How Good Is Meta's New AI Business Assistant?

0:00Meta, has just added its AI business assistant directly into Ads Manager. And this isn't just another gimmicky chatbot bolted on for show. This has the potential to fundamentally change how Meta advertisers work day-to-day. In this video, I'm going to show you what that AI business assistant can do to help you quickly and easily improve your

0:17Facebook and Instagram ad performance. So, this is big news and I've been having a play with Meta's AI business assistant in Ads Manager, getting it to analyze existing live campaigns and make recommendations. I'm going to show you some real-world examples in a minute. Make sure you stick around for that cuz what it can do is very, very interesting. First, I want to quickly highlight some of the stuff that Meta have said about this um in their article that that they released, right? Claiming

0:41what the AI assistant can help you do. So, I think there's some really important points to just very quickly run through before we dive into the specific examples of what I've been able to get it to do already. So, we've got improved ad performance with the AI assistant that works alongside you. "Everyday advertisers like you spend valuable time navigating complex campaign setups, analyzing performance data, and resolving account issues." So, it's designed to help us save time. Like all AI tools, that's exactly what they're one of their main uh functions, right? This part though is really interesting. "With data from current and prior campaign performance," yeah, that's what we'd expect, "Meta AI

1:14business assistant can instantly generate insights and charts," great, "compare your results to industry benchmarks, and optimize your campaigns." I think this is really interesting where it's talking about not just analyzing your data in your ad account, but also bringing in insights from other accounts, particularly your competitors, that could be really beneficial. So, we've got this section here called benefits. So, "Stronger

1:33results made easy with AI for Meta ads." Just want to quickly highlight these. This won't take very long. Uh "Simplify campaign management with AI." Well, I've been recommending simplifying campaign management for a long time. Um the AI business assistant is going to recommend that as well, also allow you to just simply do it with a click in some

1:48instances. "Get personalized guidance." One of the big advantages is not generic advice that you might get through like Meta support docs um or even from a marketing rep, to be honest, but it's like personalized advice for your business, your situation, which is great. Improve performance. So, this is what the AI business could do. Instantly analyze your campaigns, benchmark your results against businesses like yours. I think that's really interesting. That hints at data being used from your

2:12competitors and and other businesses. Obviously, Meta has tons of that. They use it in their optimization process. So, it's not just looking at your data, also looking at that stuff as well. And get insights to optimize campaign performance. And this one is also really interesting. Resolve account issues. So, fix issues in seconds including restoring disabled accounts, updating spend limits, and resolving delivery errors. Um I have already found this really really useful. I've actually released a video recently on getting ad accounts back uh using a variety of methods, one of which is Meta's AI business assistant, which is a relatively new option. So, if you have a

2:47disabled ad account, you go ahead and find that on my channel and check that out. But I think this is really interesting. Obviously, not just disabling ad accounts, but they specifically call that out, but also um sorting other sort of account issues that can just take ages if you're trying to get in touch with a real person. The

3:01AI business assistant can do a lot. Couple more quick things here. Drive real results with Meta AI business assistant. So, 12% median decrease in cost per result for advertisers who applied opportunity score recommendations. That's not quite what we're going to demonstrate and the AI business assistant, I think that has the potential to be a little bit misleading cuz it's talking about opportunity score recommendations, which you know, you you may have seen pop up underneath opportunity score on the right-hand side when you're in Ads Manager. Um and you can go ahead and apply those. That's basically following along with what Meta would recommend. Um but the AI assistant

3:34advice does give uh recommendations based along those lines anyway. So, it's like basically you follow these, you get better results. It's just that people have been doing this before the AI business assistant was integrated with Ads Manager anyway. But okay, but still worth having. 20% increased resolution rates of common account issues with Meta AI business assistant um like I just said. So, not going to work in every instance, you're still going to need to speak to some people, but definitely worth a go given that it's instantaneous as opposed to waiting for a a lengthy review process. Okay, with that said, let's dive into some examples. So, as

4:04far as I can tell, the Meta AI Business Assistant has now been added to most ad accounts. You'll be able to tell because you get this rather colorful little wheel Meta AI Business Assistant icon over here on the left-hand side if if you've got access to it. Um so, like I said, I'm going to now use this and apply it to live real campaigns and see what it comes up with. So, I've got some different scenarios, so I think that would be helpful. I started with a campaign that's struggling a bit. So, only three purchases generated, £357 cost per purchase. Now, just a bit of context actually on this business,

