Loading summary
Transcript70 lines
- [00:00]
Kip Bodnar
I think if you look at the last year and a half, it has felt like a decade's worth of innovation in about 18 months.
- [00:07]
Phil Agnew
That's Kip Bodnar, HubSpot CMO and expert on all things AI. Today, he'll walk through how AI can change your business and four tips that you can follow to make sure you don't fall behind.
- [00:20]
Unidentified Host/Interviewer
All of that coming up.
- [00:23]
Phil Agnew
Apparently most businesses only use 20% of their data.
- [00:27]
Unidentified Host/Interviewer
That's like reading a book with with.
- [00:29]
Phil Agnew
80% of the pages torn out. The point is, you will miss a lot unless you use HubSpot.
- [00:35]
Unidentified Host/Interviewer
Their customer platform gives you access to.
- [00:37]
Phil Agnew
The data you need to grow your business. The insights that are trapped in emails, in call logs, in transcripts, all that unstructured data can really make a difference to your business. Because when you know more, you grow more. You won't learn much reading 20% of a book. So why settle for just 20% of your company's data? Visit HubSpot.com today to get Learn more My guest on Nudge today is both an AI expert. He studies AI for his podcast Marketing against the Grain, but he's also a practitioner applying AI in his day job as HubSpot CMO.
- [01:13]
Kip Bodnar
Hey everyone, I'm Kip Bodnar. I'm the Chief marketing officer at HubSpot. I also host a podcast called Marketing against the Grain. I'm excited to be on the Nudge today.
- [01:20]
Phil Agnew
Kip was quick to tell me how rapidly AI has developed.
- [01:25]
Kip Bodnar
It has felt like a decade's worth of innovation in about 18 months.
- [01:29]
Phil Agnew
Specifically, he called out three ways that AI has evolved.
- [01:34]
Kip Bodnar
1. You have the core frontier companies, OpenAI, Claude, Anthropic Perplexity, all of them. The models have gotten a lot better and a lot smarter and so they're able to do a lot more things. Models are able to reason and really think through complex problems and become much better at strategy competitive analysis. A lot of like the core business use cases that you would have hired a consulting firm or something to do in the past, these LLMs are now very good at that.
- [02:02]
Phil Agnew
In 2020, three professors at Wharton School, University of Pennsylvania tested ChatGPT on the MBA core exam. It tested the 3.5 model and found that it scored a B to a B, basically good enough to pass. Interestingly, it actually struggled with the math heavy analysis, but did very well. Perhaps unsurprisingly on stuff like case study and writing later GPT4 models performed. But this. Well, this is just the first way that AI has evolved.
- [02:33]
Kip Bodnar
The next thing They've gotten very good at is image and video generation. And so going from an idea to kind of a prototype, a wireframe, a mood board, early creative assets or even finish product creative assets with a little polish and editing, you can now do much faster and much cheaper with AI.
- [02:53]
Phil Agnew
One recent study from the researcher Nightingale helped prove this. She showed 315 participants, 128 face images. Some of the faces were real and some of the faces were AI generated. And in the paper she found that participants couldn't tell the difference between the AI generated faces and the real ones.
- [03:13]
Kip Bodnar
The last big innovation that's happened is search has gotten completely disrupted. This is very specific to marketing. You see people using ChatGPT and perplexity and Claude and these tool Gemini to search and Google's rolled out AI mode and so the traditional 10 blue links. Google search is slowly declining both in usage and popularity. And you're seeing a rise of these new answer engines that are disrupting how people find and discover businesses much differently than it happened about a decade. Over the last decade.
- [03:48]
Phil Agnew
Evidence suggests that most people still prefer and trust non AI search more than AI search. But the tide is turning higher. Visibility reported that 71% of AI users have now used it for search and 14% of them are using AI each day to run searches that they would typically use Google for. All of these changes made me a bit fearful for marketers.
