AI Marketing Masterclass: Practical Use Cases and Real Results

Okay. If you want to grow your business in 2025, there are four key stages for you to get right. The attract stage. How do I attract people to my business? How do I convert them into something meaningful? How do I close them into dollars? And how do I. We are going to give you the practical AI use cases to implement in every stage, inclusive of the tools and real world examples from HubSpot so you can create the right AI winning plan in 2025. Let's get into today'. All right, Kip, we are going to give people the entire AI go to market journey that they should build. And the way that we are thinking about AI is basically across the modern customer journey from attract, convert to close, delight. What are the impossible problems that used to be things that we just could not do that are just now really hard and are worth doing with AI? And in this episode, you and I are going to go through these different use cases, give some examples, give some tools and and actually tell people some real results from our work here to date.

Kiran, two things. When you say really hard, what you really mean is very possible, but with a lot of iteration, right? It's not like really hard, like you're gonna have to like work for years and years. It's like, hey, within a couple months you can have a reasonable solution here, right? And then you're gonna iterate and build on that solution as you figure it out. And what's wild to your point is that these were just fundamentally not possible like a couple years ago, Right? And that's what we're trying to push everybody on today is invest in the time and energy to make things that were used to be impossible possible today. Because that's where you get a lot of leverage and growth and results in your business. The second thing is, Kieran, the slide that you're showing, I think is the simplest, cleanest articulation. If I'm running a business around, oh, how do I tactically do and use AI to attract customers, convert those customers into qualified leads, close them into customers and then delight them through a really good customer experience, support customer success, those types of things. Because what happens right now is we get all these AI vendors out there talking about bits and pieces and talking about the technology. What we're doing here and what we're going to try to do today is obsess about the use cases. So why don't you walk us through the some of the use cases that you've outlined here and then we're going to show some examples of how folks can go and do them right away. Right.

Let's start with how we attract customers. So there's two core places I think we want to start implementing AI. The first is actually in paid advertising. So paid advertising. You know, we typically have these AB tests where we test the control, which is the ad that is in the wild, that is producing clicks, is producing customers for you. That's what we call the control. And then we do variations of that control, which means we, you know, vary in the ad creative, maybe the text, maybe the imagery. And we do that second version, a third version, a fourth version. And we can run these tests through platforms to say, hey, what version is working best? And then the platform will pick the best version. The interesting thing about AI is AI will allow you to do that at such scale that you have not ever thought that was possible.

Right.

It's going to. I like to think about going from a B test into A to Z testing, because the AI can do many different creative versions. So really what you do is you teach the AI a template. You basically say, this is a template of the kind of ad I want to produce, and the AI will produce all the variations. And our friends over at Ramp have a, I think an investment or a company that they're excited about in the space, and we should show a quick. You're going to show a quick demo of that?

Yeah. So what happened is, you know, we've been talking about these things. You and I care. And then I saw our friend Eric Gauman. We both huge fans of Eric and the company he and Kareem are building at Ramp. Ramp is an awesome business and they invested in an AI ads creator called Icon Karen. And what's really interesting is what, what you can do is you do some testing. I recorded one video, made it into 20, and then we found this was the best video. After some testing, here's how it works. Start by entering your brand and product name. We'll then identify your target customer and set your product categories. With this info, we're going to match you with the right creators for your brand. Key signals like number of followers, engagement rate and audience demographics are taken into account. Once matches are made, creators record one video, which is used by Icon to make many videos. That's pretty magical, right, Karen? Where you can basically take one video, they're even matching you with a creator to make that one video. Then. Then what they're doing is making 20, 50, however many versions of that video and then testing them all against each other to see what performs the best. Well, they and it's a perfect example of what you're talking about.

They did something actually even different than I had been thinking about with AI, which I think if we can go back to, we can go back to start to check. But what I think he was, he said, we tested a bunch of ads to begin with, found a winning ad and then took that ad and added a creator to that ad. So first of all, they used the ad test and defined a ad format and then they use that ad format to match with a creator and then they ran variations of that ad with creators. So he's, he's actually doing two core things there, I think, which is pretty cool.

What's unique about what icon's doing? It's not just the A to Z testing, but it's bringing a creator in for a trusted voice as part of that creative so that you, you have somebody who people are familiar with in, in your industry in your space because they're using AI to match your, your product with creators. And then you can run ads in this kind of A to Z way that we're describing and get much better ad performance because of that. Because you've creator alignment and you have 20, 30, 50 variants you're testing versus two or three. Because before you would have had to find this creator manually, you would have had to email them, get in touch with them, then you would have said, cool, can you make a couple videos about my product? And then you would have tested those couple videos and that would have been it. And all you would have been able to do is kind of test titles and, and intro edits and captions and that type of thing.

