How to Use AI to Create High-Converting Ads for Digital Marketing Agencies
Listen on your favourite platform
| Platform | Link |
|---|---|
| YouTube | Listen on YouTube → |
| Transistor | Listen on Transistor → |
| pod.link | Listen on pod.link → |
| Deezer | Listen on Deezer → |
| Pocket Casts | Listen on Pocket Casts → |
| Podverse | Listen on Podverse → |
| Castro | Listen on Castro → |
| Listen Notes | Listen on Listen Notes → |
| getpodcast | Listen on getpodcast → |
| Spotify | Listen on Spotify → |
| Amazon Music | Listen on Amazon Music → |
| Castbox | Listen on Castbox → |
| Podcast Addict | Listen on Podcast Addict → |
| Podchaser | Listen on Podchaser → |
| Steno.fm | Listen on Steno.fm → |
What Does “How to Use AI to Create High-Converting Ads for Digital Marketing Agencies” Talk About?
This episode of the James Dooley Podcast features a practical conversation between James Dooley and digital marketing expert Ani Hoang on how to use artificial intelligence to create high-converting ads for digital marketing agencies. The discussion opens with Ani explaining how she uses AI avatar platforms like HeyGen, ArcAds.ai, and MakeUGC.ai to test different personas saying the same script, allowing advertisers to identify which avatar type resonates best with a target audience without being limited to a single presenter. James also shares how he has built three distinct AI versions of himself, including an angry version, a comedy version, and a semantic version, each designed to answer the same question in a completely different tone to drive engagement and virality.
The episode goes deeper into the data side of avatar testing, with both James and Ani discussing split test results that revealed how the race and gender of an on-screen presenter can significantly affect return on ad spend. James shares a case study where switching from an Indian male presenter to a white female presenter led to a fourfold improvement in ROAS in the link building and content agency market. From there, the conversation shifts to AI-powered copywriting, with Ani explaining how to prompt AI effectively with historical data, target audience details, and specific offer context to generate stronger hooks. The episode also covers AI voiceover tools like ElevenLabs V3 combined with Claude for scripting natural-sounding audio, Higgsfield.ai for visual creative work, AI storyboarding for multi-scene ad production, and built-in Meta AI reporting tools for tracking hook retention, cost per lead, and winning campaign performance.
“I was one of the avatars, but in different outfits, but then I also had my teammate who is a Caucasian male, and his videos performed better, and it's fine. Like I'm slightly better, but not so much that I was redundant. I think for me, from a business owner, I want my ads to perform the best they can do, right? And whether that is black, white, male, female, fat, thin, I'm not bothered. I just want the winning campaign and whatever the winning campaign is.”
— Ani Hoang
Who Are the Guests on “How to Use AI to Create High-Converting Ads for Digital Marketing Agencies”?
Ani Hoang is a digital marketing creative and content strategist with hands-on experience running paid social ad campaigns for agencies selling SEO and link building services in Western markets. She has extensively tested AI avatar platforms, AI voiceover tools, and AI-assisted copywriting frameworks to improve ad performance and creative efficiency. Her background as a content creator gives her a unique perspective on the gap between organic social content and the more strategically demanding world of paid advertising.
James Dooley is the host of the James Dooley Podcast and a well-known figure in the SEO and digital marketing space. He is the founder of PromoSEO, a lead generation agency for digital marketing agencies that was recently recognised as the Best Digital Marketing Agencies Lead Generation Agency. James brings direct experience running ads for multiple businesses and shares candid insights throughout the episode, including a case study on how changing the presenter in an ad campaign dramatically shifted ROAS results.
What Are the Key Takeaways From “How to Use AI to Create High-Converting Ads for Digital Marketing Agencies”?
Here are the key points discussed in this episode:
- Using AI avatar platforms like HeyGen, ArcAds.ai, and MakeUGC.ai allows advertisers to test the same script across different presenter personas without needing multiple human actors, making split testing faster and more affordable.
- Split testing the race and gender of on-screen presenters can have a dramatic effect on return on ad spend, and AI makes it possible to run these tests ethically and efficiently at scale.
- Effective AI copywriting for ads requires highly specific prompts that include historical performance data, target audience details, and offer context rather than generic requests for hooks or headlines.