- [04:15]
Unidentified Host/Interviewer
I think what's interesting about that is there's a number of different marketing roles that are almost being affected by this changes. You've got writers, you've got video creators, you've got SEO. Are these individuals from your own experience today, are they becoming obsolete or are they just changing how they work?
- [04:31]
Kip Bodnar
I think AI is changing dramatically how people work and what they focus on and the skills they need to be successful. So for example, it used to be if you were a graphic designer, you would have to do, oh, I got this brief. I really enjoyed getting, doing the concepting, figuring that out. But then I kind of had to do a rote build out of these hundred assets that once we aligned on the design and style and everything, I had to go and just do pretty manual build out work. That manual build out work is largely gone if you are a true adopter of AI. So a graphic designer is still very important. You need the taste, you need the sensibility, you need to be the design knowledge to refine. What's gone away is that kind of rote. I just got to manually do stuff for the next two days to pump everything out that has become much, much easier, much faster.
- [05:27]
Phil Agnew
But what about content creators, blog writers and social media managers?
- [05:31]
Kip Bodnar
Content creators, kind of similar. Content's actually more important than ever. If you think about this transformation to aeo, you need a whole nother level of specific content. Like, I was talking to one of our HubSpot customers, Wilderness RV. They rent RVs across New Zealand and they're like, you know, how do we move to this new search landscape? Well, and I was like, well, before, you know, you would have had a maybe a listicle of the 10 best animals to see while you're exploring New Zealand or the 10 Ultimate Road Trips, things like that. Right now it's very different. What you're seeing now is I need content so that that's much more niche and much more focused even than the old search engine playbook, where, oh, a family of four from America who's visiting New Zealand for a week. These are the five animals they must see and how to see them, right where there's a whole nother level of granularity. And that level of granularity and detail isn't possible unless you have a really good content creator working with AI tools to help do the research and the collation and the organization of that content. Because you need much more and much more targeted information than you did in the last era of the Internet, before AI.
- [06:46]
Unidentified Host/Interviewer
That's fascinating. I'd never really thought about SEO evolving in that way, and I think that's really useful for the listeners to hear that too. One thing I think what the Listicle maybe would have done is helped you can build more sort of trust and visibility if people become familiar with your content. So HubSpot traditionally created the inbound marketing model, built a following that people believed in. I used to work at a company called Buffer who spent a lot of time creating content, building that trustworthiness of visibility and authority. How do brands do that in a world where perhaps AI is the interface, but where you get this information? You're not going to HubSpot.com to learn about inbound marketing going through AI. How would you advise marketers to build trust and authority in that world?
- [07:31]
Kip Bodnar
This has been the question I've asked myself since ChatGPT launched. Information and answers are now very democratized and commoditized. What does that do to inbound marketing? What does that do to the profession of marketing? I spent the last couple years thinking about it and working on it and getting feedback from lots of people and kind of we came up with this new playbook and it's called loop marketing.
- [07:52]
Phil Agnew
This loop framework is rather interesting. It's essentially Kip's four steps to help marketers adopt AI and use it effectively. However, Kip did have a very quick clarification.
- [08:04]
Kip Bodnar
The principles and some of the core foundations of the way we've been doing marketing for the last 10, 15, 20 years are still exceptionally relevant. It is the tactics, the channels, how we think about executing that has actually changed a lot.
- [08:20]
Phil Agnew
I won't change the basics. It just changes how marketers operate. And Kip says these changes will come in the following four steps.
- [08:28]
Kip Bodnar
The first step is the step that I'd argue got ignored a lot in the last decade of marketing. It's called Express. And that is how do you work with AI to get a really clear story and point of view and style guide for your brand for your campaign? In the last decade, search was kind of easy. PPC ads were kind of easy. And people would just go straight to the tactic versus saying, oh, this is what makes my company different and remarkable. And here's how I can tell it in a really great story.