What happens in a lot of cases is these creator, the creator companies match with creators and allow that creator to work with brands in different ways, whereas they're going to just build for a single channel and match the creators themselves. So that's a version of the, the kind of creator matching and working with brands I hadn't thought about before. But it makes a ton of, it makes a ton of sense. I do wonder long term, like how many of those creators are going to be AI generated creators, because that is a whole other space.

You're starting with human creators and eventually those will become AI, you know, creators, AI influencers and avatars for your ad variations. And this is, they build it in a way to make that transition very smooth, which is probably very intentional on their part, I imagine. We haven't talked to Keenan, but I imagine that's the case.

If you want to get started there, I think the takeaway for the audience is you should definitely try some of these AI tools that allow you to run creative ad tests and at scale because there is a real lot of value in doing that. And then obviously if you have a, if advertising is a real core part of how you run your attract phase and how you acquire new demand and you ha. You're in a space where there's a bunch of creators, I think that's a really great next step is actually figuring out if you can use a platform like Icon to attach your paid advertising to creators and see how well those ads run. I can tell you from, you know, zapier days and even some of the things that we do in HubSpot those create those ads through creators tend to perform better than anything else because it really brings in an authentic voice.

Completely agree. So you can go check out Icon. There's other people in this space but we wanted to show one example of how you could go and do this today. It's like an actual thing you can go and do. It's not us just being, oh, we'll, we'll be able to do this cool thing in the future. Hey everyone, quick stat for you. AI usage has increased 53% since last year. And you know how I know that? Because I just read our brand new 2024 AI trends in marketing report. This report is packed with original data. That data not only highlights the latest in AI trends, but also shows how businesses today are using AI to grow their business. You'll learn the top five AI tools for marketers, exactly how teams are investing in AI and where they're seeing the return on their investment. And you get a data backed answer to the big question, is AI helping or hurting my marketing career? I invite you to steal and take all of our hours of research for free. Grab your copy from the link in the description below. Now back to today's show.

So the next thing really in the attract phase is we are, we have been a very text driven bunch of marketers and the reason that we're very text driven is because it's really hard to create video. And video is much more impactful than text in a lot of ways. Especially when you start to think about search is going to be heavily disrupted, which is the kind of core reason that most people produce text. So you have these two things happening at the same time. Google being disrupted, search being disrupted, and then video becoming much easier to create. And these channels that are growing, YouTube, Instagram, TikTok, all these channels that are growing much better, better for social video. So we believe the next thing you want to invest in is how do I integrate more video into my marketing mix? And AI allows you to do that.

I am going to show Captions AI today because Captions is one of the hot companies in the AI video space. And basically you can generate videos with AI so you can do text to video, which is kind of the core part of this. And so all of these AI video companies are trying to position themselves as, hey, you don't need a video camera anymore, right? The video camera is obsolete. You can generate video with text. You can select editing styles and everything and do editing really quickly and cheaply, which is also important. The other thing is global potential. What we found at HubSpot is that AI video and AI dubbing of languages and videos is very, very good, very cheap, but very effective in terms of we can take one video in one language, put it into multiple languages for very low cost, hundreds of dollars and it performs very close to what a, you know, natively recorded and produced version of that video would be. And so the global use case when you're talking text to video is very, very important.

This is one of the areas that I cannot wait to till there's like a real unlock. I know I call this year the year of Video. I do think that is somewhat the case because we have seen a huge increase in the amount of video tools, video capabilities. I'm going to show a. I think I'm allowed to show this. So I just will say I bookmarked this and then I have not looked into the details because it is a Chinese AI model. And I know there's always like it's. And it's also on X. So whether this is misinformation or not. Yeah, but you know the cool thing about. So it's called Minimax, but it's a text to video model. And the reason I want to show it is because, you know, the world I want to live in is not just when I can bring my ideas to life easily through video, but can you imagine when video tools are good enough that you can go from text to video pretty accurately? Wouldn't, wouldn't most of us just not go then to like Claude and talk about the kind of TV show that we like and then just cut and paste the script from Claude into text to video and produce our own tv? Let me show you this. So this is like text to video. Again, no person involved here, just the AI model taken text to video. And this is a little, a little story again, if there's like mad stuff in here. I haven't looked at it. Don't blame me. Wow, right? Pretty unbelievable.

