Read Podcast Transcript

Jack Rawlings (strategicabm) – A quick note before the podcast starts. We recorded this one back in May, and so some of the information that we discuss in this episode may be a little bit out of date, with the pace at which AI is changing and developing all the time.

We are sure there's still plenty of information and points in here that you'll be interested to listen to, but likewise, we will be looking to record a new episode on AI very shortly, to try and catch up with some of the most recent developments and changes in the industry.

But I hope you enjoy this one, and keep your eyes peeled for the next one coming out very soon. Thanks.
___

Jack (strategicabm) – Welcome to another episode of ABM Under the Hood. My name is Jack Rawlings and I'm joined again today by Josh Weale. Thanks for joining, Josh.

Josh Weale (strategicabm) – Hey, Jack, thanks for having me again.

Jack (strategicabm) – No worries. So, today we are going to be talking about specifically AI and ABM. How to use AI in the context of ABM and essentially kind of really digging into different use cases that we've seen from sort of different areas, but also that we, we've been using ourselves in recent weeks and months from, yeah, with clients and with campaigns that we're running, and all of that kind of stuff.

So, there's lots to cover. I don't think we're gonna necessarily be able to go into all of the intricacies and the details that we, that we could, today. So I think, you know, this potentially is almost like a kind of an intro to the topic. I'm sure that there'll be a point in, you know, a couple of weeks or months, or so, where we might wanna revisit it and talk about other aspects of it, 'cause it's, well, it's changing so quickly and all the time, you know, there's things that have developed even in the last few days. Which has kind of almost dramatically changed the scene, and what's possible and what's feasible with the tools that we're starting to use.

So, I guess first thing for us to kind of discuss really is the basics of AI. I mean, maybe we wanna just quickly go into some of the sort of tools that we've been seeing that are, kind of, out there. And a little bit about what, you know, what they can potentially be used for.

So, I mean, obviously first one, one big one, is ChatGPT. That's something that I think everyone will be familiar with and is probably the most, yeah, well known and well used at the moment. One of the more, most exciting tools on the market. But any other tools or platforms that you are using, or you've seen from an AI perspective?

Josh (strategicabm) – Yeah, I mean, there's loads. Too many to mention really, but I think…

Jack (strategicabm) – Yeah!

Josh (strategicabm) – … some of the ones that I've been trying out is obviously Google's competitor to what OpenAI are doing with ChatGPT, called Bard, been testing out that recently. There's also kind of all of the generative AI stuff around imagery and graphics, videos, voice creation. Things like MidJourney and stuff like that.

And also some of the stuff that I've been testing as well around kind of presentation decks. There's a couple of tools called Gamma and also Decktopus, which I've been trying out recently, trying to see how well I can create a presentation and yeah, it's, it's pretty stunning actually, that things can happen in 30 seconds, which would ordinarily have taken me about, kind of, two hours. So!

Jack (strategicabm) – Yeah! 

Josh (strategicabm) – Although there's still a long way to go before it becomes really that easy. But definitely, it's amazing to see kind of how, how these kind of tools have arrived on the scene, and how they're impacting our day-to-day, really. 

Jack (strategicabm) – So, I guess, you know, the big thing that we've been sort of thinking about and trying to figure out, I suppose is how can we use AI tools in ABM programs? What, what is it that, you know, what benefits is there that we can kind of glean from these, these different tools that we've got available?

Obviously it's still very much early days with, with a lot of them, and I don't think we've quite sort of worked out the full capabilities yet, but I know that you are doing some, some kind of work at the moment. You're doing a bit of an, an investigation into the different tools and the different use cases in the ABM sort of framework. Did you want to give a bit of detail on that at all?

Josh (strategicabm) – Yeah, I mean, as you say, from, from my perspective at the minute, it's kind of just messing around, seeing what works, seeing how we can kind of optimize things. Do things in a more efficient way. And yeah, one of the things that I've been really focusing on is just trying to understand kind of around insights and, and data specifically around kind of how do we categorize insights and data, and find patterns in data. Both statistical and also qualitative, in terms of if we take for example, a customer interview or an interview with a salesperson, and trying to find kind of common characteristics and patterns around kind of what are the key things that they're talking about that we can then translate into something meaningful for our insights process? That's something where I've seen kind of a lot of help.

So yeah, I mean one of the, the use cases that I've been kind of using most recently is, kind of, using things like ChatGPT for analysis and statistical analysis and qualitative analysis of like, during the insights process of taking a kind of bunch of data that I've collected either through interviews or through kind of building out spreadsheets through, through quantitative data. And then actually running that information through ChatGPT to kind of get a head start on, on what the kind of clear trends and patterns are.

Obviously I can do that over time, but what I've found is that by having this kind of little assistant that supports me with it, it's allowing me to kind of quickly pull together some trends and some patterns, which I can then use as almost like a starting point for me to then go and do the more human element of researching those particular trends, and making sure that it's valid, and also that it's correct. 'Cause obviously we can't just take the output for granted. We have to apply a level of interrogation to what we get out of these tools.

Jack (strategicabm) – Yeah.

Josh (strategicabm) – And I think specifically around kind of, what am I doing on a day-to-day basis? A lot of it is around using these tools to avoid kind of what I'd maybe term as like 'blank page syndrome,' where like you're sat there kind of staring at a blank piece of paper and wondering, well, what's the best way of tackling this problem? Or approaching this piece of work?

Jack (strategicabm) – Yeah.

Josh (strategicabm) – And yeah, sometimes you can kind of overthink things a little bit too much. And what I've found is that these tools are a really good way of kind of overcoming that and giving you a starting point that you can then go and take and build it into some really kind of meaningful work.

