Brent Weaver: And welcome back, hopefully you all had a quick second to grab some coffee or tea, I know personally I’ve got a double espresso queued up for this next session, so I want to be on my A-game for taking notes and getting the most out of the session I’m about to introduce. But big thank you to Tim and Elizabeth, the new agency model, let’s give them some additional love there in chat, such great content, such great Q&A, thank you all for engaging and just being such amazing guests on Agency Advantage 2025.

A couple of announcements before we jump into our next session, just a reminder on the Easter eggs, be on the lookout for those Easter eggs, there’s that pinned message in chat if you just joined us and you have no idea what I’m talking about. There are these little treasures kind of hidden throughout the two days of the event, and for those people that submit the Easter eggs the most of them accurately, you’re going to be entered to win potentially some gift cards, some hosting credits, we’ve got 50 and $100 gift cards, so keep on the lookout for those. And keep the engagement coming as well, we’ve got the leaderboard, we’re going to be racking up those points over the couple of days and giving out some amazing prizes to our most engaged, highest participants over the course of this event.

And one last thing before we introduce Christian, our next speaker, I’d love to take a moment to recognize our sponsors. So we’ve got some amazing folks supporting this event, we have ProfileTree, we have TemplateMonster, WP Umbrella, and of course the company that I’m repping today, E2M. So if you haven’t checked out the expo for those companies, make sure you make some time over the course of these two days, jump in the expo rooms, go to their website, look at what kind of services they offer, see if those companies can help out your agency. Support our sponsors because they make events like this possible, and I hope you’ve gotten so much value over the course of just today, I feel like my notebook is filled with insights and actions that I’m looking forward to implementing at E2M.

So our next speaker is going to be talking about his topic, which is from spray and pray to strategic growth, how AI agents are powering account-based marketing for agencies. So I’m really excited to welcome Christian Banach to our stage, he is the founder of a consultancy helping agencies and consultancies predictably land six and seven figure opportunities, with a background in experiential marketing and top-of-the-funnel business development. I don’t know, Christian, I’m a huge fan of ABM, so I’m sure AI has been transforming this business and making it better, faster, more effective, so I’m super pumped for your talk, welcome to our Agency Advantage stage, Christian, it’s great to have you here.

Christian Banach: Great to be here, Brent, I appreciate that. I hope this session is going to live up to the hype there. I’m a big fan of ABM myself, and I just find what’s available to us now with AI has just really been remarkable, and it’s consistently evolving as you’ll see in my presentation here, so excited to share and hopefully add some value back to the audience here.

Brent Weaver: All right, friends in chat, let’s give Christian some love, let’s hype him up, let’s get him ready. Christian, the stage is yours.

Christian Banach: All right, let me go ahead and share my screen. So as Brent mentioned, the topic of today is from spray and pray to strategic growth, how AI agents are powering account-based marketing for agencies. So a little bit of an introduction here to start things off. I’d like to actually start a poll, and I’d love to know from the agencies that are on, how difficult has new business been for your firm compared to the prior 12 months? So if we could fire up that poll and see what the results are, we can come back to those results, but I’ll move ahead here while everybody’s answering that question.

And I can tell you, we have run a similar poll for our own audience, and we did it first in February and then we did it again in May, and in February it was 70% of agencies said that it was either somewhat harder or significantly harder, and when we did it in May it jumped up to 76% saying that new business is harder or significantly harder than it was 12 months ago. So I’m not sure yet what the results are here, but the bottom line is it’s much more challenging than it had ever been, and it feels like it’s getting even more hard. So, do we have any of the results yet that we can show, or should we come back to that? All right, well let’s come back to that, I don’t see the results right now.

But bottom line is, I think we would all agree that it’s not easier for sure, and particularly with sort of the spray and pray approach just not working what it used to work before. And there’s a lot of reasons why spray and pray hasn’t been working, and we kind of have tracked what those things are. And one of the big reasons is sales tool accessibility. I’ve been doing this for 15, 20 years now, and the tools that are available today didn’t even exist 15 or 20 years ago, and slowly these tools started to get introduced, so things like sales engagement platforms, marketing automation, contact databases. And over time those tools have become more available and more affordable, and that has really led to a surge in use, as well as then a surge in email and email overload.