It is wild.

First of all, like business idea. If you can upload your, your child's.

Photo and then character.

Yeah.

What we do next, Captain.

All right, so you get it. I just think that, that that is going to be a total unlock for market marketers. Like not just marketers. Like human civilization is going to change, I think a lot when you can just go from text to video. In terms of how we consume things. I don't even know how disruptive it is going to be for the average TV show and the average movie. I, I can't even wrap my head around the fact that you could, I do think you could upload characters either from your family or from whatever you want and create TV shows around them. Let's move into convert because we have a fun thing for convert. So this is a huge unlock this chat text, AI multimodal. And I want to explain this with a little demo. We have the amazing Amanda Kahlo coming on the show. She's ex founder of six Sense. She now is the CEO of this very cool company called One Mind. I'm sure she won't mind just doing a sneak peek of One Mind. She's going to come on and actually demo it in full. We're going to do a quick little snippet, a little tease of One Mind because Amanda is going to come on and do the full thing and we're going to have Amanda's a avatar. Come on with us. Why am I showing you this? Because I want to quickly show you an interaction and then tell you where I think chat is going to go. Hey, Mindy.

Hey, Kieran. Great to see you. How's everything going over at HubSpot?

Hey Mindy. I would love to know how we can make our chat a lot better using One Mind and an AI avatar.

That's a great question. Buyers want a human like experience and our AI avatars are designed to provide just that.

All right, so let's leave it to Amanda comes on. So she will give us an incredible overview here. But what I'm really excited for here is today chat is very much like a text based interaction and you either answer the question, a support question and log a ticket or you say, hey, this is a sales thing and I'll book time with my rep. Or it's such an uncomplicated conversation that you can just have that person upgrade touchlessly. That's kind of how all chat Works. What AI will, I think will allow us to do is to go multimodal and do more of it, more of that and do more of the sales process all in one interaction, right? So you can imagine being able to go in here and start with text and then that person can say hey, which is an AI avatar. Hey, why don't we transition here to, you know, I don't even know you have to say transition but like pops up on the screen and says hey, let me show you some visuals, right? And, and then can start answering your questions. So what one might allow you to do is train on anything you want to train on. Your customer calls your visuals and can always bring up the right visual for the right question. So you basically can do the entire pre sales part in the chat experience. And that means everyone can kind of get a walkthrough of your product, all of their questions answered and do a lot of the pre qualification before they ever have to talk to a human. Now there's a whole other sidebar here which I think we might want to pause and wait for Amanda to come on, which is, and the thing I, I'm really curious about because we're going to see if we can pilot some of this is let's say you had two experiences side by side, right? You had a world class experience because the avatar understands everything about the product, the service, the company can answer perfectly. All of your questions is pretty much like none. What's the right word? Like isn't in your face, you know what I mean? Like it's kind of just answering like and, and does open in the questions but in a none like kind of aggressive.

It's reactive. It's reactive versus super, super proactive and aggressive.

And so you have that perfect experience. And the other side you have typical, you have a sales experience, right? And they're, you know, some are great, some are not great. And that person might not know all of the questions. They might say, hey, I have to do a follow up call, I have to do a follow up email. I'll have to get the sales engineer to come on. I'll have to do these other things. But it's a more human experience, right? I still don't know what people will think because this avatar is so new. I think that's one of the more interesting things.

If you're watching, put your comments and questions, if you have stuff you want us to talk about with the new avatar, chat with one mind with Amanda, post that, post those below. We'll bring those up in the show where if you would be fine with it or you would hate it. Post those in the comments below. We'd love to hear from you.

So that's gonna be a great episode. I would combine so I. So I think I've called these two things out separately. Checks text to AI multimodal. Multimodal. The ability to go from text to video. I think you can also just do audio because if you didn't want the video to be on, the person could just chat to you in audio. The avatar product demo is not too dissimilar.

Right.

You can actually imagine a world where the avatar can walk you through your freemium product, can walk you through the within app, can walk you through your product out of app, can actually do a full demo, which means the person would come to your website, you would say hey, get a product walkthrough. And the avatar would walk you through the product and be able to like answer questions. This is a really good example of where you can do things with AI that you traditionally have not been able to do. Like an impossible problem because the unit economics are not there. You cannot afford to give everybody a complete demo and walk customized walkthrough of your product. In a world with AI, you can. And so again the original like theses of Inbound was the consumer is way more in control of the buy in cycle and you just need to provide more value. That is never going to be more true in AI because I can literally get the entire walkthrough of the product and customize to the way that I want that walkthrough to go and the questions answered via AI. And so there are two things that we are working to pilot. The chat part in the product demo on the email. I can throw it over to you Kip, but that's like we've had some pretty great results in terms of we B2B marketers have traditionally in this convert stage, personalized email by segment and AI allows you to personalize much better towards like a one to one type experience.