4:38the average customer value is like just over £5,000, so this is still very profitable. It's still a 15x return on ad spend, but it's a lower conversion volume than we'd like. The cost per purchase higher than we expected. There's a really really large warm audience. This campaign is struggling, just the sort of scenario you'd want to use the AI Business Assistant to help work out what's going on and what we could do next. So, the way it works is you want to either select specific campaigns and then use that to get the AI Business Assistant to analyze it, make recommendations, or you can do it across

5:08the board. I'm going to go ahead and select this campaign then make sure that the ad set and then the ads within here are all selected. Then I'm going to go ahead and click on the Meta AI Business Assistant. You can see we've got five selected here. So, we've got simple questions that we could go ahead and ask or we could type our own in. I like the idea of give me some ideas for ads. You see when I went through and selected the various assets, we've only got three ads in this ad set. Clearly need more, so let's see what the AI Business Assistant can do in terms of

5:38giving ideas. So, what it will do is it'll analyze existing performance to see if any conclusions can be made that could then be used to direct future ad performance, which is exactly what you want it to do. That's exactly what we would do if we were analyzing ads manually as we'd be looking, okay, well,

5:53which ones have got the good hook rates? Which ones have got better click-through rates? Can we use the hook similar hooks to the ones with the highest hook rates, etc., etc. Okay, so let's take a look at this. We've got ad performance analysis and creative ideas. So, it's just talking about what we're after, you know, optimize it for cost per result. Yeah, yeah. Performance analysis. I've analyzed the performance of your campaign associated ad. Here's some of the performance. Yeah, yeah, yeah. Key insights. High engagement, low conversion. All ads show a very healthy CTR above 3%, which indicates that your creative barn video concept, that's just where the the ads were recorded. That's

6:25the location. Just it's in the naming convention. So, Meta's grabbed it. It's highly engaging and relevant to accounts that are ads seeing it. However, this engagement is not translating into a high volume of results. For example, this ad reached just under 14,000 people, only generated three purchases. This suggests a potential friction point on the landing page or a mismatch between the ad's hook and the final purchase intent. So, maybe we need to re-go through and have a look at the landing page and see if there's a disconnect or if the ad itself isn't convincing people enough. Learning phase challenges. Both active ads are

6:56currently in learning limited state. Well, only three purchases, that's not a surprise. Just saying you're going to need some more. Creative fatigue risk. You're currently relying heavily on the variations of the same concept, barn video. While this concept has proven it can stop the scroll, relying on a single creative style can lead to performance plateaus. Okay. So, some really succinct useful analysis there. Creative recommendations. To address the learning limited status and improve your conversion efficiency, I recommend diversifying your creative approach. As we discussed previously, while we've been chatting with this a while, it's learning, you're recommended to iterate on the winning barn video

7:27concept and diversify creative formats. So, test problem/solution static imagery. Since your videos have high engagement but low conversion, testing a static image that clearly calls out a specific pain point and your solution can attract accounts that are accounts, it's just another term for like Facebook and Instagram users with higher intent. This provides a clear immediate value proposition that might convert better than a long-form video. We can click on that and then go ahead and set up an AB test to create some of those. Although I don't necessarily think you would need an AB

7:57test to set this up, you could do it. Implement social proof carousel ads. As mentioned in our previous summary, testing social proof is a powerful way to build trust. Create a carousel ad featuring customer testimonials or before and after results. This format allows account center at accounts to interact with multiple proof points, which can lower your cost per result by addressing objections early in the funnel. I really like this problem solution static imagery. I quite like this social proof carousel ads. These are good ideas. Additional recommendations: leverage advantage plus

8:27creative enhancements. Eh. Less interested in that. Creative enhancements are those things that Meta can do automatically to your ad creative. Um They're more for images than they are for video. I've created a whole video on that if you want to work out which ones you should use and how it should go through. Strongly recommend going through it. I think it's a really um useful video. But because these are video ads that are already edited, I think adding in creative enhancement additions is just not going to make much of a difference. Then we've got broaden reach with advantage plus audience. Um which is you know, it says like given