Jack (strategicabm) – Yeah, yeah, no, that's, it's a great point. And I think, yeah, so, there's that sort of side of things, which is the, the kind of, it's almost like the, the instigator – it helps you kind of almost brainstorm, and you know, bounce ideas off and, and kind of get, yeah, get some, get something sort of started.

And I think also the other, another really kind of key element that I've seen used for with particularly ChatGPT is the kind of its ability to do some of the kind of more mundane admin-y type tasks that are involved in setting up a program, an ABM program, you know, things like that kind of like data. Not, not analysis, certainly not analysis – yet! Although that's, I think, potentially, you know, there are ways that that could be coming, but, but more so compiling of data, you know, into whatever kind of format you're looking for. Even things like, I've been recently using it for helping me with like Excel and, and Google Sheets formulas and stuff.

Josh (strategicabm) – Yeah.

Jack (strategicabm) – I'm probably, I would probably call myself sort of intermediate when it comes to Google Sheets and Excel. Like I know some formulas, I know the basics, but I wouldn't by any means suggest that I'm an expert. But, with ChatGPT, I've kind of been able to put in, you know, requests and kind of prompts about how to, how to sort of slice data up or how to, kind of, you know, get a different picture on things in certain ways.

And actually that's been incredibly useful for me. It's almost like having kind of, it's like increased my Excel and Google Sheets skills without necessarily actually learning it. I mean, that touches on something, which I think is more of a kind of, there's an 'ethical' question as such, but more of, like a, you know, it's, it is a bit, a bit of a deeper question, which is by doing this, are we kind of changing how we work and our brains work? That we're not necessarily gonna be learning as much –  you know, how to do things from scratch – in the way that we would otherwise?

But, I guess it's sort of similar to like, you know, the transition from it to using, like, search engines, right? Like there was, there was a time when people had to learn a lot of facts and trivia and the answers to things and they had to just kind of learn it, you know, rote.

Whereas now, or certainly in over the last little couple of decades, it's changed to being effective is to know how to find the information that you're looking for, which usually, you know, used to be using, using search search engines to do that, and the web. But now, potentially, it's taking it even a step further and it's knowing how to use AI tools to, to give you that information, or to do the work for you, essentially.

Josh (strategicabm) – Yeah, absolutely. I think, I think what I've been trying to kind of instill in my own processes and practices when it comes to using tools like ChatGPT and Bard, and other kind of generative AI things it's around, kind of, I try to avoid doing anything that's, I would call a lazy task. So, if it's like a piece of work where I can do it and it takes a certain amount of time, I will still do that work!

Jack (strategicabm) – Okay.

Josh (strategicabm) – And approach it in the same way that I would, because I think that point you made there around, kind of, and I think people will have different opinions, like some people will see kind of efficiency as the number one priority over everything.

Jack (strategicabm) – Yeah, yeah.

Josh (strategicabm) – But I do think there is something to be said around kind of maintaining our own kind of approach to things, and not just relying on technology. And it's about kind of using these tools as kind of leverage rather than, kind of, a replacement for us.

And it kind of touches on this whole big question around will AI replace jobs, and will it replace people? And maybe I'm jumping ahead a little bit, but I think, yeah, there's certainly a level of kind of more basic simple tasks that I can see how ChatGPT, or something similar, could actually fulfill those roles. But I just don't see a world where we get to a situation where humans aren't required in the process any more.

Jack (strategicabm) – No, yeah.

Josh (strategicabm) – I think ultimately, there will always be a human element that's required, at least, unless I could be proven wrong by the advancements in technology, but I think there would always need to be that human element implemented within any of these processes to kind of really validate the work and make it meaningful.

Specifically for Account-based Marketing, 'cause it is about understanding what the target accounts need and actually turning that into something meaningful. And that still needs a level of kind of rationality and imagination that I think AI probably still doesn't have.

Jack (strategicabm) – Yeah, I completely agree. I think it's, there's certainly elements that I think are potentially gonna change. You know, there will be jobs that will not exist as a result of this, a hundred percent. I think there will be certain jobs that will be replaced by ChatGPT. Actually, and AI generally. I actually don't, I actually don't think we know yet which jobs those are – because I think that it's very hard to tell which direction things are gonna go in, right?

I think, I don't think a lot of people expected there to be such an immediate kind of influx of creative use cases for AI as there have been recently, in the sense like, you know, using it for images and video and music, and all of that kind of stuff. But, I think there's also gonna be a bit of a pushback on those sides, types of things as well.

And, you know, aside from regulations and that kind of stuff, and you know, intellectual property-type concerns. But from an almost a consumer perspective, I think people will potentially not be, not be as keen to engage with content or, or creative kind of output that is AI generated, if there hasn't been some kind of level of human input or overlay onto it.

I think it's gonna be, you know, in the same way that potentially people kind of switch off to, to anything that's kind of cookie cutter, that's sort of, you know, mass produced, the same kind of same kind of thing on, you know, I think AI from a creative perspective could go down that route, and therefore I'm not sure that that's the, the way that people are gonna, the biggest benefit or the biggest change that's gonna happen from, from AI. I think it's gonna be much more around the way in which it augments our current processes, and strategies and techniques, and not about, you know, replacing what we currently do. Essentially. 

So, let's talk a bit about some different potential use cases specifically to ABM. So we'll talk about like predictive analytics, personalization, segmentation, that kind of stuff.

Josh (strategicabm) – Yeah.