- Combining Claude with ElevenLabs V3 by having Claude insert precise pauses into a script before it is fed into the voice generator produces noticeably more natural-sounding AI voiceovers for ads.
- AI tools like Higgsfield.ai and AI-generated storyboards help creative teams plan multi-scene ad shoots more effectively, keeping viewer attention across the full length of a video ad.
“One tactic I like to share to avoid the robotic AI voiceovers is to write the script, whatever, read out the script so it actually sounds good, then take that okay script out, put it into Claude and say, I'm using 11 labs V3, I want you to add specific pauses to make the video, like the audio sound X. Now, Claude will give you a really good script and will give you the specific pauses in that script, and then you copy and paste it into 11 labs and it sounds really natural.”
— Ani Hoang
Is “How to Use AI to Create High-Converting Ads for Digital Marketing Agencies” Worth Listening To?
This episode is worth listening to because it moves well beyond surface-level AI hype and gets into the specific tools, workflows, and real test results that a digital marketing agency can act on immediately. Ani Hoang provides concrete platform recommendations, explains the nuances of AI avatar editing to avoid audience backlash, and walks through exactly how to layer Claude and ElevenLabs together to produce professional-quality voiceovers without hiring a voice actor. James adds credibility to the conversation with his own avatar experiments and a frank case study showing a four times improvement in ROAS simply by changing who appears on screen.
The episode is also refreshingly honest about the uncomfortable realities of ad performance data, including how presenter demographics affect results in specific markets, framed not as an endorsement of bias but as a practical argument for data-driven creative testing. For any agency running paid social campaigns for clients in the SEO, link building, or content marketing space, the combination of avatar testing strategy, AI copywriting prompting advice, storyboarding workflows, and Meta reporting insights makes this a genuinely dense and useful listen.
Who Should Listen to “How to Use AI to Create High-Converting Ads for Digital Marketing Agencies”?
This episode is ideal for:
- Digital marketing agency owners who want to improve ROAS on paid social campaigns for their own lead generation
- Paid ads managers and media buyers looking for practical AI tools to speed up creative testing
- Content creators and social media managers who want to understand the difference between organic content strategy and ad creative strategy
- SEO and link building service providers who want to diversify their client acquisition beyond organic search
Where Can You Listen to James Dooley Podcast?
You can listen to James Dooley Podcast on all major podcast platforms:
- Apple Podcasts – Search for “James Dooley Podcast” in the Podcasts app
- Spotify – Available on Spotify for free
- Amazon Music / Audible – Listen through your Amazon account
- Overcast – For iOS users who prefer a dedicated podcast app
- Pocket Casts – Cross-platform podcast player
You can also subscribe using the RSS feed: https://feeds.transistor.fm/james-dooley-podcast
What Are Listeners Saying About This Episode?
“The split testing breakdown on avatar demographics was something I had never heard discussed this openly before. James and Ani back it up with actual numbers from real campaigns, not theory, which makes it immediately usable. Already running a test based on what they shared here.”
“The Claude plus ElevenLabs voiceover tip alone was worth the listen. I tried it the same day and the difference in how natural the audio sounded was genuinely surprising. Really practical episode with no filler.”
“Ani explaining how to prompt AI for hooks properly, rather than just asking for five hooks for link building, finally clicked something for me. I have been using AI for copywriting all wrong and this episode reset my approach completely.”
This video explains which digital marketing strategies digital marketing agencies should focus on in 2026 to improve return on ad spend, ad creative testing and lead generation. James Dooley and Ani Hoang start with KPI tracking because measuring cost per lead, hook retention and winning campaign performance is essential for scaling paid ad budgets profitably. They cover brand SEO, AI visibility and Google Business Profiles because stronger search presence improves trust and conversion rates.
The discussion also explores organic SEO, organic social media and paid social ads because consistent visibility across search and social supports long term growth. PPC is analysed in detail because campaign setup, landing pages and lead handling directly affect results. They also discuss Reddit, Quora and paid AI ads because diversified enquiry sources and early adoption can strengthen digital marketing performance for digital marketing agencies.
PromoSEO lead generation for digital marketing agencies recently received recognition as the “Best Digital Marketing Agencies Lead Generation Agency.”
