NOTES

• Phil Ilic's LinkedIn 
• Kiin — LinkedIn Ads MCP for reporting
• Join the LinkedIn Ads Fanatics  and get access to our 4 courses to take you from beginner to expert.
• Subscribe/Rate/Review 
• Contact us with any questions, suggestions, or corrections!
• Resty Concepcion podcast  and video  editor that we love!

SUMMARY

In this episode of The LinkedIn Ads Show, AJ Wilcox dives into how AI is transforming LinkedIn Ads: from the good, to the bad, to the genuinely useful. He breaks down LinkedIn's new AI features like Draft with AI, Brand Kit, predictive audiences, and auto-targeting, explaining which tools are worth using and which ones deserve caution. AJ also shares how his team is using AI behind the scenes to build custom reporting systems, automate ABM list processing, and dramatically reduce manual work. Plus, he covers several new LinkedIn Ads updates, including personalized Message Ad subject lines, Lead Gen Forms for Thought Leader Ads, and practical AI workflows that can save advertisers hours every week.

TRANSCRIPT

LinkedIn's baked AI into everything: ad copy, imagery, even targeting. Let's dive into the good, the bad, and the ugly on how AI is affecting your LinkedIn ads on this week's episode of The LinkedIn Ad Show.

Welcome to the LinkedIn Ads Show. Here's your host, AJ Wilcox.

Hey, hey, hey there, LinkedIn ads fanatics. As he said, I'm AJ Wilcox. I'm the host of the weekly podcast, The LinkedIn Ads Show. I'm thrilled to welcome you to the show for advanced B2B marketers in their evolution of mastering LinkedIn ads and achieving true pro status. I'm getting asked a ton about how AI is affecting LinkedIn ads.

I'm actually putting a presentation together on the same topic, so I wanted to actually take this opportunity to share it with you all. The connection with LinkedIn and AI runs super deep. As I'm sure you know, Microsoft owns LinkedIn, and it's also the single largest investor for OpenAI, that owns ChatGPT.

So you can imagine how tight of a connection we have here. Since the very beginning, LinkedIn has worked really hard to bring AI into every facet of every one of their products, and the ads platform is, of course, no different. We'll talk about all the AI features so far, and then I'm gonna flip the script and show you some genuinely cool things that AI has built for me and my team behind the scenes.

The LinkedIn Ads Show is proudly brought to you by B2Linked.com, the LinkedIn ads experts.

That's right, B2Linked is the ad agency 100% dedicated to LinkedIn ads, and we have been back since 2014. You know, before it was cool. We build a custom strategy for every account we work with. You get to work directly with me and my local team, and you won't get any sort of a cookie-cutter approach or a standard account template from us.

Plus, with the strategies that we've developed and our mastery of the platform, we always save our clients more than we charge, so it's kind of like getting the best in the biz for free. If you'd like to explore partnering with us for your LinkedIn ads management, schedule your free discovery call at b2linked.com/discovery.

First off in the news, one of our super fanatics in the community, Anthony Blatner, he shared about how in testing right now is the ability to personalize subject lines in message ads, which I think is awesome. Forever, we've been able to insert things like first name, last name, company, title, into the message portion of a message ad or a conversation ad.

But now it looks like we're gonna get the ability to customize and put someone's, let's say, first name, right in the subject line. I think this has been coming for a long time, and I'm really glad we're gonna have it. Also, message ads seem to be getting the same treatment as conversation ads now, where LinkedIn is going to add a call to action button of Not Interested.

It doesn't hurt your performance when people hit this, but we also don't get any sort of reporting on how often people are using it. I'd sure love those analytics. It'd be super helpful as a metric. But if someone clicks it, it means they're not gonna get that message again, and you're also sending a signal to LinkedIn that, you know, maybe this offer isn't interesting to them.

This is also right fresh off the press. I was just on a consulting call with a large spending account who has access to something I've been waiting for for a long time. This is the lead generation objective being able to be paired with your thought leader ads. It was so cool to see it in action. I really hope that we're gonna get thought leader ads in the website visits objective as well.

I'm crossing my fingers. All right. Each week I'm shouting out a company and wearing their swag. This one, as you can see if you're watching the video, is Restream.io. They're a streaming platform that I really liked for running LinkedIn Lives back when I did more of those. LinkedIn won't let you natively go live from their platform, so you need a platform like Restream, and they do a fantastic job.

