. It is Advantage+, a set of automations layered through the campaign you were already building. Meta has been folding those automations into the default path rather than offering them as an alternative one: Search Engine Land reported the simplified Advantage+ setup that removes the choice between manual and automated campaigns, and Social Media Today covered the same AI options moving into the standard ad flow, on by default for audience and placements. For anyone running ecommerce campaigns today, automated is not a mode you switch on. It is the road you are already on.
Meta also grades your setup now. Opportunity score, which Meta made available in Ads Manager globally in June 2025, is a 0 to 100 rating where a higher score indicates better optimisation and a greater likelihood of improved performance over time. It is worth reading before you publish. It is also worth understanding what it grades: your configuration, including creative variety and signal quality. Not whether the thing your ad claims is the thing your buyers care about.
You supply an audience suggestion rather than a hard target, and Meta treats it as a starting hint before looking wider. It is switched on by default in new campaigns. In practice this retires most of the manual interest stacking teams used to do, and it means the lever you still control is what the ad says, not who sees it.
Custom audiences built from your customer list, site traffic or engagement still matter, and lookalikes still work off them. Under Advantage+ they behave as strong signals inside a broader delivery rather than as a fence. Exclusions are the part worth keeping tight, because excluding recent purchasers is still your job, not the model's.
One ad set can serve Facebook, Instagram, Messenger, Threads and the Audience Network, with the split decided by predicted performance rather than by you. Reels, Stories and feed each want different framing, so the practical requirement is supplying assets that survive every crop rather than opting placements out.
You upload multiple images, videos, headlines and primary texts, and Meta assembles and personalises combinations per person, including automatic enhancements to copy and visuals. It optimises the permutation. It cannot invent a claim you did not upload, which is why variant count is a weak substitute for genuinely different arguments.
Campaign budget optimization distributes spend across ad sets toward whatever is converting, on a daily or lifetime budget, with optional cost and bid caps. Caps are the honest way to enforce unit economics. Setting them too tight starves learning, and the failure mode looks like a delivery problem when it is a maths problem.
Three levels, and the split of responsibility has changed. The campaign holds the objective and usually the budget, the ad set holds audience, placement and schedule, and the ad holds the creative. Automation absorbed most of the middle layer, which is why creative at the bottom now carries more of the result than it used to.
Each of them optimises distribution. Not one of them generates an argument. Advantage+ creative will test your headline against your other headline and personalise the pairing per person, and it will do that better than any human. It cannot write the headline that names the objection stopping your buyers from checking out, because that sentence exists only in your reviews and your support inbox. As automation takes the targeting lever away, the remaining lever gets heavier: Nielsen credits creative, not targeting or media spend, with up to 89% of a digital ad's in-market success, and Meta's own research on creative quality and ad ROI agrees. Getting it wrong costs more each year: Facebook cost per lead climbed almost 21% year over year in 2025, per WordStream's benchmarks.
Sign in with a Facebook account that has a role on the ad account, through Meta Business Suite or adsmanager.facebook.com. Confirm the account selector top-left before you touch anything: launching into the wrong account is the most common and most expensive first-minute mistake on agency logins.
Sales for purchases, leads for form fills, app promotion for installs, awareness and engagement for the upper funnel. The objective decides which optimisation model runs, so an ecommerce campaign optimised for engagement will faithfully buy you engagement instead of revenue. Sales, leads and app promotion now enter the automated flow by default.
Supply an audience suggestion, keep Advantage+ placements on, and set a daily or lifetime budget with campaign budget optimization handling the split. Add your exclusions here. Budget high enough that the campaign can clear its learning phase, because a budget too small to gather conversions never stabilises.
Add your videos, images, headlines and primary texts, and supply a vertical crop that works in Reels and Stories as well as feed. This is the step where the campaign is actually won or lost, and it is the one step no automation does for you: the model chooses between arguments you wrote.
Check the preview in each placement, confirm the pixel or Conversions API event matches the objective, and read the opportunity score before you publish. Meta made that 0 to 100 score available in Ads Manager globally in June 2025, and it flags setup problems that are far cheaper to fix now than after spend.
Watch cost per result, hook rate and hold rate on video, and frequency. Give the campaign its learning phase before judging it. Reacting to hour-two numbers is the classic way to reset learning repeatedly and then conclude the platform is broken.
Pause at the ad level and leave the ad set intact where you can, since large edits restart learning. Rising frequency with falling click-through is creative fatigue, not an audience problem, and the fix is a new argument rather than a new interest stack.
Use the built-in A/B test to compare one variable at a time. Test different claims and hooks against each other rather than six colour variants of the same claim: the model already optimises permutations, so a test that only varies the artwork gives you information Meta would have found anyway.
Read those eight steps back and notice the distribution: seven are mechanical and take minutes, and one, uploading creative and copy, is the one nobody can do quickly. Teams do not lose weeks configuring ad sets. They lose weeks deciding what the next test should argue, reading reviews in a spreadsheet, screenshotting competitor ads, and arriving at a brief built mostly on instinct. The platform got faster at everything except the part that determines the result.
That is the step Selzee removes. It connects to Meta Ads, Shopify, Klaviyo and your review platforms, ranks the objections and desires your customers actually raise by how often they raise them, and returns hooks and briefs with the customer quote attached to each claim. Your strategist edits and approves instead of starting from an empty document, which is what collapses the cycle from a week to an afternoon. The launch still happens in Ads Manager. You just arrive at step four knowing what to upload. The same discipline applied to rivals is covered in competitor ads analysis.
"Advantage+ took targeting off the table as a source of edge. Every competitor in your category is pointed at the same broad audience by the same model. The only variable left is whether your ad says something true about your customer that theirs does not, and that answer has never been in the ads platform."