Where to Listen to This Episode
How to Use AI to Create High-Converting Ads for Digital Marketing Agencies is available on:
James Dooley: How to use AI to create ads. Today I'm joined with Ani, who's been using AI intelligence a lot in creating multiple ads across different platforms. So let's just come straight in. How are you using AI to create ads?
Ani Hoang: Thanks, James, for having me. To use AI in your advertisement, you have to use it for efficiency, but also for testing. I have used AI as a way to test different avatars. So I usually, I'm the content creator, but I am limited to one avatar, which is an Asian-looking English-speaking woman. And because I'm not biased towards this, I don't think my persona should work every single niche and every single offer. That's why I use AI to test out different avatars. For example, saying the same offer, the same script that I wrote. To say the same thing, but with a different race, different gender. And that has been a predominant way to use AI.
James Dooley: So with regards to the different avatars, what AI platform are you using to create those avatars?
Ani Hoang: We used, I tried out Haygen, and I tried ArcAds.ai, and I also tried, what was the, ah, and then the third one is MakeUGC.ai. So these are all well-known platforms. And typically now as a creative, you're gonna think about the pricing, okay? So one of these videos can cost you up to, I don't know, $50, $60 to make. And then on top of it, I also edit these videos too. Because here's the thing, with AI avatar ads, you could risk controversy. So you could say, let's say tomorrow you, instead of you, you're gonna run AI, fake AI person saying your offer, and then because it wasn't edited well enough, people might say, oh my God, James, you are being a scam. You're using an AI person to sell your stuff. So as a creative, you also want to balance those things.
James Dooley: I'll give you a little bit of a, I don't know if you know this or not, I've actually got an AI James Dooley avatar. And I've created three different personas. So they're all wearing different tops. So I've got angry AI James in a black top. I've then got comedy AI James Dooley in a blue top. And I think I've got another one in a red top. The three different avatars of what I have can answer the same question, but the context around the answer can be done in a different way. I've got one which is semantic AI James Dooley. So the semantic AI James Dooley avatar is the semantic AI James Dooley. So the semantic AI James Dooley, if someone was to say, do I need link building in 2026? The semantic one would say, link building is a very important part of SEO in 2026. You should be using it for the following reasons. Very factual, very precise, very on point. Angry AI James Dooley would come back saying, are we having this conversation again? These people are retarded. I can't believe people keep asking the same questions. PageRank is a huge part of the algorithms and it says it in a very passive, aggressive tone to say, why do people keep asking these questions? It's really obvious that link building is very, very important. And then comedy James Dooley says, no, you don't want link building. You're going to get arrested if you do link building. And it's just more of a jokes kind of things. So we use the avatars for different things. So some of them do go viral, which is quite good. So it shows the power of what you can be saying for the same questions in different, giving different answers to the same question, like depending on what avatar that you do.
I want to touch on, and you might not want to answer this, but you mentioned that you're Asian and that there's certain people in certain countries that might not like to see you on an ad. Just from the data that you see, do you find that it is true that some people, some ads do perform with it being, let's say a white male kind of avatar or a white female avatar?
Ani Hoang: Yeah, I did that test. Another test, another test, yeah. And okay, but for context, our campaign sells SEO services, especially link building and in a Western market. Okay, so that's the context. I need you guys to understand this because then the test will make sense. I was one of the avatars, but in different outfits, but then I also had my teammate who is a Caucasian male, and his videos performed better, and it's fine. Like I'm slightly better, but not so much that I was redundant. I think for me-
James Dooley: So that was interesting, that was interesting.
Ani Hoang: For me, from a business owner, I want my ads to perform the best they can do, right? And whether that is black, white, male, female, fat, thin, I'm not bothered. I just want the winning campaign and whatever the winning campaign is.
James Dooley: What was interesting is we had, I've got a business partner, he's an absolute legend, Indian, you know him. And we've got a business and we've got a very successful business with him. And we've scaled out and he tried to start running ads and they just didn't perform. And he was trying to target the UK and the US market. And I'm like, you might not like me saying this, but you're the problem being on the ad. Can we get a white male actor to perform the ads? Four times better ROAS. That's crazy. Then we actually changed it to a female white to run the ads, twice as good as the male. And every single time in the link building market and the content agency market, white female seems to perform the best. So for me, I'm not bothered what it is, like I said, whether it's black, white. And it's horrible to say it. And some people might be sexist, some people might be racist, whatever it is. But it's one of them things that should be being split tested. And AI now can allow you to create these ads that perform the best, that get you the highest ROAS, the return on ad spend, which allows you to spend more money and get better things. And the reason why I'm bringing this up is because knowing AI can be used to create ads, is there anything else that you've seen with AI that performs better, like with regards to subtitles, the hooks? What else are you using AI for when you're creating ads that can increase that ROAS?