Jack (strategicabm) – When it comes, so when it comes to AI and ABM, there's, you know, there's lots of potential use cases even, even right now in this sort of early / early-ish stage of the kind of industry. But I think it's already clear that there are some use cases that are, that are already possible, or are certainly becoming very, very plausible in the, you know, in the next couple of weeks and months.

So, I mean, one I think is especially interesting from an ABM perspective is personalization, when it comes to, you know, personalization at scale in particular, which, you know, sounds a bit like a, it sounds a little bit like a kind of, you know, it's a bit of a paradox, isn't it? Almost personalization at scale. But actually, that's kind of what we look like, what we're aiming for sometimes with like One-to-many programs, that kind of stuff. Make it feel like it's, it's unique and personal to that, to that person, but actually it's something that's been sort of done, you know, across, in a more scaled fashion.

So yeah. How do you think that we could use AI tools from a personalization perspective? What kind of ways would we be able to do that?

Josh (strategicabm) – Yeah, I think one of the ways that I've seen it kind of used effectively - and I've been testing this out myself - is kind of through building messaging matrices in and structuring messages in kind of a particular way to match up to kind of different seniorities and different functions.

Jack (strategicabm) – Yeah.

Josh (strategicabm) – Essentially, once we've created a messaging hierarchy, and a value proposition for our target accounts and who we're trying to reach, it's quite a manual task to actually take that and then actually build it into something that's, I suppose, personalized at the account level. Especially if you're trying to do this at One-to-many; if it's One-to-few and it's 15 accounts, then it's something that we can do.

Jack (strategicabm) – Yeah.

Josh (strategicabm) – But AI kind of opens up the opportunity to do this in a more scaled way. So taking that messaging hierarchy, applying kind of job function and seniority levels to the data that we have, the contacts that we're looking to, to engage with, or even just the personas that the people that we're looking to reach into in terms of the decision-making unit, and actually taking the messaging hierarchy and personalizing it, and asking the AI system to do that. And then obviously in comes the human element to refine and to make sure that it all makes sense, and it's aligned to our original thinking.

Jack (strategicabm) – Yeah.

Josh (strategicabm) – But the capacity to kind of make that whole process more efficient, and to lessen the amount of time that's involved – but with the potential upside of the, the additional level of personalization that you can achieve. I think that's, that's kind of a really powerful use case in terms of how, how something like ChatGPT could actually empower greater personalization within a target account list.

Jack (strategicabm) – Yeah, and I think the other thing that I've thought about as well from a personalization perspective, and we're gonna, we're gonna come to the, one of the latest developments this week with ChatGPT in terms of plugins. We'll go into a bit more detail on that shortly.

But I was thinking about Zapier. So obviously you can, with Zapier, you can connect different, different tools. You can, you can basically bring in inputs from different, you know, quite a wide range. I think it's 5,000 different sites or, or apps. So, you know, one of which is potentially CRM. So you know, HubSpot or Salesforce or whichever CRM you're using, connecting that to ChatGPT, bringing in first-party data from your CRM into ChatGPT and then, and then using that overlaid on the messaging matrix-type process that you just, you just discussed could be, could be an absolute game changer.

Because essentially what you are, what you are potentially able to do there is take not necessarily real-time, but real-world data, you know –  your interactions, the conversations that the sales team are having. Any kind of keywords or pain points that have come up in those conversations, potentially: What emails have they opened? What, you know, what content have they engaged with? Bringing that into the messaging matrix.

And, so you've got kind of that, default messaging matrix that you've created and then overlaying the first-party data that, you know, at the moment, we can do that. As you say, you can do that manually with 15 or so accounts, it's not too bad. But when you get into the One-to-many kind of level, trying to do that is, that's just not possible basically at the moment. Unless you've got absolute, you know, infinite time and money – but this could actually be very feasible. And something that I think could be done as of now, basically, if you've got access to ChatGPT Plus, and you connect up the plugins with Zapier and HubSpot say, you know, you could start doing this now. Right?

Josh (strategicabm) – Yeah, exactly. And I, I think from an account selection point of view, you made a really interesting point then around kind of being able to analyze at scale what is resonating with an audience. 

So, we have a campaign type in the agency called Kickstart Accelerator Program. And this is essentially taking our clients' best thinking and getting it to market as quickly as possible. Almost in kind of a Demand Generation style, to really just understand what kind of messages, content types, channels and tactics are, are really resonating with the target accounts that we're trying to reach into.

And we do this for two reasons. One being to kind of warm up the audience and build awareness for the client, but also to gain an understanding of what is resonating with the target accounts. So we can use that knowledge and intelligence to actually build that into our more comprehensive One-to-many, One-to-few, One-to-one campaigns.

So we can see, well, what, what is it that we're putting into the market that resonates the most? And at the moment, we obviously we do this at, at a campaign level to understand what are the assets that are performing well, what are the messages that are performing well? And we can do that at a general level in terms of understanding the market, and which accounts it's resonating with, but not maybe as granular as we would like because it's just too resource intensive.

But something like AI could actually help massively here in terms of account selection. So we can actually load in all of that data that we've collected from the market, look at kind of the different copy that was used even down to like a single ad level and run some analysis through, through ChatGPT, for example, and understand kind of is there a common trend or a common pattern between different pieces of copy? Is there specific groups of accounts that resonate with certain copy or design? And understand kind of what are the best-performing ads?

And it just adds kind of an additional layer of intelligence to what we're already doing, and being able to do it at scale. Whereas before it, it wouldn't make sense from a point of view in terms of how much resource you'd have to invest in terms of what you'd get back.