8:55it's struggling learn limited, enable advantage plus audience. And that's not very good advice for this campaign specifically cuz advantage plus audience is already enabled. Um so that's just doesn't apply here. And that's fine. They're in additional recommendations. I would say that one's just incorrect cuz it's already applied. That one's limited upside. These two are really valuable. And Meta's kind of ranked those as you'd expect. Like with all AI tools, you have to be the filter as the advertiser. You have to have a base understanding. And that's a good example cuz you'll be like, "Hang on, isn't it already?" It is. Um but these two I really like. So

9:26we can go ahead and implement these and help improve performance. Okay, that's example one. Let's take a look at some others. Okay, so here I am inside another Meta ad account. It's only operating at like low levels with a small volume because they only so much they can handle capacity. It's just them at this point. So, it generated 11 leads, £14-ish cost per lead so far, so

9:42doing quite well from that standpoint. So, we've got all those selected. Let's go ahead and select the Meta AI business assistant. And let's say we wanted to reduce the cost per lead. We could say uh something like, "How do I decrease my cost per lead?" And then give that a search. And then Meta's obviously going to analyze how things are uh performing and then make recommendations based on what they find, which is might be a little bit different. Okay, so let's go through this. To decrease your cost per lead for your campaign, we need to focus on improving the efficiency of your lead generation. Currently, your primary ad

10:13set has a cost per result that is 154% above the industry benchmark for lead generation prospecting. Well, that's interesting. While the trend has improved by 52% recently, brand new campaign, sticky in the learning phase at the beginning. It is getting better at optimizing in a brand new ad account. I'd imagine that this the fact that it's a lot more expensive a bunch above benchmark right now is simply that and needs to learn. Also demonstrates, you know, that Meta's tracking this and able to see the trend has improved by 52% recently as it started to learn and generate more conversions. There is significant room to optimize your spend

10:45and lower your cost per lead. Okay. Performance analysis. Analysis of your campaign reveals a clear disparity in how different creative assets are performing, which directly impacts your overall cost per lead. Interesting. Creative efficiency gaps. There is a wide range of CPR across your ads. For instance, this ad is performing exceptionally well with a CPR of £3.79, which is £3.14 below the internal benchmark. That's very interesting. Uh so, significantly lower than the average. Um in contrast, this other ad is struggling with a CPR of £30, £23 over above the benchmark. I think this benchmark information is really helpful for people to know, how well are we

11:19actually doing here? Ad quality concerns. Several ads, including these two, uh flagged as below-average quality ranking. This indicates the audience may perceive these ads as less relevant or engaging, which often leads to higher cost in the auction. Okay, great. So, what should we do? Less ads like those, maybe turn those ones off. Um audience performance. Well, this is a

11:38local business in a specific area. There's not so much we can do here, but it does say, "Look, ad set 18 to 24 demographic is your most efficient segment with a CPR of £4.31. However, cost rise significantly in older brackets, £25 in the 55 to 64 group." Um since you already have Advantage Plus audience enabled, it worked it out this time. The system is working to find the best opportunities, but the creative relevance for these older segments may be a limiting factor. So, if you do want to reach these older segments, then produce more ad creative that talks specifically to them. If you're happy going with younger demographics, then by

12:09all means leave it as is. Uh is basically the conclusion I'm drawing. This is part of the reason why campaigns get so much better, because Meta's worked this out, so more leads are coming through because more of the budget is being spent to those age demographics. Budget constraints. Your current budget is likely a limiting factor uh for performance. When a campaign is budget limited, system cannot explore, etc., etc. Yeah, we can spend more. We don't want tons of leads to this campaign, like I said. But again, I thought this would be a good example for a lot of people watching that are new campaigns, low conversion

12:35volumes. We'll have a look at a very different campaign in a minute. Recommendations. Increase campaign budget. Okay, that's good advice. You have to be a little bit careful whenever the platform says spend more money and that'll help. I think it's good advice in this instance, but you know, take it with a pinch of salt. Apply standard

12:51enhancements bundle. Um as previously discussed, applying visual touch-ups, overlays help improve ad relevance. These ads are quite edited already. They already have text overlays. That's been done manually. I'm less keen to allow Meta to just do whatever it wants to the ads because of that. I think they could look wrong. I don't think it's going to make a big