Statista actually reports that there’s 376 billion emails predicted to be exchanged daily in 2025, which is a 23% increase from five years ago. So despite all these other channels and social media and Slacks, email is still increasing. So because of that, people’s inboxes are overloaded, the email service providers like Gmail and Yahoo and Microsoft have begun making the inboxes stricter to get into, and they’ve been introducing new sending requirements. And this is all a good thing, because we don’t want spam hitting us, but at the same time it has made legitimate emails that we’re trying to send also harder to get through.

So there’s more emails being sent, it’s harder to land in the inbox, deliverability is more challenging, and because of that reply rates are plummeting. Some research done by Belkins found that the reply rate has actually plummeted from 6.2 to 4.2% from C-level executives from ’23 to ’24, and I haven’t gotten the data yet for this year, but I can tell you we’re seeing on our side it’s also getting more and more challenging. So that type of spray and pray approach is just not working anymore.

And I think we’ve all been on the receiving end now of what I call crappy AI emails, and that is also in some ways leading to some of the challenges that we’re seeing, because a lot of people are using some of these tools that maybe aren’t ready for prime time yet, and we are getting really bad emails. And what is an AI email? I think we can have all gotten tuned in to understand what one sounds like, you could just sort of see it and feel like, oh, this was written by AI. And what are those qualities? Well, it could sound like a robot, not like a human would actually sit down and write an email, it sort of says nothing specific about you or the situation or the business, it’s just very general.

It also seems like it just takes too long to get to the point, it’s just a lot of fluff that’s included, a lot of these emails have what I call fake personalization, so you could see kind of in this example here it’s like, you know, exciting marketing and advertising space, like, yeah, it is personalized, but it doesn’t really add any value to it. And then just in general, something just feels off about the message. And what we’re going to be talking about here today is not this, I think the system that we’ll walk you through is going to really explore how to do this, but do it at a level that really is elevated from, in many cases, what even humans could sit down and do.

So the bottom line here is agencies don’t necessarily need more tools, but they need better systems. What I’m going to provide here today is really using a ChatGPT account, you don’t need expensive tools to do what I’m about to show you here. But that being said, I want you to start thinking about AI not just as a tool but really as a co-worker that is a part of your team, just in a digital sense.

So before I get into it, I just want to give you a little bit of background on who we are, why we have the authority I guess to talk about this. So we are a growth consulting firm, and we help agencies, consulting firms, adtech, martech companies land six and seven figure opportunities predictably. They come to us when they have challenges, maybe with low brand awareness, insufficient leads, stalled deals in the pipeline, or retention and growth challenges. And we’ve worked with everything from large global holding companies down to small boutique agencies, and our job really is to help them land opportunities at large enterprise organizations.

And then a little bit about me personally, I have been in the sales and marketing space now for really my entire career. I actually started a concert promotions business out of high school, worked with Lady Gaga, worked with Pitbull, some really great artists, and along the way got introduced to experiential marketing, and my business eventually morphed into half concerts and half experiential events, did that for about 15 years. And then sold off assets to that business and went all in on just the agency space, where I have been focused now for the last probably 12, 13 years, and really helping clients generate millions of dollars of revenue. I’ve really kind of found my superpower is helping these companies break into these large enterprise organizations.

So our business has fundamentally changed. This particular business of mine started about five years ago, and what we were doing five years ago and what we’re doing today is vastly different, and one of those things that we’re doing differently is what we’re going to present here today. So I want to talk a little bit about what we call the AIM process, and it’s an acronym that will really describe how we’re approaching using multiple agents to create these ABM type programs. So the A stands for assemble research, and using agents to do so, I is integrating those insights, and M is messaging strategically. So you’ll see how each one of these different agents fall into those categories here in a moment.

So one thing that I want to make clear here is that you don’t need a big human team to run this, but what you do need is smart AI agents. And the agents that we’re going to be walking through today, for this example, really fall into four categories. There is what we call our account and industry research agent, there’s the prospect research agent, the strategist agent, and your email copywriting agent. And the idea here is that you have specialized agent roles that will replicate, in many cases enhance even, human strategic thinking, it will save time on your research and writing process, and really enable you to hyper-personalize messaging at scale.