- [08:59]
Phil Agnew
Put simply, Express is defining your brand identity, sharing it with your AI so the AI can better help your business give you tailored and more personal results. Here's an example, and one of the.
- [09:11]
Kip Bodnar
Things that I do all the time in this Express stage, I take all my HubSpot customer data and then I ask, hey, I'm thinking about running this campaign. How would my customer think about it? How would they feel about it? Where am I missing the mark? What would make it clear for them? And Phil, I've run those types of AI queries and I've done like full traditional focus group surveys, and they've come back very similar. Just happens that the AI version is much faster and much cheaper and allows me to really get that feedback from my customers, but do it in a way that instead of doing it maybe for my one big campaign of the year, I can do it for any.
- [09:49]
Phil Agnew
Now let's move on to the next step that Kip recommends.
- [09:52]
Kip Bodnar
The next step of the framework is tailoring. You know, it's like buying a suit off the rack versus having a bespoke suit made for you. It's a very different experience. One feels good, one feels perfect. Right? That's what we're talking about with tailoring. In the last. Over the last decade, we talked a lot about personalization and marketing. How do we use personal personalization tokens or a little bit of data to kind of make it feel like, hey, we know a little bit about you and we can do. And because of that, you should trust us a little bit more. And now with AI, you can really be personal instead of personalized. You know, at HubSpot, like one of the core examples of this is we use LLMs to write all of our emails now. And so instead of writing one email template and having personalization tokens, we can have a base email. We have data about the person and the company and the problems they're trying to solve. We send that to a large language model and it writes a very bespoke crafted email just for that person at that company. And it's a whole different experience. We see engagement rates up 100 to 400% because it's just, it is deeply personal versus like oh, this is a form marketing email where somebody put in my name and my company name and a couple other things that I filled out in a form.
- [11:09]
Unidentified Host/Interviewer
And on that email example, if I'm on the HubSpot mailing list and there's an event in London coming up, the personalization is that looking at the sort of stuff I've interacted with with HubSpot in the past and tailoring the email based off that, like how are you getting that personalization level from the, from the AI?
- [11:25]
Kip Bodnar
Yeah, so we're, we have the data that obviously your, your behavior data with us. Then we have our firmographic and demographic data. So we have data about, about you that you've either given us or we've enriched based off of third party data sources. So we know you, we know your company, we probably know your job title, we have some approximation for like your priorities, your challenges and everything there. And then based on that we can say, okay, we're running this campaign to have Phil attend this event in London. What, what is the best framing of this event? So we might call out a very specific session of the event that's relative to you and we won't just call it out and say hey, you should check this out. But it would be like, hey, we know you're working on this problem. We've got this world class speaker who's here addressing your problem and this is how we think it might benefit your company. One, two, three in terms. And that's like just a completely different experience.
- [12:24]
Phil Agnew
I found this really interesting. HubSpot use AI to tailor most of the emails they send to. But HubSpot don't stop there. Here's Kip explaining the third step.
- [12:33]
Kip Bodnar
The third stage is Amplify and that's how do you, you've got this personalized message that you've really gotten good feedback on. How do you get that out? And there's a lot of ways that the distribution channels out in the world have changed. And I think that's what's really important for people to know. The biggest one being we've moved from search engine optimization to answer engine optimization, and that has been a massive transformation. And the things that you do to get cited in these answer engines are different. If you look at ChatGPT search results, a lot of them come From Reddit, because OpenAI has a partnership with Reddit, so you have to participate on Reddit. A lot of them are multimedia. So it's not just about text, it's about having video. Like, I've done podcast episodes that have been cited in ChatGPT hours after they were published, just because it was the most relevant video on that topic.
- [13:23]
Unidentified Host/Interviewer
I'm conscious that a lot of the people who listen to this episodes, many of them won't work at companies which will have teams deployed ready to help them understand AI. They might be solopreneurs or they might be and a team of a few. When you're thinking about answering optimization, what advice is there? One bit of advice you would give to these people with a bit less time on their hands as to where they should start is that Reddit, is that video? Where would you advise them to spend their time?