13:09difference. Even Meta says this. Implementing this could lead to a 3% lower cost per result. Not massive. Utilize generative AI for creative variety. Okay. Using AI to generate background variations or image expansions to help your creative stand out and stay fresh. This helps combat creative fatigue and can 10% higher click-through rate. Um I If you're just a product-based image ads, I would agree, but what's being offered here, lead-based, is a service. It gets a little bit sticky in terms of where you're just creating stuff that didn't actually happen. Like, these are images of the service being implemented. No, we want them to remain true. So, again, I don't particularly

13:42like those as much. You have to be the filter with this stuff. But, the creative analysis, like I've drawn different conclusions, you can tell as I went through that, but the analysis part was great. So, I've drawn different conclusions, that's why it's so important to have a basic understanding and the fundamentals. Additional recommendation, test creative tailored to high-cost demographics. Since your CPR is significantly higher for 45 to 64 age groups, consider AB testing new creative assets that specifically feature people or messaging relevant to that demographic. I already mentioned that. So, I really liked that recommendation, and that was something that I thought as I was going through

14:13it. Um or we could just decide to specialize in the the younger demographic. In this instance, clearly the takeaway I have from this is we need to do more creative production that specifically speaks to those people, uh probably the messaging in the ads as well. One of the biggest challenges with Meta ad creatives is figuring out what's actually going to work before you spend money testing it. That's why I want to introduce you to Higgsfield supercomputer. It's an AI agent that can watch a winning competitor ad frame by frame, deconstruct exactly what makes it work, the hook, the pacing, the structure, and then rebuild it with your

14:43product and your face. So, instead of guessing what creative will perform, you're already starting with something that's proven to work and making it your own. And that's just one part of what it can do. Drop a product link and supercomputer will build out a full creative testing matrix, generating AI avatars and multiple ad formats all in

14:58one go without switching between tools. If you want to stop guessing what creative will work and start testing from a much stronger starting point, there's a link in the description to get started with Higgsfield supercomputer. Okay, let's move on to the the third and final example with a very different business, much larger scale. So, this is a an ad account that's generating far, far more purchases last 30 days, hundreds of purchases, uh operating with a significantly larger budget. Um let's isolate this top campaign for now. Let's go ahead and select that. Let's make sure we grab the ad set

15:27and the ads as we've done previously. Then I'm going to go ahead and select the Meta AI business assistant and then we can decide what question we want to ask it. How can I improve my ROAS in this campaign? Okay, so let's go through this. So this first section I've just read through is basically just it's saying this is what you want, right? Higher ROAS and how it's going to go about doing it. So we'll skip down to the performance analysis. So the campaign has generated 709 results with an average cost per

15:54result of $94. Campaign is actually a significant variance in performance across your individual ads. This is impacting your overall return on ad spend. I mean, that's fairly common to be fair. Top performers, so this is currently leading with a purchase ROAS of 1.62. Oh, by the way, purchase ROAS of 1.62 doesn't sound great, but this business is very much one that generates tons of repeat purchases. So the eventual customer value ends up being far, far larger than that initial transaction, which is the only thing that's accounted for in this return on ad spend number here, right? Is is that initial transaction. So this company, as long as

16:27they're they've got a ROAS above a one, they are very happy with that. So 1.62 is is great for them. 1.62 cost per result of $50. Similarly, this one's doing very well, but you've got underperformers with a ROAS of 0.69. So basically the conclusion here is, right, more like these, less like this. Again, not super helpful analysis, but doesn't hurt to know. Creative engagement, I just said this video has a strong CTR, yet its ROAS is low. This suggests high initial interest that isn't fully translating into high-value purchase compared to your top-performing video. So what I would draw from that is the hook's great. So the initial 3

17:02seconds or so is really good on this one, but the body of the video ad might not be as good. So maybe we could take the the main body from other video ads that convert better, combine it with this hook, and that could be a top performer going forward. Again, that's my conclusion based on my understanding. Uh as you mentioned last time, again, it learns you're looking for creative suggestions to improve ROAS. You were previously recommended to turn on creative plus enhancements. We've talked about that elsewhere. So, it's got enable advantage plus creative enhancements. Only so much can be done,