Jack (strategicabm) – Yeah.

Josh (strategicabm) – But in this scenario, if you can do this in a much more efficient way, and then the outcome is something that's quite kind of tangible and it gives you kind of intelligence that can be built upon and act as kind of, I suppose, fuel to enable, kind of, your insights and messaging development work, then I think it's a really valuable exercise and something that can only kind of elevate what you're already doing.

Jack (strategicabm) – Yeah. I think you're absolutely spot on. And actually I was thinking, you know, similarly as well as, as well as being able to bring in that kind of campaign data into your thinking and analyze that campaign data in a much more kind of tangible way, you've also got tools that you can use. You know, you might currently be using, and I'm thinking the likes of intent data platforms or also, you know, potentially social listening tools as well, where that's, it's more kind of third-party data that you could essentially bring into the, into ChatGPT or whatever it might be.

I think, I was trying to think about how you would do that. ’Cause I'm not sure that Zapier necessarily syncs up with a lot of intent data platforms or social listening tools. It might do, but the alternative being, if you've got, for example, like Bombora integrated with HubSpot, you can bring the data from HubSpot in, from Bombora into HubSpot into ChatGPT.

That could potentially be a way to, so talking more about the kind of the account selection side of things now. Using it from almost like a predictive analytics perspective. Which accounts are gonna be interested? Are gonna, are gonna be, are in market? Which accounts are trending and engaging with which, you know, which topics?

But that analysis normally, you know, when we currently do it, we do it for, we do it in a One-to, we can do it in a One-to-many process, in the sense that we can take a list of, of accounts, you know, that are trending in an intent data platform and then manually kind of go through those, align, compare and contrast with the first-party data, compare and contrast with campaign data.

You know, you almost have to take these things in, you know, bit by bit in isolation, combine it manually, kind of go through. There are some tools that you can get that will bring stuff together in one place, but they tend to be quite expensive. Or they tend to be not fully comprehensive; they don't have all of this data that you are, that you are...

Josh (strategicabm) – Yeah.

Jack (strategicabm) – … that you could potentially look at. But with ChatGPT or with a similar tool, you can essentially bring all of this data into one place, have it evaluated for you, compare and contrasting the first-party with the third-party with the campaign data.

And then suddenly, you've got this list of accounts that are like, you know are, know who you are, have engaged with your content, engage with topics that you are writing content about, are in market - you know, all of this stuff that you can just pull out. And then suddenly you've got a really, really hot target account list essentially, that you can --

Josh (strategicabm) – Yeah.

Jack (strategicabm) – That you can work with, right? That for me is, is where it's gonna get to very, very soon. And it's pretty, pretty astounding, really, that we can get to that level for the price that it's probably gonna cost the investment that's gonna cost to get there. Right?

Josh (strategicabm) – Yeah. And, and I think there, there's a couple of really key points that you made there and I think I want to tie it back to kind of this, the whole trend of why it's important for, for strategic marketers to kind of get to know the capabilities of these AI tools and leverage them on a day-to-day basis. Because...

Jack (strategicabm) – Yeah.

Josh (strategicabm) – ... specifically in kind of Account-based Marketing teams, we know from working with our clients that quite often, these programs, they start out pretty small in terms of resource, maybe one or two people, and you've got kind of one Marketer who's trying to do six or seven different things all at once. And it becomes kind of a question of prioritization in terms of: How much time have I got to do something? How much resource have I got? How much budget do I have? And can I achieve everything I want to do?

And I think for me, everything that we talk about kind of ties back to something like ChatGPT or Bard being used as that kind of accelerator for that one-person marketing team to actually enable them to do Account-based Marketing without having to try and justify additional budget to kind of invest in other people.

Like, to have a technology or a tool there that can actually help you with a level of data science, or data analysis like that, that's massively helpful because if you work in an organization where you have a team of data scientists, they've got their own work going on. And a lot of time you'll be able to kind of build relationships with different teams, but ultimately, the work that you need to get done from an Account-based Marketing perspective, often, is something that you need to get done yourself.

So to have that kind of just additional level of assistance through something that you can control and you can kind of build your own processes around, I think it's just a massive enabler for, for being able to be more efficient and do more with less, which is obviously quite a common theme that we're hearing all the time at the moment!

Jack (strategicabm) – Yeah, absolutely. And on the other side of things as well, right? You've got the plugins option, and so you've got this inputting in, analyzing the data, bringing that all into ChatGPT, getting it or the equivalent, you know, I'm kind of using ChatGPT as a bit of the shorthand at the moment for those sorts of tools. 'Cause, you know, we don't know which one's gonna end up being the one that everyone ends up using.

But you know, you can bring all of that data in, but also with the plugins that are starting to come through – and again, Zapier being one – you can also then bring that data out as well. So you can then filter that data out automatically into, for example, your HubSpot or your, you know, ad platforms potentially even, you know, things like that where you can essentially... it doesn't, you don't even need to take that, take that manual approach of then building audiences and building lists, and all of this kind of stuff. You can, you can basically just do it automatically.

On top of that, you can then also do the, you could create email sequences and sales sequences and whatever else in those platforms directly from GPT based on the data that's been fed in, from your own campaigns, from your first-party data, from your third-party data. You're suddenly not even, so you're sending out the list of accounts that you need to target from GPT, you're sending out the emails, everything. Like, you're not having to do any of that as such.