- [13:49]
Kip Bodnar
Most of the time, some of the basic opportunities will honestly just be a rethinking of your content strategy and using maybe these AI text to video tools. You know, for example, I talked to Gaurav, who is the founder of Captions AI, which is a tool that lets you put in text and generate a video. And he said the number one reason people are using that tool is so that they can have more video content on their pages to improve their search engine and answer engine optimization.
- [14:19]
Phil Agnew
So we've heard about express, we've heard about tailoring, we've heard about amplifying. The last tip that Kip had is evolve. Here's why.
- [14:27]
Kip Bodnar
When I was putting this framework together, I was really focused on evolution as being the last step. Because AI lets you get insights and learn faster than ever before. And I thought one of the big problems of marketing over the last decade has been I run these big campaigns or channel programs every month. I have like this slide deck that I either look at or my CEO looks at or my team looks at, and we don't really learn that much and we kind of move on and do the next thing where instead AI, while you have a campaign in progress, you can look at that information through natural language. Instead of needing some fancy data analyst or just not getting the information, you can just ask your AI question, get the insights you're looking for, and you can actually tweak that campaign in flight and get better results. And the thing about AI as your kind of work companion in this, it learns right with you, and it will learn what could have been better from that campaign, and it will apply that to the output of that next campaign in that express stage. It will tell you, hey, we learned that your. Your audience is really sensitive to jargon, so we need to reduce jargon even more, for example. Right? Those types of insights really compound overall.
- [15:40]
Phil Agnew
However, Kip wanted to point out that he thinks AI will allow marketers to just do the things that they always wanted to do.
- [15:47]
Kip Bodnar
I think marketers over the last year or last decade, really, they just wouldn't ask a lot of questions because they're like, oh, my data's a mess and I don't have an analyst. And yeah, I'm looking at this chart, but I always look at it wrong because some underlying data isn't right. And now with AI, yeah, you can ask any question you want and it can contextualize that answer. Not only is your campaign getting better, you as a marketer are getting way better every month, every day, every week.
- [16:14]
Phil Agnew
It all sounds simple. Kips almost painted a picture that AI will do nothing but benefit marketers in businesses. However, he was quick to point out that that is not always the case and he had a warning for me.
- [16:26]
Kip Bodnar
I believe we are now at the point where humans care and capacity to learn are the real limiting constraints.
- [16:34]
Phil Agnew
All of that coming up after the break. Inclusion and Marketing, hosted by Sonja Thompson.
- [16:41]
Unidentified Host/Interviewer
Is brought to you by the HubSpot Podcast Network.
- [16:43]
Phil Agnew
The audio destination for business professionals. Inclusion and marketing digs into the important topics like belonging and customer experience and helping you practice inclusive marketing authentically. Sonja's most recent episode is on Buyer Personas, and I think it's a very good one to get started with. So go and listen to inclusion and marketing wherever you get your podcasts to check it out. Hello and welcome back. You are listening to Nudge with me, Phil Agnew. Before this break, Kip said humans care.
- [17:13]
Kip Bodnar
And capacity to learn are the real limiting constraints.
- [17:17]
Phil Agnew
And here's what he means.
- [17:19]
Kip Bodnar
The technology is already far, far better than most people think. They just don't really take the time to learn it and use it very much. And so they use it for a one off thing and they don't give it a lot of context or the right data and they think it's bad. And I guess what I would say is that is fundamentally wrong. And the thing that I always tell people who maybe not aren't that clear yet is AI is a remarkable at making great educated guesses. And anytime you need to make a great educated guess, you should use AI. So like for example, email segmentation. We used to like put groups into people, thousands of people into one segment, and just send them one bulk email. AI is much better at guessing what each other of those individual people want. What AI is bad at is if you need to give each of those people a perfectly exact specific message and make it exactly the same across all of them. AI is not going to be good at that.
