The Real ROI of AI Agents, Quizzes & Strategic Thinking

Speaker A: Welcome to the Marketing Automation Moment Podcast. I'm Hannah Kelly.

Speaker B: And I'm Mike Maynard. Um, this is Napier's podcast to tell you about the latest news from the world of marketing automation.

Speaker A: Welcome to the Market Automation Moment podcast. I'm Hannah Worley.

Speaker B: And I'm Mike Maynard. Today we Discuss HubSpot's Breeze AI agents

Speaker A: using quizzes to profile customers.

Speaker B: Hannah lets me talk about geeky updates

Speaker A: from Salesforce and we discuss why strategy is a must for marketing automation. Hi, Mike. Welcome to another episode of Market Automation Moment podcast. I mean, do you notice anything different about your surroundings today?

Speaker B: Oh, ah, it's so exciting. We've actually gone hired a proper podcast studio to do some recording face to face rather than being remote. So it's great to be with you, Hannah, for this and I'm really looking forward to this episode.

Speaker A: Me too. I mean, it's a bit unnerving to have three cameras facing us, but I'm getting used to it. So let's kick it off because we've got some really cool things to discuss. The first is an update from HubSpot. Now, it was really interesting because HubSpot basically released this article talking about their Breeze agents. Now, it's classic doing a different type of title for what is basically an AI agent, but we love market automation's platforms trying for it. Now, it was quite interesting because the thing that they were really driving hard in this article was that, uh, actually 50% of conversations can be solved by their AI agents. I mean, you had a read. What did you think?

Speaker B: Yeah, I mean, thanks for sending through the article. It was really interesting. Obviously HubSpot's not alone. Everybody's trying to generate these AI agents that take work away from humans, make it be done by AI. I think, you know, always the question is, is how effective are they? And certainly talking to our clients who tend to be B2B companies, they tend to have very technical products. They are usually quite cynical about the ability of AI agents to solve problems. And I think the perception is, is that quite often AI agents are great at solving simple problems, for example, quite often with humor brands. But when it gets to much more complex things, you know, particularly if someone's, for example, coming onto a website and asking for recommendations about which semiconductor to use, maybe an AI agent isn't the right way to go. I don't know. I think, you know, increasingly we are going to see B2B companies starting to use these agents. And I think, you know, the more B2B companies that experiment, the better. But I mean, the problem is, as you and I know, there's such a huge cost that's involved in training the agent. Don't you think that's really one of the biggest issues for B2B?

Speaker A: I think that's a brilliant point, Mike. And actually, you know, we talk about HubSpot. I actually use the HubSpot AI agent, which they use in their own platform the other day, and I got so frustrated with it so quickly. In the end I had to be like, speak to a real person. Speak to a real person. And I was actually asking a really simple question around their different packages and the pricing of things. And so I think it's a really valid point that, yes, they can help, yes, maybe they can cut back. So it's 50, it helps 50 of conversations, but it has to be those easy ones. And even then to get it trained up to that point is going to take a lot of work and it is going to take a big time commitment. Once that time commitment's been made, then absolutely, you know, I'm sure it will help, but I think there'll be some frustrations from customers if these companies are just straight away switching to these AI agents and actually they're not getting the answers they need.

Speaker B: I mean, I completely agree. We obviously work with a client in the space. And I think, you know, one of the most insightful things they said to me was there's actually two types of companies when it comes to deploying AI, uh, agents for customer service. The first is all about saving money. And so they just want to move conversations from humans to AI and they want to save money that way. And they said that generally doesn't work. What works is if companies are focused on using AI to actually make the customer experience better. And I thought that was really insightful because then what you do is you actually have your AI agents and they're really focused on doing the things they do well. And then also the company is not trying to hold customers like you were with HubSpot, hold them on the AI agent. They're very happy to let them go to a human if that's the best customer service. So I agree with you. I think it's an interesting trend, certainly something I think B2B companies should be watching. But I think, you know, in the long term, um, probably what we're going to see is we're going to see a fairly slow rollout in B2B because it does tend to be more challenging.