The only thing that you have to do is QA it; is evaluate it. And make sure that everything is up to scratch and up to par. And that for me is like, I don't... it's groundbreaking, really. But what it doesn't suggest to me is that it's gonna replace, you know... it replaces functions that we are currently doing. Or we currently wish we could do at that scale. But it doesn't, it doesn't replace, it doesn't replace the Marketer, the person doing it, it just means that the Marketer can do more, and be more effective, and do more, you know, more of it, right? With a smaller resource and a smaller budget.

So suddenly, instead of being able to only target a hundred accounts with a One-to-many campaign, you could target, you know, thousands realistically. As long as they, as long as there was the right amount of data coming in and all of that, it suddenly becomes feasible to, to really scale that up. So, you're then looking at potentially more, you know, more from a results perspective as well. So the whole thing is, is quite kind of like, it's a magnifier, right? It's like a 10x on our abilities as Marketers, rather than a replacement for us.

Josh (strategicabm) – Yes.

Jack (strategicabm) – And I say that, I'm not saying that out of a personal kind of hope that my job's not gonna be taken, you know, it's more a case of I'm excited by the possibilities that we can actually really do more with the tools that we've got at our disposal.

Josh (strategicabm) – Yeah, no, absolutely.

Jack (strategicabm) – So, you know, it's not gonna be all plain sailing, right? It's not gonna be, it's not gonna be a completely smooth and simple process. So what is, what are some of the challenges, what are some of the challenges you've had already with using AI and ABM, and what challenges can you foresee coming as we kind of use this technology more?

Josh (strategicabm) – Yeah, I think the biggest one from my perspective at the moment is, is actually getting the output that you want. And it's about kind of knowing, knowing what you want and having a strategy for how you're going to use these tools.

I think if you just go into using something like ChatGPT or Bard and just start typing prompts, you are, what you're likely to get back is something that is quite generic and it's not really fit for purpose. I think a challenge to overcome for a lot of Marketers is going to be, well: How can we write the best prompts, or engineer the best prompts, that actually achieve what we need it to do? How can we train it to sound like us? How can we train these generative AI platforms to feel authentic?

Authenticity is going to be kind of a real big barrier to actually using a lot of these tools in-campaign, because especially for an Account-based Marketing campaign where we're talking about we need to be authentic, we need to be personalized, we need to really get across the value that the product, the solution offers – it's not good enough to just be quite generic. It has to feel real.

I mean, from my perspective, it's about kind of having a strategy to know what you want to get out of it and having prompts that are engineered to be able to achieve that. That's one challenge.

Jack (strategicabm) – Yeah.

Josh (strategicabm) – And secondly is more of kind of a big-picture ethical challenge around kind of ownership and also regulation.

Just this week, we've seen kind of the founder of OpenAI, in US Congress, answering questions around kind of the regulation that needs to be put in place on these kind of generative AI platforms to kind of, I know it's, it's a cliché to say, but to avoid kind of one of these 'Skynet' situations that people think is gonna come from AI!

Jack (strategicabm) – Yeah! There's also, there's also this week, or it might have even been yesterday, I think, I read where the EU is now looking at new legislation around, specifically around LLMs – so Large Language Models – and being able to essentially the, who's gonna have the rights to the data, and who's gonna license, who's gonna get a license to be able to build them, and all of this kind of stuff.

So there's already legislation being worked through the European government to, Parliament to, put a blocker on some of this kind of, you know, at the moment it's very much Wild, Wild West territory. Right?

Josh (strategicabm) – Yeah.

Jack (strategicabm) – It's real, like, it's just, yeah, it's got. It's a gold rush! And I think at some point regulation's definitely gonna come in, you know.

Josh (strategicabm) – Yeah.

Jack (strategicabm) – Sooner than later.

Josh (strategicabm) – And then this is why it is so important that as Marketers, we learn how to use the tools, but we don't base everything we do around the tools. They can become part of our day-to-day, but they shouldn't be everything that we do.

Jack (strategicabm) – Yeah.

Josh (strategicabm) – Because, what's here today could be gone tomorrow based on kind of regulation that comes in. So yeah, so it's really, really important that we, we are considering kind of how we can best use it, but also be aware that if we're building, building services and building products around exactly what a tool can do today, that can quite easily be kind of regulated, and not be possible moving forward.

Jack (strategicabm) – Yeah. Yeah, a hundred percent. I think, you know, another challenge that I've had recently with it is, it's - too much! That is the way I put it! Like, I don't, I sometimes dunno where to start with it. I'm so, I'm so in awe of the possibilities and the different possibilities that I'm kind of, and I'm not particularly good at, at focusing on one thing at the best of times, but it's, it's very sort of, you know, 'new shiny thing!'

Josh (strategicabm) – Yeah.

Jack (strategicabm) – And there's so many, there's so many different options with it that I don't really know, I don't really know what a good starting point is, and which thing I should focus on and get kind of skilled at.

At the moment, I'm being quite general with it and just kind of testing things out, trying different things. Doing a little bit here, a little bit there. Like, I did this, I think you might have seen my LinkedIn post the other day where I kind of built this like video that was entirely done through AI. So it was ChatGPT script and then put into a tool called ElevenLabs to bring out the voiceover, and then used DALL-E to create the images. And then the only bit that I didn't use AI for was the video editing, which I did.

But like, that was just a little, fun little thing. It wasn't, you know, there was no purpose to that other than just to be able to experiment and see what was possible and, and to sort of see what the outputs was. 

I don't, I haven't yet nailed how to bring it into my day-to-day workflow. You know, without, like I'm thinking of all these things, but I'm not sure which one to start with basically is what I'm saying. Where, which is gonna be the most beneficial to me in the short to medium term.

Josh (strategicabm) – Yeah, I mean...