- [18:18]
Unidentified Host/Interviewer
There's a study that I spoke about with this professor Matt Johnson on the show a few episodes ago. It was this really interesting study where he shared with me where people were shown two pieces of art and it was replicated with lots of different types. But basically one piece of art is human created, one is AI generated. And for these specific pieces of art, the AI generated one is the one people prefer. They then tell people this is generated by AI. Now what do you think they find? That when people know that the art is AI generated, their opinions change. That actually a lot of people suddenly prefer the human generation. We're human first. I guess we've got a bias towards the human authors, which I think all of us can understand.
- [18:56]
Phil Agnew
Where does that.
- [18:57]
Unidentified Host/Interviewer
Because when you talk about this capacity for caring, it's like, where does that leave us with AI?
- [19:02]
Phil Agnew
Do you.
- [19:03]
Unidentified Host/Interviewer
Because Matt Johnson on the podcast was making a point that this might just change over time. That give it a few years, we'll get used to AI generated content a lot more and we'll be less fussed by it. Or do you think that is sort of a bias that we'll retain that there are some things. Maybe writing emails is absolutely fine, but maybe creating art and music, or I don't know, writing a marketing book, it might be something that maybe it isn't possible for AI to do it because humans just won't want to read that content as much.
- [19:31]
Kip Bodnar
I think I'm in the camp of I think it's going to evolve over time. The reason for that is that there's a lot of experiences in life. Like would you rather have nothing or would you have something that was AI created? Is the answer. And so like, for example, like a very common issue with a lot of Businesses. If you're a small business and you don't spend that much money with a company, you often get very little human attention, which is like the paradox of economics, right? So if you could have basically an AI salesperson who can give you time and attention and you know, you could interact with over the course of 20 minutes and get all of your questions answered, you're gonna think that's good because you weren't like the. The alternative was pretty poor. The question is, if you're used to having an hour long phone call with a human, would you be okay with an hour long conversation with an AI?
- [20:28]
Phil Agnew
I found this really interesting. Kip doesn't think that AI will totally replace human expertise. He thinks it will add in areas where there's no economical sense for humans to invest. AI can be a great salesperson for perhaps the cheap five dollar Netflix subscription. Netflix could never hire a salesperson to sell that package. It's just too cheap. But perhaps AI could do it at scale. It makes sense. But it did make me wonder how HubSpot will make sure its junior salespeople, you know, the people who are selling the cheaper HubSpot plans. How will it make sure those people are still supported and still valuable in the business? So I asked Kip.
- [21:06]
Unidentified Host/Interviewer
It does make me think. Think about HubSpot as an example. Just because you're here, CMO of HubSpot, thinking about salespeople at HubSpot, it feels like there's going to always be a place for those sales executives who are working with the biggest brands. You need to develop that strong relationship is always going to be the case. There's the question that leads me to.
- [21:23]
Phil Agnew
Think about, is the junior salespeople, the.
- [21:25]
Unidentified Host/Interviewer
Ones who would maybe traditionally be doing the more lower order stuff, learning how that works and then progressing upwards?
- [21:31]
Phil Agnew
Where would you advise junior marketers, junior.
- [21:34]
Unidentified Host/Interviewer
Salespeople in the world of AI and how it's changing a lot of the entry level jobs? Where would you advise for them to spend their time? How can they stay relevant in this new AI world?