Speaker A: Absolutely. Well, let's move on because I want to have a quick chat about ActiveCampaign. Now, I was scrolling through their webinars. I do this when we're looking at our podcast recordings, and I came across a really interesting one. Now, it was talking about using different ways to basically get data. And I mean, it had the basic things like, you know, gate your content if you're on your Contact Us page, maybe use a form there. But the one thing that I found really interesting is it actually talked about using quizzes. So using quizzes that are maybe a bit more fun, that have a resource at the end, but actually help provide the data in order to be able to personalize. I mean, what do you think about that type of concept?

Speaker B: So I think it's really interesting, you know, I mean, when we started with marketing, uh, automation, everybody was super excited about progressive profiling. Let's keep asking more and more questions, and then we can personalize. I think the idea of activecampaign to kind of gamify it. Yeah, I mean, that's not a bad idea. Although, haven't you seen too many quizzes on Facebook? I mean, I'm kind of bored of them, and I don't even go on Facebook that often. I think, you know, if you're going to create a quiz, the challenge is not how it's going to help you as a marketer personalize the content, but much more how that quiz is going to help your customer or your prospect learn something. So, I mean, we do something slightly different at Napier, um, and we're very into creating, you know, complex analytical tools that really understand customers problems and then produce customized reports. So we've done that quite a lot, you know, from, from an abm, uh, evaluation tool to decide your ABM maturity all the way through to helping, uh, some of our clients develop tools that lets their customers see, for example, how well they maintain systems.

Speaker A: Yeah, I mean, I think the thing I would add to that, Mike, is yes, maybe we've seen too many quizzes, but I think me as a marketer, it's sometimes just a little bit of a relief. You're like, oh, okay, cool. Something just to take off this hard work, you know, that I've got going on during the day. But I think also as well, it can be done in smaller ways. So, I mean, one of the things that we do is we use Turtle. So Turtle is an interactive content planner platform and similar to quizzes, but we actually put in polls. So we put in polls within the pages related to the content to say we're talking about marketing automation. It's like, you know, what percentage do you have of people that convert from your emails? You know, maybe something like that. And we actually get people to fill in the data. And again, it's progressive profiling, but it's actually a really easy way where they probably don't even think twice about doing it. And so I do think there's actually something not unique, but just a little bit different that makes you go, okay, I'm gonna give you my data. Why not?

Speaker B: Yeah. And maybe the other thing I. You kind of alluded to it when you said quizzes are quite nice to have a break. I think the problem is that, uh, as B2B marketers, we're always a bit too serious, and we always want to know technical information and data about your plant or whatever it is. Um, and actually, sometimes we should look at being a bit more fun. I think it's interesting. Cambridge Analytica, and obviously what they were doing was not necessarily a good thing, but the way they were approaching profiling people through questions like, which Game of Thrones character would you be? I think having that fun is the way to do it. So maybe what we need to do as, uh, you know, marketers in B2B is we need to be thinking ways to create fun quizzes rather than necessarily always be too serious.

Speaker A: I think you've summed it up perfectly, Mike. Now let's move on, because I want to have a chat about Salesforce, and this is going to make you laugh because I read this Salesforce article, and it's basically an update on what they're doing to their AI, their more technical specifications in the back end, and I didn't understand a word of it. And I think that solidified that you are the geek and I am the marketing expert. So I don't know, do you want to have a quick chat about that? Because I think it has some really useful things in there. I just do wonder if it's going to go over, uh, B2B marketers heads a little bit.

Speaker B: Well, I mean, I think anyone who listens to this podcast knows that you are really the marketing automation expert. You're finding all the stories, and I'm just here sort of, you know, giving my opinion on them. But I really appreciate you finding a geeky one for me. I think, you know, this is really interesting. I was actually listening to, uh, you know, somebody else on a podcast, and they said that they don't believe that, uh, AI in terms of large language models are really going to move forward very much. And actually, if you look at what's happened you know, honestly, the jump between ChatGPT 3.5 and 4 is way bigger than 4.0, uh, and all the other 4.1 and all the other different versions we've seen now in ChatGPT. So, yes, it's true. You know, large language models are definitely plateauing. And what they said was the way things are going to improve is that you're going to have software around the AI to kind of control and, uh, manage it. And if you look at what Salesforce is doing, it's doing a lot of software around its agents to really monitor, evaluate and control. And I think that's the idea behind it. Super geeky. Totally agree with that. But I do think it makes sense and I think it actually reflects really the trend of what's happening.