17:29not too bothered. Optimize for value. This is interesting. So, for the ad set, duplicating it to use the maximize value of conversions performance goal will shift the focus from just getting any purchase to finding accounts that are accounts likely to spend more. That is absolutely our recommendation. This campaign has been running for ages um since before we had that as our default approach, and we've been reluctant to mess with it. Obviously, we introduced new ads and stuff, but to change optimization just because it's been doing so well, hence part of the reason why we have other campaigns running to test out other stuff. We want clear uh

18:01demonstrable improvement performance from optimizing for value in this instance before we switch over. But, if I was to create this campaign fresh today, I would absolutely be optimizing for maximize value of conversions instead of just uh maximize number of conversions. 7% higher ROAS in my experience is higher than that, um but that does depend on uh the type of business um for sure, right? So, yep, I really like that recommendation. It's interesting it says how to do it as

18:24well. It says duplicate it. It's It's like it's a good recommendation there. Uh scale high-performing video concepts. So, effectively spend more money on the ones doing better if you want to improve ROAS, and spend less um on the others. So, it says here that this one's a good performer. This suggests the elite video style is resonating well with your audience. You should consider creating new iterations of this specific concept, perhaps with different hooks or call to actions to maintain momentum without causing creative fatigue. We do that all the time. You stumble across a winner during your testing, right? How can we milk this as much as possible? New

18:55variations, new hooks, etc. etc. Really like that. Refresh underperforming video creatives, the ones that are struggling, uh updating the creative assets, add copy to better align with the messaging found in a more successful ad um before testing again. Basically, turn it off or make adjustments. Uh test new creative variations for static images. So, this image has a decent ROAS of 1.43. Its CTR is quite

19:19low at 0.55%. Indicates that while it converts well once clicked, it isn't stopping the scroll as effectively as your videos. Testing a new version of this image with more vibrant colors or a clearer product focus could help boost engagement and drive traffic more efficiently. So, really good advice there. Testing a new version of this image with more vibrant colors because it's not stopping the scroll well. So, there's some stuff in here that's kind of a bit fluffy, but there's also some stuff like that that's like really specific tactical

19:46recommendations um that I really like. So, that last point on this one was probably um the most valuable. So, you can see some really good stuff in there. So, a couple of things to mention about the advice given from the Meta AI business assistant. You need to be careful with it. Some of the advice is

20:00really good, some of it isn't. Particularly be skeptical around the advice on increasing budgets. Like, who's that serving? A lot of the time advice and recommendations from Meta as a platform will help you as an advertiser, but sometimes it helps Meta out more than it does the advertiser. So, this is where you need to understand the fundamentals. You need to understand best practices. You need to stay

20:21up-to-date. Subscribe to my channel. Watch my videos one a week. Uh doesn't take up too much time. Will absolutely help you with that and you'll understand, okay, I should take this advice, I shouldn't take it this advice, that advice probably won't make any difference. Uh and you'll get to to learn that stuff. Um if Meta ads are important to your business, like I said, subscribe. Make sure you keep an eye out for new videos. I think you'll find that really, really helpful. What I think this is great for is new ideas and time saving. So, it's doing analysis that would take significantly longer to do yourself manually, and it's also going

20:51to come up with things that you might already know, but you just didn't think of in the moment cuz there's so many different things and so many different options to think about. So, I know for example that more vibrant colors on an image is a great solution if your uh click-through rate is low cuz it's not stopping the scroll. But, I might not think of that cuz I'm looking at all the data and all the various elements and I might just miss it. Whereas, boom, it just went and pulled that out for me and then I can see that clearly, low click-through rate, high conversion rate, brilliant. Let's Let's jazz up the

21:19image and make it clearer, more product-focused, or or more vibrant colors. Um some of the other examples, like this ad is doing really well from a click-through rate standpoint, but it's not converting. Is there a disconnect between the ad and the landing page? Like again, I know that that's good advice, but it just being able to pull that out for me means I can just immediately go have a look instead of

21:39having to do loads analysis um first. And if you don't want to do that work, you don't want to go through that process and stay up to date, my agency can run your ads for you. If there are other things you want to be cracking on with, you can leave that up to us. Most

21:52likely get significantly better results. There is a link in the video description to inquire if you're interested. Next, I'd recommend you check out this video on the best way to scale your Meta ads with the help of the AI business assistant and maybe even more so, my content. Hopefully, you'll be generating fantastic Facebook and Instagram ad results very soon. If you are, you will want to scale. You want to make sure you do it the right way. I lay it all out in

22:14this video here.