Jack (strategicabm) – So.

Josh (strategicabm) – ... yeah, but it certainly can feel, feel quite overwhelming. I think it'll be, it'll be interesting to see kind of like how, how we can build strategies around how to leverage AI – specifically in processes and bringing in kind of specific use cases that are kind of the default approach. This is how we use it.

Jack (strategicabm) – Yeah.

Josh (strategicabm) – Because, yeah, you're right. If we just kind of turn to it every opportunity, there's a risk that we just become kind of distracted and don't actually achieve what we aimed, what we set out to achieve.

Jack (strategicabm) – Yeah. Yeah. And that's a good point, I think is sometimes you've gotta start with the objectives and work back a little bit, right? Like what's the overarching objective of whatever you're doing, you know, your program, and then does it make sense to bring in AI to help me with this? If so, how do I do that?

That, that is the only, that's the way so far that I've been able to kind of get some, some stuff, you know, into my workflow. But it's not, it's by no means refined yet; it's still very kind of rudimentary.

So, let's talk about plugins. So this week, I dunno if everyone's got access to them yet, or if you've got ChatGPT Plus, but I think most people have now. They were rolled out for me, I got them I think on Tuesday. We are, so we're recording this in the middle of May. So it was, it was, yeah, it was a Tuesday night, I think that I got them, got access to them. And at the same time also got access to the web-browsing functionality.

Josh (strategicabm) – So obviously ChatGPT was previously, even in the Plus subscription, was limited to data up to 2021 was its cutoff, but now, there is a beta version or an alpha version of the plugin of the browser functionality that essentially means it can search the internet up, you know, present day.

That plus the range of plugins that are already available; there's 70-odd plugins for different things. I mean, that's a game changer, right? That's, even just from, from what it was two weeks ago, which was already, you know, ripe with possibilities. It's now even bigger. The scale is even bigger, right?

Jack (strategicabm) – Yeah. Yeah, absolutely. And I think it's only going to continue to grow. I mean.

Josh (strategicabm) – Yeah. There's like 70 plugins, or so that are available today. In two weeks' time, as more and more businesses start to think about, well: How can we integrate with something like ChatGPT's API? How can we make this usable for our kind of audience? It's only gonna get bigger and bigger. So again, it kind of goes back to that point about around being focused on what your strategy is, what your objectives are and what you're trying to achieve.

Jack (strategicabm) – Yeah.

Josh (strategicabm) – And selecting the plugins that are gonna help you to do that. So yeah, as you mentioned, that kind of web browsing. There's also a couple of plugins that I've been using called Web Pilot, and I think it's something to do with prompting as well. Where the kind of approach that I'm taking is, is if I need to kind of go and let's say I've been doing a little bit of work around kind of competitor analysis and looking at kind of what are the key words and key topics they're talking about online.

For me to do that it would be a case of kind of having to go through and manually look up all of the different articles, search kind of their webpages and, and do all of that. Which we do for One-to-few programs and One-to-one programs. But when we're trying to do that at kind of a One-to-many scale and understand kind of a larger spectrum of competitors, and understanding kind of all of the different sub-competitors as well, there's a lot of work that goes into that.

Jack (strategicabm) – Yeah, yeah.

Josh (strategicabm) – So having something that can access the internet and go and pull that data back and put it into something that's coherent rather than just, obviously web scrapers and stuff like that have been around for years. But a lot of the time you get that and it's incoherent, a lot of gaps. You have to kind of turn it into something usable. But to have something that you can kind of now send out and go and pull that data back, and it automatically builds it into something that you can kind of take...

Jack (strategicabm) – With sources as well.

Josh (strategicabm) – ... yeah, with sources, and being able to kind of validate it as well. Yeah, it's just massively powerful in terms of accelerating some of that work.

Jack (strategicabm) – Yeah. Yeah. A hundred percent, yeah. It was a real eye-opener for me when I got access to those. And also there's a few others that I think are, that are really potentially very intriguing from an ABM perspective. So you've got, there's two, there's one called Chat with PDF, I think it is, yeah. Chat with PDF, and there's one called VoxScript, which are both kind of similar in the sense that one is to, one goes through a PDF.

So what is currently kind of, there's, you know, loads of PDFs online, but actually they're quite hard to take anything from beyond, you know, having to manually read through them, right? You can't, you can't scrape them. Potentially you can, I don't know specifically, but they're very hard to kind of get the data, the information and data from, right? This you can use to almost, you almost ask ChatGPT questions about the PDF – it will answer based on the PDF, what's in the PDF.

Same with the YouTube one, VoxScript, and we'll share a link to all of our kind of thoughts on the plugin side of things, but also we're gonna, we're gonna share a prompt cheat sheet for ABM and AI prompts as well. So that will be in the resources, but the YouTube prompt, the YouTube plugin, VoxScript, as well, that basically means you can go through videos and you do the same thing. You can ask questions based on information in a video, which is, you know, that feels really quite futuristic.

Those sorts of things potentially from an insights-gathering perspective could be enormous, right? If you've got PDFs that are, a company's been publishing, you can take information from that and to build a picture of up of, you know, I'm thinking things like financial reports and statements and stuff like that come from there. But there's other tools as well, other plugins like, so it's BizTok, Public and Golden – three other ones that I've seen, which are all similar types of tools, I think. Essentially they're like, company analysis tools, financial analysis and stuff like that. So, that's several things that you can bring into it.