So Restream, thanks for the awesome shout. All right. Do you have a question, a review, or feedback for the show? Message me on LinkedIn, or you can email us at podcast@b2linked.com. I would encourage you, please attach a link to a voice recording from you so I can play you right on the show. But if not, I'll take dumb text as well, no worries.

I'm happy to keep you anonymous or shout out your details. I hope you know I wanna feature you. Okay. Without further ado, let's hit it

Initially, LinkedIn launched Accelerate, which was their AI-based ad set builder, and I'm being as kind as I possibly can here, it blew chunks. It basically made me ignore anything that came out afterwards that LinkedIn said was AI-based. I shared in last week's news that Accelerate had been retired, and all of its functionality has now been brought into the classic ad set creation flow, which I think is actually a great move.

I didn't hate all the features, I just hated that it was a restricted way to build your targeting, and it didn't play well with the classic version, which most people are comfortable with It was also the default option, so it confused a lot of people who were just trying to go in and create a normal ad set.

Now we actually get all the same functionality out of it, we just get to choose when we want it and when we don't. So when you're in the ad set creation flow, one of the first steps it's gonna ask you is to enter your landing page URL, then it will go and crawl it, and then AI will come back and tell you what sort of audience targeting it thinks you should use.

I wouldn't recommend it, but we'll get into that here in a little bit. Then when you go into the ad creation flow, you'll now see a button called Draft with AI. When you click that for the first time, it'll prompt you to create a brand kit, where you can give it your logo, your color palette, custom fonts that you use.

You can also prompt it on your brand voice, which should help it sound a lot more like you and a lot less like AI. You can either skip the brand kit or fill it in, but then it's gonna show you a message for about 15 seconds that says, "Analyzing your company page and ad set insights." Then it popped up five intro texts for me to review.

Any that you like, you can just hit accept, or you can edit them first. I would probably recommend editing them first, 'cause they definitely sound like AI, but I have to say it's a pretty slick way to create an AB test, or at least get some ideas for your copy. So that's the intro section. The headline section also has the same Draft with AI feature, although I will say anything it drafts is going to be replete with em dashes and it's this not that comparisons that you are so used to seeing with AI.

But at least it gives you something to start with. For imagery, I know this is something that LinkedIn is actively working on. For right now, it'll just take you into, like, a window of Microsoft Designer, which doesn't currently have AI capabilities, but we know that's coming. As you know, image creation with AI is still very problematic, it's pretty unpredictable, and it's also very expensive for whoever's paying for it.

With Microsoft Designer, you can do some basic stuff, like adding text or adding shapes, throwing your logo in the corner, that kind of thing. For a test, I decided to let LinkedIn take over my targeting and just see what it came up with. When I clicked Auto Targeting in my ad set, it set my audience to all of the US, all language English, and it did actually exclude my own employees, which I thought was a cool touch, it's really thoughtful.

But that was all the targeting it set up. It didn't tell me my total audience size or anything like that. It does have a button down below that says Add Signals. What it wants you to do is give it some signals of the kinds of things about your audience that would help guide it, but it is just guiding.

LinkedIn makes it very clear that this is going to add it to LinkedIn's current insights, and they're gonna decide who they target So it essentially turns your own creation of your ad set targeting into predictive audiences, where you hardly have any control over who sees it. So I don't recommend it. I think it's gonna do like Meta does and just make your audience huge.

But when we're paying LinkedIn prices, I like to be really precise about who I'm targeting. I feel like this could have been vastly improved by simply just asking for a description of the kind of professionals that someone wants to target, and then setting up a very simple combination of their job function with seniority, company size targeting, maybe throw some industry in there, and just do some really basics that I think most marketers would use.

I think that's the functionality that I would've designed if I were being consulted, but of course, I was not consulted. That's the functionality I wish were there that, you know, if I were the one designing it, I probably would've set up, but that's not what we get. We basically get predictive audiences So on the subject of predictive audiences, LinkedIn calls this an AI feature, but I'm pretty sure it's more of a machine learning feature.

We've talked about it on the show before, but it's actually pretty good at taking a seed audience of people who are, you know, let's say your, uh, current and past customers or something, and it's gonna go and find a lookalike of those people on LinkedIn. But it's not just a lookalike, it's more of an act alike.