Speaker A: So let me clarify that, Mike. So if we're going to take it out of the geeky terms more into marketing terms, are you saying that they've put in this software that's going to manage the AI and then that's going to help basically how to improve the AI as the platform moves forward to evolve?

Speaker B: I think you put it perfectly there, Hannah. Uh, I mean, another way to look at it is at the moment, anyone who uses AI knows AI needs a boss because it can be lazy, it can get things wr. Not really very reliable. And typically today that boss is a human. If you can put software in to kind of make sure that the AI stays within certain guardrails, then actually that boss can be a computer. So I think what they're doing is they're installing, you know, certainly at least the line manager to be a computer for the AI. And that's really going to help improve the quality of AI and reduce the amount of work people need to do to actually get results from AI.

Speaker A: Well, I love how you describe that, Mike, and I think it's something that marketers are going to absolutely love. So thank you for explaining it in not geeky terms because I really appreciate it. But let's move on. I really want to talk about our insightful tip of the week. Now, we've talked about this and I would say, uh, many, many months ago, and I think it was due a refresh. So I want to have a chat about strategy. Now, strategy is something that we always talk about, Napier. We have a four step process. We have our first two steps that are literally focused on strategy. Who you're trying to reach, what are your messaging. And I really wanted to talk about, when it comes to actually working with a market automation platform, why do you need strategy in place. Why should you have a strategy in place of how you want to use that platform? Because we've unfortunately seen many companies just spend thousands of pounds a month and not get any return of investment for it. So what do you think?

Speaker B: Well, strategy is easy, isn't it? It's just two emails a week.

Speaker A: Done. Yes.

Speaker B: I mean, in all seriousness, you've seen it, I've seen it. You know, people who really don't think about the strategy, so they've invested in marketing automation or they've bought in a marketing automation team and then the goals are all outputs. You know, what are we going to do, how many emails are to send, how many form fills we're going to get. And they're important, but they're not strategy. And I think strategy is always the why, you know, why are you doing it? What do you expect to get as a business result from it? Mhm. And so, you know, you pointing out the Napier four step process. Absolutely. I love that. It's a great point because what you need to do is sit back and understand the situation. That determined phase we do at the start. So understand what you want to change and how you want to change it and then you've got to do the focus stage, you know, looking at who you're talking to and what you want to say to them. So the audience and the message and do do that before you rush into the tactics. So really it's simple. Take a breath, step back, don't think about what you're going to do, but think about the strategy first.

Speaker A: Absolutely. And I want to add something to that Mike, because I think one of the things that gets overlooked is that relationship between the sales and marketing teams. Now I love how you describe the determine and the focus stage there, but I think the key thing as well is that when you take that breath and you take that step back, you are working together as one. Because as a platform, market automation platforms are amazing to support both sales and marketing teams. If you've set up a platform and it's just for marketing's goals, it's not going to work. If you set up a platform on the other side and it's just for sales goals, again, it's not going to work. So I just want to put that point because it's so important that actually when you take that step, make sure you're working well together internally because the results you will see will be double because you're all going towards that one unified goal.

Speaker B: I think that that's so insightful, Hannah. And I think if you're looking at it from both the sales and marketing point of view, inherently you've moved above an individual department's KPIs. And you're absolutely right. It's a great way to start off is to make sure you're covering that whole customer journey.

Speaker A: Absol well, thank you for another interesting conversation today, Mike. It's been so great to do it in person.

Speaker B: It's been fun. Let's hope we do it again like this.

Speaker A: Thanks for listening to the Marketing Automation Moment podcast.

Speaker B: Don't forget to subscribe in your favorite podcast application and we'll see you next time.