If you run a clinic, a dental practice, or a trades business and you advertise on Facebook or Instagram, you already know the drill. Meta sends you recommendations. In the ads manager, in your inbox, sometimes on the phone. Apply this suggestion, raise this budget, broaden this audience.
Now that same advice is getting an AI voice. Search Engine Land, a search marketing news site, reported on August 20, 2026 that Meta AI can now analyze and optimize Meta Ads campaigns. Meta is the company behind Facebook and Instagram, and Meta AI is its AI assistant. In plain terms, the assistant can now look at your ad account and tell you what to change, or change it for you.
Meta is incredible. Meta recommendations are not.
Let me separate two things, because owners mix them up all the time. Meta the ad platform is genuinely one of the best lead generation machines ever built. We have run campaigns on it for local and professional service businesses for over a decade, and when it is set up right, it prints appointments.
Meta recommendations are a different animal. For years, agency owners like me have received them through the ads manager, through emails, and through phone calls from Meta reps. In my experience, following them blindly grows one number reliably: your spend. Now small business owners will receive those same recommendations through Meta's AI assistant, delivered in a friendly conversational tone that makes them feel like advice from a trusted expert.
The incentive problem an AI assistant does not fix
Here is the core issue, and it has not changed. Meta earns money when you spend money. So its optimization suggestions naturally pull toward more budget, broader audiences, and more automated placements. None of that is evil. It is just the business model.
You have a different job. You should be optimizing for ROI, meaning the actual profit that comes back from what you spend. For a service business, that means following a lead all the way from the ad click to the signed contract or booked treatment plan, and knowing what that deal was worth. Meta AI cannot see your closed deals unless you deliberately feed that data back to it. Left on its own, it optimizes toward what it can see, which is clicks, form fills, and spend.
The gap between a form fill and a paying customer
This gap is where service businesses lose the most money on Meta. A campaign can generate a hundred cheap leads that never answer the phone, and the platform will call it a win. Meanwhile a campaign with fewer, more expensive leads might be producing every real job on your schedule.
The fix is offline conversion tracking. That is the practice of sending your closed deals from your CRM or booking system back into Meta, so the platform learns which leads actually turned into revenue. Once that loop exists, the machine starts hunting for buyers instead of clickers. Setting it up properly takes real technical work and clean sales data, and in my experience it is the single biggest difference between accounts that look busy and accounts that make money. You need an agency that knows what it is doing to build it.
What I expect this to change, and what it will not
My bet is that Meta AI will make it very easy to run a mediocre campaign. An owner with no ads experience will be able to get something live, get some leads, and feel like a marketer. For businesses currently running nothing, that is probably a step up.
Running a top 1 percent, high ROI campaign is another matter. The way I see it, that will take as much finesse as it ever has: sharp offers, creative that speaks to your actual customer, landing pages that convert, and offline conversion tracking wiring the whole thing to revenue. An AI assistant trained to keep you spending does not replace any of that. My take is simple: the floor just got higher, and the ceiling did not move an inch.
At North Digital we have managed multiple millions in ad spend across Google and Meta for service businesses, and the winning accounts share one trait. The owner or the agency measures what a lead is worth after it closes, then makes the platform chase that number.
Meta AI advice vs real ROI
How to prepare your business
1. Treat any Meta AI recommendation as a sales pitch, and check it against your cost per booked job before applying it.
2. Start recording which leads turn into paying customers and what each deal was worth, even if it is just a spreadsheet this week.
3. Ask whoever runs your ads whether offline conversion tracking is set up, and if the answer is vague, dig deeper.
If you want us to look at your Meta account and show you what it is actually returning per closed deal, book a call with me.