- [21:46]
Kip Bodnar
I find this to be one of the most fascinating questions of this whole era because I'm a student of the Internet and I read Reddit and I read X and I read all these things and it's kind of bleak for a lot of people who are early in their career who haven't built a lot of skills. This might be controversial, but this is my fundamental belief. Those what were the last generation of entry level jobs were very kind of like apprentice type jobs. Right? Like, come on board. We'll show you how to write, we'll show you how to package, we'll show you how to design these things. We'll have you do these social media graphics and all these things. Like all of that still exists, but the mentor is now artificial intelligence instead of a human. I can learn those things better and faster and cheaper if I really am focused with AI, But I don't have somebody but like, it's way more self directed, you know, I don't have somebody coming in and saying, hey, it's really important that you know these skills. And I think that's the big challenge. And so if I were in university and I was a year from graduating, like, what would I do? I would say, I would first look inward and I'd say, what am I really passionate about? What are the things that I enjoy? I love design, I love creative problem solving, what have you. And then I would figure out how to build those skills with AI and I would literally just tell AI, it's like, hey, this is what I'm passionate about. These are the skills that I'm going to. I want you to build me a curriculum that I can practice with you every day to learn those.
- [23:15]
Phil Agnew
We're at an interesting point with AI. Like Kip said, there's no doubt that the technology is improving at an unprecedented level. The quality of the outputs from large language models are only improving. High quality art, music, books, I think, and definitely brand work. While all of these things are being created every day by AI. Well, I would say every second, in fact. And yet part of me acknowledges that psychology won't allow us to view everything in black and white terms. Books written by AI will never have the same impact, even if they are as good as human written books. They won't have that impact because we know they're written by AI. They will lack that essentialism that Dr. Matt Johnson spoke about a few episodes ago. However, this view might lead you to ignore AI, to assume that it's not worth spending time on to underplay its impact. I've had the temptation to do just that before, but this chat with Kip convinced me otherwise. AI is here to stay, and in many cases it'll augment and improve our work. I don't think art will be replaced by AI, but that doesn't mean that AI can't improve art. Your next marketer definitely won't be an AI agent, but that doesn't mean your marketers can't learn much faster using AI. There's a lot of hyperbole about how AI will change our lives. And I think it's safe to ignore most of that. But that doesn't mean you should ignore AI entirely, because one way or another, it will be part of your job in the coming months and years. And if you want to learn more about AI, here's where you should go.
- [24:46]
Kip Bodnar
Yeah, you can check out Marketing against the grain on YouTube or wherever you get your podcasts. You can learn more about me and.
- [24:53]
Phil Agnew
HubSpot@HubSpot.com on HubSpot.com, you'll also find the loop framework that Kip talked through today. I've left links to all of that in the show notes. Now, finally, before I go, I wanted to make a request. Nudge is the number one marketing podcast in the uk and that is because thousands of people like you tune in every day. And I'm incredibly grateful for that. But in the world of AI, it can be hard to see the humans behind all of this data, behind all of these listener numbers. So if you do regularly listen to Nudge and you're enjoying the show, perhaps you could take 30 seconds just to leave a review. It's very easy to do on Apple and Spotify. Just go to the Nudge listing wherever you listen to Nudge and leave a review there, leave a rating. And on Apple especially, you can let me know what you think. But in your review or on Spotify, you can just leave me a comment to let me know what you think there. I'd be incredibly grateful if you do that. Otherwise I might just become convinced that you're all a bunch of AI bots. I'm joking.
- [25:45]
Unidentified Host/Interviewer
I don't think that's the case, but I would really appreciate a review if.
- [25:48]
Phil Agnew
You have 30 seconds after today's episode. Thank you so much for listening, folks. Coming up on Nudge, I've got some really exciting guests. I have the best behavioral insights team's chief behavioral scientist Michael Hallsworth coming on the show. Richard Shotten is coming back on to talk about his latest book, the Columbia Business Professor Adam Galensky is coming on. Plus legendary marketer Adam Ferrier and award winning professor Dr. Sunita Saar. Loads of very, very good guests coming shortly on Nudge.
- [26:15]
Unidentified Host/Interviewer
I really hope you're enjoying the show.
- [26:17]
Phil Agnew
And there is lots more to look forward to. Thank you so much for listening. I'll be back next Monday. Cheers.