Yeah. Earlier the year when Kieran was out on parental leave with baby Hope, Emmy, Jonathan on our team and I sat down, we walked through like a whole email experiment where basically we used AI to generate customers email content and content recommendations. In this case it was like a course on HubSpot Academy that we thought would be most relevant to that person. And so we created a custom email for every single person versus like hey, we think these 10,000 people are going to broadly care about this thing and we're going to use Some personalization tokens. And that it, that's it. We saw our conversion rate from the hyper personalized email increase over 80% and so it was very, very dramatic. And literally what that is doing is passing some information at that time, at that time. And we're still doing it now to OpenAI and this case right now we're doing GPT4O to basically create those emails in OpenAI, put them back into HubSpot and then send them through the HubSpot email workflows product. So you can do do this with any of your email marketing requires some technical work to actually do the integration with OpenAI. Get the Content up and running and you're going to have to do a bunch of testing and iteration. Go check out the episode with Emmy if you want like the full deep dive of exactly what we did. The technology involved all those things. But here today you can have deeply personalized email that converts at a much higher rate than ever before. And I think that is going to be a huge part, especially the next year or two of this kind of transformed AI led buyer's journey that we're talking about.

All right, this one here, virtual bdr, I just changed on the fly. We had to call that email. It's not really email. So let me kind of go through this one a little bit. So today, and a lot of B2B companies, you kind of acquire new contacts into your database and then you say, okay, well we have some sales reps who will sell into that database and we will have some sales development reps who will look in the database for good fit best companies, reach out, do some outreach to them and try to book meetings for those sales reps. So the sales reps are not having to do all of their own prospecting, right? That's a way a lot of scaled B2B companies work. And those business development reps, they are, you know, really talented, but they can't get the entire coverage across all of the companies in your database because again, when we think about scale, humans always have scale problems. Like it's just unit economics. So you can try to use these business development reps against your best fit companies because that's where you're going to get most return on your investment. What's interesting about AI is AI is starting to be able to do all of the components of that role, right? So if you think about every role, when you think about AI, you should just break it down into a list of tasks. That's the best way to think about jobs. That that task is how do I research companies, how do I find good fit companies, then how do I get enough data about them to craft really great emails? How do I actually do follow up calls with the right person? There's some date, there's some company mapping to find the right decision makers and then there's probably like some social outreach. You create some sort of sequence where you kind of email, you call, you do some social outreach, you get the right people, the right data, the right companies. So AI is able to prioritize companies pretty well like identify what are good for companies with all of the data we give it. We've already just talked about the fact that I can craft and very personal emails to be able to outreach to not only the, you know, one person but going to cover all of your decision makers and actually can personalize them to each decision maker. If you give it the right data around those individuals, it can then I think at some point do lightweight calls. Right. We talked about the company that I've talked about before called Ayres AI. There's another one called Bardin AI. Um, these are voice AI enabled apps but there are starting to be some really good voice enabled agents. We've talked about them previously. They're not there yet. They still can glitchy, they can still don't sound human like at times, but I suspect at some point we'll move from text to calls. I don't ever want to see social outreach because I don't like thinking of AI mass saturated in people's social channels. I think that one should be much more human led and you can craft really great sequences across those three things and then I think over time you can integrate and I want to throw it back to you because traditionally video has been one to many because they're so costly to make. So they're kind of just when we're trying to close a customer. They're just like one to many product demos or one to many videos. There's a company called hey Gen, we're going to speak to Josh from the main stage of Inbound this year. He's doing the podcast live and they are really going to transform video to be the spray to spray and pray to Context Driven which means you can treat it much more like an email and integrate these dynamic parameters to customize it for the individual.