What else? You've got Yabble, have you looked at that one? Yabble. It's like a survey creation. So you create surveys based on the information you want to gather. It will create a survey that kind of specifically, you know, take to gather that information. I showed you one the other day, didn’t I, Show Me, which is a pretty interesting one. That's the, it actually draws diagrams for you. So I used it to draw a sort of dummy org chart, which was, quite interesting.

So I think one thing that anyone, businesses that are planning on using AI in their kind of ABM journey should be aware of and should be kind of thinking about is data privacy. You know, we mentioned earlier about bringing in you know, first-party data from a CRM and other, you know, data tools and stuff like that. There's some considerations that businesses need to be thinking about, right? When they're putting that data into ChatGPT, you know, what should we be sort of aware of when we're inputting stuff into ChatGPT, would you say? 

Josh (strategicabm) – Yeah, I think it's a really, really good question and I think it's really important that this is kind of a prioritized consideration for anybody looking to use AI in particularly kind of generative AI Large Language Models.

Because these tools, they're obviously trained on information that's submitted to them. And the more data they have, the more they know, the kind of better the quality of the output is. Which is obviously a great thing in terms of usability. But from a company's perspective there's a lot of considerations around data privacy. And as a disclaimer by no means are we kind of data privacy experts. But, in terms of like some common sense things to be aware of – particularly with tools like ChatGPT – anything around kind of proprietary data, data that's not in the public domain, there's a setting in ChatGPT which you can switch off kind of the saving of that information into ChatGPT. And there's actually a few tools I think there's a Open GPT playground I think it's called which doesn't, it has all the kind of same functionalities and same kind of capabilities, but it's not kept as up to date.

So, you have a payoff there, but there's tools like that that you can use to kind of do analysis on data that you don't want to save into the Large Language Model, because ultimately it's data that belongs to you. You don't want it to be used in the training, kind of, of these models and certainly don't want 'em to kind of become public knowledge or in the public domain.

So that's one thing. And kind of the second thing is around, I guess from a... not just a privacy perspective, but also moving towards kind of intellectual property perspective; when we're thinking about kind of generative AI, about being aware that some of the kind of, certainly some of the things that we're seeing at the moment in terms of creation of videos, creation of audio files, even to an extent creation of documents... If you are kind of claiming that those documents belong to you, or they've been created by you, and you've used some of these tools, just be aware that that may not actually be true.

And there is currently, as you mentioned earlier, kind of legislation that who actually owns the rights to what is created by these tools. So, just a couple of common sense things to consider if you're looking to use some of these tools in your day-to-day. 

Jack (strategicabm) – So yeah, I think the other thing as well as that sort of privacy side of things I think actually people have to be really careful from a legal standpoint, I think again as you said, not data privacy experts and also not a legal expert by any stretch, but, that area is gonna grow massively in the coming weeks, months, years around the legalities of everything. And particularly when it comes to something like uploading sensitive data to ChatGPT or any tool LLM, any tool or Large Language Model, you might be at risk of actually kind of, you know, falling foul of some legal, you know, legal issues and legal frameworks, right?

You know, if you think about things like... just one example I've got off the top of my head is if you were to upload sensitive financial data ahead of an acquisition or a, you know, going public or something like that. That would be very much a, well, very illegal – and you know, bordering on fraud. So, that kind of stuff you, even if unintentionally, there are consequences at play if you go down that route.

So it's not even just privacy, but even, you know, serious real legal issues that you could run into. So just be very careful about that and make sure that you are, as you say, Josh, kind of having that button. If you're putting anything sensitive in there, switch the button to 'don't save the chat' and make sure that none of that information's being uploaded into the model. 

We have had some questions as well this week. We posted, or I posted on LinkedIn about recording this episode and we had a couple of questions from some of our network, and the first one is from Juan Berganza Setien, and he asked us will AI be able to choose a list of specific accounts? And I think we've kind of answered that to some extent. You know, we would say that we are not necessarily right at that point quite yet, but it's very much almost there. And with these plugins now and the ability to bring in first- / third-party data, all of that into ChatGPT, it's almost there basically. It could be done, right?

Josh (strategicabm) – Yeah. Yeah, it could. But I think it again, to reiterate a point that we've kind of spoken about already in this podcast, it's not necessarily about getting the AI tool to select a group of accounts; it's about using that tool as part of your overall process. And having a strategy, having a kind of defined criteria for how you're going to set those accounts, and then using the tool to actually identify which are the best fit accounts, and using it to analyze data, analyze all the different kind of information that you have available about those accounts to build the list.

So, although it can certainly help in terms of making that process a lot more efficient, I would certainly at this point consider how it is part of a bigger process rather than just, can we just plug the tool into our account list and say: Spit out which accounts are we gonna go after next quarter?

Jack (strategicabm) – Yeah, yeah. Agreed. Yeah, I think all of this stuff that we discussed today needs that level of human oversight and strategy. Yeah, ultimately strategy, you know, tying it back to your objectives and being clear about what you are looking to achieve with it. So yeah, spot on.

We also had questions from, or question from Corrina Owens, who is a good friend of the podcast. She was on Let's talk ABM as well, our kind of sister podcast. And she wanted to know about, specifically, technologies that are already kind of existing that are helping Marketers, and ABMers in particular, kind of identify moments in a prospect's journey in real time.

So, you know, identifying anonymous traffic, bias sentiment, whether or not something's kind of converting, you know, the amount of clicks it's taking to get to a certain action, that kind of stuff. Is there anything that you're using or you are aware of that kind of helps with that?

Josh (strategicabm) – Yeah, I mean I haven't certainly come across one tool that can do all of those things, but I think - certainly not in one package - but I think one of the areas where I think you can tie multiple things together is what you were talking about earlier in terms of connecting things with tools like Zapier.