Meta ships a dedicated Ads Manager app on iOS and Android, described as a way to create, edit and get insights from ads across Facebook, Instagram, Messenger and WhatsApp. It signs into the same ad accounts as the desktop interface with no separate setup, so edits made on a phone appear on the web immediately and vice versa. Push notifications cover budget exhaustion, disapprovals and delivery drops, which is the honest case for installing it: you find out a campaign stopped spending while you can still fix it.
Treat it as a triage tool rather than a workstation. Checking cost per result, pausing a fatigued ad and topping up a budget all work well on a phone. Building a campaign, comparing creative properly or handling anything involving payment settings does not, and several controls remain desktop-only. On privacy, the App Store listing declares an unusually broad set of data linked to the user, including purchases, financial info, precise location, contacts, browsing history, identifiers and usage data, used for third-party advertising, Meta's own marketing, analytics and personalisation. Worth reading before you install it on a personal device rather than a work one.
The app holds a 4.5 star average across roughly 50,000 US ratings on the App Store, so the reception is broadly positive and the loudest reviews are not representative. Those reviews do cluster on a real theme worth planning for: account compromises, sudden restrictions with thin explanation, and difficulty reaching a human at Meta. None of it is specific to the app, and all of it is survivable with basic hygiene. Enable two-factor authentication on every user with account access, keep at least two admins on the Business Manager so a single lockout does not orphan the account, run ads through a Business Manager rather than a personal profile, and keep creative and audience assets in your own systems so a restriction costs you access rather than the work itself.
Run Advantage+ sales campaigns the way they are designed to be run: broad, with exclusions rather than restrictions, and with enough budget to clear the learning phase. Fighting the automation by narrowing the audience is the most common way teams underperform it. Give it room and spend your own effort on the input it cannot generate. Feed it genuinely different arguments rather than more variants of one, because six colourways of the same claim tell you nothing the model would not have found on its own.
Build for Reels first. Vertical video is where attention sits and where the cheapest impressions are, and an asset that only works cropped to a square feed placement is an asset the automation will struggle to place. Shoot the hook for sound-off viewing and put the claim in the first three seconds. Organic content earns its place here too: a post that already performed has proof of concept attached, and promoting it costs less than producing a new asset that has never been tested on a real audience.
Turning that organic content shoppable is the step most teams skip. Connect your product catalogue in Business Manager, tag products in the posts and Reels that already earned attention, and those posts become buyable surfaces rather than reach for its own sake. The same catalogue then feeds Advantage+ catalogue ads and dynamic product retargeting, so one integration serves both the organic and the paid side. The reason to do it in that order is evidence: a post your audience already engaged with has demonstrated its claim works, which makes it a far safer thing to put budget behind than a concept that has only ever been approved in a meeting.
On account safety, most restrictions trace to a handful of avoidable triggers: advertising from a personal profile with no Business Manager behind it, sudden large spend increases on a young account, landing pages that do not match the ad claim, and prohibited-category language in copy, particularly around health, weight and income claims. Warm a new ad account up rather than starting it at full budget, and keep the landing page consistent with the promise.
Finally, read the real-time reporting for diagnosis rather than for reassurance. Rising frequency with falling click-through is creative fatigue and needs a new argument, not a new interest stack. A high hook rate with a poor hold rate means the opening promise and the body do not match. Both are creative problems that show up as media metrics, and both are cheaper to fix when you already know which objection your customers keep raising. Grading the argument before launch, rather than the asset after it, is covered in scoring ad creative, and Facebook ads strategy covers the testing cadence around it.
It is not a separate product. Meta Ads Manager is the tool for creating, running and measuring ads across Facebook, Instagram, Messenger, Threads and the Audience Network, and the AI layer inside it is branded Advantage+. That layer chooses audiences, placements, budget splits and creative combinations automatically, and it is on by default for new sales, leads and app promotion campaigns.
For most ecommerce advertisers it is now the default path rather than a choice, because Meta merged the manual and automated flows. The automation is genuinely better than manual targeting at finding buyers. It has no opinion about what your ad should say, so the advantage it delivers depends entirely on the quality of the arguments you feed it.
A 0 to 100 rating of how well a campaign is set up, which Meta made globally available in Ads Manager in June 2025. A higher score indicates better optimisation and a greater likelihood of improved performance over time. It grades your setup, including creative variety and signal quality. It does not grade whether your message is right.
Yes. The Meta Ads Manager app is on iOS and Android, syncs with the same ad accounts as desktop, and lets you create and edit ads, check performance and respond to notifications. Some settings remain desktop-only, so treat it as a monitoring and triage tool rather than a full replacement for the web interface.
Selzee sits upstream of it. Ads Manager reports what happened and automates delivery. Selzee reads your reviews, comments, product data and campaign results, then returns the hooks, angles and briefs worth testing, each carrying the customer quote behind it. You still launch in Ads Manager. You stop guessing at what to launch.
No, and nothing does. Ads Manager is where the money is spent and the results are reported, and no third-party tool substitutes for it. Selzee replaces the manual research that happens before you open it: the days spent reading reviews and competitor ads to work out what the next test should argue.
Book a demo and bring your Meta Ads account, your reviews, and your toughest creative bottleneck. We will show you exactly how Selzee turns that raw data into hooks, briefs, and angles you can test this week, with the customer quote behind every claim and no manual digging required. You keep launching in Ads Manager. You just stop guessing about what to launch.
Request a DemoPrefer to explore on your own? Get started or browse free ecommerce tools.
Keep exploring: the Meta ads tool for ecommerce teams, what Meta Andromeda means for your ad strategy, Facebook ads AI tools, Facebook ads strategy, Meta ads templates, Meta competitors statistics, scoring ad creative, competitor ads analysis.