Part of why we're talking about hey Jin Kieran and why we know them the most is they just build a pretty sick HubSpot integration so you can get all of your data out of HubSpot which is what you're talking about, right? Like, you can have all your CRM data and instantly turn that into a personalized outreach video for your avatar. So, like, you and I can record a video, one video, and then it can use all of my CRM data to personalize a hundred, a thousand, however many versions of videos that you need. Right. So if you look at this video here, it's like, get some information that's in the CRM and you can take that information and dramatic, you can take the information provided by your prospect or customer and you can give them a very specific personalized video. So this is kind of like the A to Z ads thing that we talked talked about in the attract stage, but for the kind of deal closing and influence stage of the buyer's journey. And it's all about using text to video in that use case. And I think it's actually one of the best use cases for text to video because at this point, people are not used to getting personalized videos in their inbox for like a very focused thing. Oh, it's like this is about my company and how I'm thinking about your product. This is very niche thing. I would have never expected it. So one, they really like it and it really resonates well. And the second thing is like, they're a little bit more forgiving of any errors or any maybe lags or anything from a technology perspective because it feels so magical. Right, Right.

So you can imagine the, you know, the world of the business development rep actually has an agent or agents that they manage for them with their, you know, their actual personality, their likeness. And it's actually doing this thing at scale. And that's able to do the text, it's able to do the voice, it's able to do the personalized video through. Hey, Jin. And again, these are all use cases. We've mapped out the kind of use cases that we're bullish on, that we are invested in. And this is one of the ones that we are also investing in because they have such a killer integration with HubSpot. Okay, I'm gonna round out with two really quick ones. One of them is like a really cool use case. I shared it on LinkedIn the team. I just love this one because it's such a big problem for SaaS. Businesses like churn is the killer of every great SaaS business. Every SaaS business deals with Churn. One of the big things that is such a, you know, mundane problem, but a really hard problem to solve is when a company Churns, they don't really tell you why, they just churn, right? And churn means that they stop using your product and for the most part when they stop using your product, they're out. And there was some reason they were out, but you don't know. And if you knew the reason, you can make your product much better and over time save customers who have similar problems. And so what the company did is they built a little lightweight model that was able to go through the emails and the calls of all of the customers who would churn and left no reason and figure out a reason for them, for them churning, which means they, you know, when they log support tickets, when they emailed a customer support rep, when they were on calls with the customer support team, there was a thing that they were frustrated with, there was a thing that they were not happy with and AI was able to pull that out. And then over two thirds of the customers who originally had no reason for churning now have a reason for churning. And that means that HubSpot can actually go be proactive and fix those problems. I think that, that those kind of use cases of AI are so impactful and so I love that one from the team. The, the last one I'll end with is support.

Right.

Support is, we talked about it before is like a big part of your brand. Like it's how people really get turned off by your brand. In a lot of cases, if you have great support, it's something people are happy with. But if you're a terrible support, it's something that people will leave you over. And AI is going to be able to give companies their coverage. And this one's been covered to death. So I'm not going to spend a lot of time, but over 25% of all of our support today is managed by AI, which is equivalent to more than 100 heads. And so you're going to be able to give everyone the support coverage that they deserve. And I think that is actually going to become a net positive for your brand. So that is a, definitely a place that you want to invest in today.

Wow. We just covered how you can use AI every step of your buying and customer success process. We gave you a ton of real life apps that you can go and use as well as some summaries of some experiments that have worked really well with some real data. It's pretty incredible, you know, and I think the overall theme for me, Kieran, I want to know what you think is that we've gone from like a lot of these being Hypothetical things that maybe you can do to very tangible things that most forward thinking marketing, sales, business leaders can go and do now.

Exactly. There are now practical use cases. There are now data. That's what we want to share with you all. We want to share some of the, wow, look at these things happening. But we also want to share, get into the practicalities. Like we shared an agent with you, with Dharmesh that now has over a thousand runs and people are getting a ton of value. We're going to release more agents over the coming months. You can just go use these things. We're going to give you the practical use cases with the data. And I think my call to action here is if you are at Inbound this year, we are going to be on the creator stage. We are going to do a deep dive on AI use cases. Some personal, some fun and some of these. And so you should definitely come join that stage. If you actually enjoy this stuff, be.

On the main stage with Josh from Heygen. We got all kinds of stuff. So if you are coming to inbound 2024 in Boston, which is our big event at HubSpot, you're going to see Kieran and I on a bunch of stages, but also walking around, please come say hi, talk to us. We'd love to chat with you. And until next time, this has been marketing against the grain. We'll see you on the next episode. This data is wrong every freaking time. Have you heard of HubSpot? HubSpot is a CRM platform where everything is fully integrated.

Whoa. I can see the client's whole history. Calls, support tickets, emails. And here's a task from three days ago. I totally missed HubSpot grow better.