Jack (strategicabm) – Yeah.

Josh (strategicabm) – So you could take a platform like Dealfront, for example, which is identifying anonymized traffic to a website, connect that up to Zapier so that it has a kind of constant stream of information that's feeding into Google Sheets, and then having kind of something like ChatGPT which is running on a script on that Google sheet to kind of analyze, well, who are these accounts that are coming in? What pages are they looking at? How long are they staying on the platform for?

And actually then connecting that piece of work to something like a notification or Slack, even. You could in theory create a Slack channel which is directly being sent to yourself or to the Sales team, to say rather than, kind of, I suppose to an extent the manual process today of having to kind of keep that feed going and checking in on it every now and then, you could actually create a working notification system that is not just informing the team that somebody's been on the website and this is what they've looked at, but it's actually going one step further and analyzing how long they've spent there.

And in some ways you could actually set up to start recommending: 'Share this piece of content with them.' 'Share this piece of content with them.' Depending on what they've been doing and what level of kind of interest they've shown in the pages that they've been on. So that's certainly one area where I think AI could be explored in terms of connecting several tools together, and using something like ChatGPT as a kind of an enabler to facilitate a workflow like that.

Jack (strategicabm) – Yeah, yeah, absolutely. And actually, I was thinking, you know, potentially you could do the same with something like Hotjar or, you know, there are other similar tools like that where it's kind of that page analysis-type activity, you know, heat maps and stuff like that, but also specific kind of like link, you know, data clicks on pages, all of that kind of stuff.

Combining that, it might be that it would make sense to use – I've been thinking about this in terms of this sort of, what's that centralized hub gonna be? Like, is it best to build everything into a spreadsheet, as you say, or would it make more sense to have something like, if you've got HubSpot where you can get most of that data integrated into HubSpot, you know, or other CRMs, Salesforce and Marketo and stuff do similar.

Or, is it that, you know, almost you've got a, you know, ChatGPT is almost like the centralized place or whichever kind of, you know, tool you're using is a centralized place where that's where everything goes and then it feeds back out again. It's kind of hard to know exactly which one is the best and I guess it probably depends somewhat on what you are used to using, what the context of your business and all that kind of stuff. What would you say? Do you think it makes more sense to be working from a spreadsheet or from the CRM, or how would you approach that?

Josh (strategicabm) – Yeah, I think, wherever possible in ABM it's important to have a single source of truth.

Jack (strategicabm) – Yeah.

Josh (strategicabm) – And the easiest way of kind of facilitating that is often to get your CRM right, and connect your CRM to the tools that you need to. So I would always say, kind of, focus on getting the CRM in a place where it can do those things first, in terms of storing the data and capturing data. And then using those kind of augmented tools to enrich the data and build workflows that then keep everything up to date, keep everything kind of in as real time as possible in terms of activity.

So yeah, I would say that's probably the best approach is focusing on your CRM as your single source of truth, and then using the tools around it.

Jack (strategicabm) – Yeah, 'cause I guess most of these tools, you know, most of that kind of tech stack that we sort of talk about when it comes to ABM things like Dealfront things like HotJar, things like, you know, SparkToro, or these kinds of like sentiment-analysis tools – to kind of go back to to one of Corrina's points – all of that can, most of it can sync up and integrate directly with the major CRMs.

So it kind of makes sense to get all of that data into that consolidated place. And then if you need to do the manipulation, if you need to do the analysis, you can pull it out and you can do that and then put it back in. But yeah, having that single repository where everything kind of feeds in as a starting point, I think is a good shout. Cool.

So I mean we have probably, I would say only just kind of scratched the surface of this topic. We are doing more and more each day on this as a business and as individuals we are kind of using it and investigating continually. So, I very much envisage us having a follow-up conversation about this topic in a few weeks or months time, where we'll maybe dig into some other sort of areas and aspects that we've not necessarily covered today.

But it's exciting, it's an interesting time, an interesting opportunity for us as an industry, I think, to really capitalize on what's kind of coming. There's things that I am seeing in, you know, even short- / medium-term pipeline coming through that really are, that even again are exciting me.

You know, there's an iOS app that's just come out in the US for ChatGPT, which apparently also has voice activation, which would be for me something I'd be pretty interested in trying out. It's not available outside the US yet, but that's coming pretty soon – and it will be on Android soon as well. Likewise, there's this Code, what's it called? Code?

Josh (strategicabm) – Interpreter.

Jack (strategicabm) – Code Interpreter, that's it, yeah, yeah. Code interpreter, which sounds like it could be, well I mean that basically just means it opens up, really does open up infinite possibilities and if you can do that, right? You can literally build and spin up apps and little tools and stuff in, you know, a matter of minutes, hours or days, you know, even with limited coding knowledge and stuff like that. So that could be another game changer. But it's like there's another game changer every day, every week at the moment, right?

So, but yeah, so plenty more to discuss I'm sure on this topic. So we'll be back on this one, but in the meantime we'll also be sharing a link to a resource alongside this. We'll be sharing a ChatGPT prompt cheat sheet for specifically kind of geared towards ABM use cases as well. So that will be something that you can view and download as well.

So, thanks very much Josh for chatting to me again today. Looking forward to catching up on this topic again in the future and some other interesting ABM sort of points and topics to discuss soon as well.

Josh (strategicabm) –  Yeah, thanks for having me, Jack. Great to chat all things AI.

Jack (strategicabm) –  Nice one. Cheers Josh.

Josh (strategicabm) –  Cheers.