The reason why I like this so much, I haven't yet experienced a predictive audience that didn't have higher click-through rates and higher conversion rates, and even lower costs than native audiences on LinkedIn, which is pretty incredible. I have found lead quality to be quite poor unless you put some heavy guardrails on it.

But it's a feature like this that when you're looking to just expand your audience and you already have a good idea and a good-sized list of people who are already great, it can be a great way to expand them. Just don't forget to exclude students and retirees. I speak from experience on this one. I think it's pretty well-established that any AI management feature released by the ad platform itself is probably never going to be very good, since at its core, the ad platforms do not have the advertiser's best interest at heart.

We have countless examples of this. But I will say, watch third parties for tools and development here, which are absolutely exploding. One cool example I can share is an MCP released by Phil Ilic from Kyne.co. I went ahead and linked to it in the show note, so check it out. For those of you who don't know, an MCP stands for a Model Context Protocol.

Basically, it's a way for AI to connect to some sort of a system, and then you can communicate with it in normal language, and it can do something for you. So for instance, this MCP that Kyne has built, they connect to your LinkedIn Ads account, and then right inside of Claude, you can ask it questions like, "Tell me how I spent this last week," or, "What ad sets were my highest performing?"

Or, "What's my overall cost per conversion this week?" It doesn't touch anything in your account, it's just read-only, helps you with reporting. But it's pretty incredible if you have just specific questions that you wanna ask it, and then it immediately reports back. It's pretty awesome. I'm super curious to hear how you all are using AI to assist you in your LinkedIn Ads campaign Hit me up about it, and I would love to shout a few of you out.

There have been a whole bunch of fun things that I'm building with AI right now I absolutely love, so I'll share some examples here with you. Uh, number one, reporting has long been a challenge for us because we have our own set of custom metrics that we use to analyze and optimize accounts by. Those aren't metrics that you can create right within Campaign Manager, so we've always had to use some sort of, like, a custom build on the outside.

For a long time, that was just Excel. It was custom-calculated columns in Excel pivot tables. So we built a client reporting engine on a platform called N8N with AI doing the heavy lifting. I jumped into N8N myself, looked around for about five minutes and went, "Yeah, this isn't for me." But I can have an AI agent go and actually do the building for me, and that worked out fabulous.

I'm no developer at all, so it's pretty amazing to see that I've been able to build things that are actually useful, and our clients are actually loving the reports. Pretty amazing. Also, another big pain point of LinkedIn ads account management for me has been wrangling contact lists and ABM lists, uh, to upload right into LinkedIn.

So to be sure, processing lists is not hard work. It's just really simple between CSV files and in Excel moving stuff around. But it was pretty time-consuming, especially if I had a lot of them. So I built an AI agent that takes the data, formats it into LinkedIn's template super quickly. This wasn't hard to do.

I'm sure all of you can do it, but one little challenge came in where I don't wanna have AI process the personally identifiable information. If it's processing an ABM company list, I'm okay with that. I don't think if I send data about companies back to Anthropic, that they're gonna train a model based on it.

But I don't want things like their personal email address or work email or really anything like that going up to the frontier model servers. So what I had to do is create and set up a local model to actually do the processing when it's a contact list. I'm building a bunch of other awesome stuff, but I wanna hear what you're building.

Please reach out. All right, if you're a LinkedIn Ads fan, and you're not already a member of our private LinkedIn Ads Fanatics community, what are you waiting for? We've got four courses designed to take you from beginner to LinkedIn Ads expert. Plus, we've got all the top minds in LinkedIn Ads in there, sharing what they're learning and working on.

Plus, there's an upgrade to get on our weekly group call, where I get to give you direct feedback on your ad sets. I try to answer all of your questions here on LinkedIn as I can, but I always prioritize the Fanatics community. So come join us if you wanna get direct support from me. If this is your first time listening, please subscribe.

We'd love to have you back. And if you're a longtime listener, the biggest favor that you could pay me is to go and rate and review us on Apple Podcasts. It would warm my little ginger heart. All right, with any questions, suggestions, or corrections, which we occasionally need, please reach out to us at podcast@b2linked.com.

With that being said, we'll see you back here next week, and as always, I'm cheering you on in your LinkedIn ads initiatives.