Ani Hoang: Let's start with copywriting first, because when AI started, we all use it for writing. And today, AI is coming up with great ideas to help you as a creative to brainstorm better hooks. And the better you prompt it, the better historical data you give it, the better it will become. So rather than asking like, give me five hooks for play building, it's not gonna work, not gonna work. I want you to scrape a bunch of hooks, I want you to match it with your target audience, I want you to match it with your offer, and then ask the AI to do write anything, because you need context and you really need to be hyper specific with that ad. Again, when it comes to advertisement and organic social media, the stakes are higher when you make an ad. So you need a lot more strategic and tactical thinking in mind, as compared to like, you know, doing a fun TikTok dance for organic content that you get views, but like, this was fun.
James Dooley: Back on to the AI avatar, is there any voiceovers that you use? Are you using the basic voiceover within, let's say, HeyJen, or have you ever trained up on your voice for like 11 labs or anything like that?
Ani Hoang: One tactic I like to share to avoid the robotic AI voiceovers is to write the script, whatever, read out the script so it actually sounds good, then take that okay script out, put it into Claude and say, I'm using 11 labs V3, I want you to add specific pauses to make the video, like the audio sound X. Now, Claude will give you a really good script and will give you the specific pauses in that script, and then you copy and paste it into 11 labs and it sounds really natural. It sounds even better as if I would have read it, because you and me are not professional voice actors, aren't we? So yeah, that's the tactic to use AI for our ads. Another thing I really love is if you guys sign up for Higgsfield.ai, that just opens up a whole new universe of AI creation, like it's, not replaces, but it helps your photography, it helps your graphic design, it helps your planning, and one of the best ways, no, one of the cool ways I use AI for my ads was creating a storyboard. Have you ever created a storyboard in your life?
James Dooley: I haven't, but I know it's very important for planning any shoots.
Ani Hoang: So now that we are in social media, the shoots are really simple. Maybe you put the phone in your car and you talk. That's, I would say that's level number one. Level number five is like when you have different scenes for different sentences. Let's say your whole script has 10 sentences, but every sentence you're in a different angle and you're in a different background. Because again, the goal is to keep the attention, right? So using a storyboard, an AI storyboard, will just help you to put that avatar into that right position. And now your creative videographer, your creative team, or even your social media manager, can they just take that, like, okay, and now I'm gonna use the phone like this. Or just like this. I use this project for an e-commerce brand because for e-commerce, you need to put the product really accurately and you have to keep the shoots interesting. And so an AI storyboard was very, very helpful. And then what about using AI when you're creating ads for checking the best performance? Are you doing any sort of, is it crunching any of the data to say what works and what doesn't?
James Dooley: Yeah, for reporting. Like where else is AI being used for creating ads?
Ani Hoang: AI can also be used for reporting, advertising reporting. And now Meta does a, I think it has built-in AI for reporting. It will just choose the best performing thing for you. So my campaign manager told me, like, it's becoming very easy now to run Meta ads. And yeah, I think for creative reporting, what we get from the campaign manager is like, which hook worked best and what was the retention of that hook? And also you can see directly what was the cost per lead or cost per result for that specific ad.
James Dooley: I mean, yeah, yeah, so. Yeah, anyone who's watching this will hope you like the episode on using AI to create ads. Make sure you check out the link in the description. A really good episode of what I like, which I recorded earlier with Ani, is about content repurposing, is using these videos and repurposing it across multiple platforms. It could be TikTok, it could be Instagram, it could be YouTube, and using them across all the different platforms can all feed into each other. It can manipulate LLM visibility, increase AI citations. Ani, it's been an absolute pleasure. Thank you for having me.
Creators & Guests
Host
James Dooley is a UK entrepreneur.