And we’re going to unpack each one of these here, but before I do that I do want to do another poll, about what is your biggest obstacle to scaling outreach, assuming that you’re either doing it for yourself in order to generate new business, or you’re doing it maybe on behalf of clients. But I’m curious, what are your biggest scaling challenges right now? So I’ll let that poll run, and I’ll get into the next slides here.

So the next few slides I’m going to go deep into this multi-agent workflow. I know I’ve been a part of a lot of AI presentations, and it’s really 30,000-foot views, not really actionable, my goal today is to really pull back the curtain and show you what we’re doing right now and where this could go eventually, and I hope that you’re able to walk away with something really actionable from this talk. So what we’re working towards, we’re actually going to use kind of like a live case study, so I hope AI is working for me today. But this is an example of an AI-generated email that this same process walked through, so this is for a client of ours called the Amateur Pickleball Association, and you can take a time to kind of take a look at what we’re working towards, but this is sort of the output of what we’re going to get from all the inputs that I’m going to walk through here in a moment.

So to kind of dive in, we’re going to go with that first agent, and that first agent is our account and industry research agent. And what it does is, this agent’s job is to research target companies and industries to create a detailed account profile that will be later used for your hyper-personalized messaging. So this is again from an account-based marketing perspective, we are trying to zero in on messaging and pain points, challenges that specific companies are having, so this agent is responsible for that first step.

From a platform perspective, we are using ChatGPT-4o, but important to note the deep research feature of ChatGPT. In the highest paid version, which is $200 a month, you get something like 250 different deep research views, but this has really been a critical feature for us to get the level of detail that we need to build out these account profiles. If you haven’t used the deep research profile, you can get something like a 10, 20-page report about one company, and it’s remarkably good. We have experimented with some of the other ones like Perplexity and Gemini, and we’ve just found, at least right now, that we like the ChatGPT deep research the most.

But what you’re going to end up feeding this agent is an account and industry research prompt that we’ll talk about here in a second, and then a company name and a company website, and this agent is basically part of your team, their job is to research that company and that industry. So they’re looking at things like company news, marketing initiatives, key decision makers, etc.

Elizabeth Zurn Lujan: Hey Christian, we do have the poll results if you want to, I don’t know when you want to, we have both polls, so if you want to do that now or later, happy to help you cue that up.

Christian Banach: Yeah, why don’t we go over that now, before I kind of get into the real meat of it. So yeah, let’s talk about that first one, so how about the new business being more challenging?

Elizabeth Zurn Lujan: Yeah, so there we go, how difficult has new business been for your firm compared to prior few months, so you got the results there on the screen.

Christian Banach: Yeah, so 90%, so even 90% said it was harder or significantly harder, so that’s even more than what we saw just a couple months ago. What about the other one, about scaling outreach, and we’ll put that one up here as well. So, can’t personalize at scale, don’t trust AI to write well over the top two, research takes too long, lack of clarity on messaging 25%, so a little bit of a mixed bag here, but yeah, I’ll make sure that I touch upon each one of these challenges and how I believe the system overcomes it, so yeah, thanks for that.

So, I’m not going to go into great depth here, but what I want to preface here is that the prompt that we’re putting together needs to be specific to the campaign that you’re running. This is not a one-size-fits-all prompt, and I think that’s where some people get it wrong, they’re looking for an easy button, they just want to have one prompt, hey go do this research, and then they’re going to have another campaign, go do this research. Each one of these research requests has to be specific, just like if you were going to be doing this as a human, you wouldn’t just start researching, you would actually think about, okay, for this campaign, what do I want to know?

So for example, for this particular program, this is a pickleball group, so we want to know, is this brand involved in pickleball at all, what other sponsorships are they involved in, pickleball is like a health and lifestyle sport, so what else are they doing in health and lifestyle? So I won’t go through this in depth, but you can see it’s been very much tailored to what specifically we want to know about this brand, for this particular client and this campaign. And then you’re going to give it some instructions as far as the output that you’re looking to get from it.

So that’s the first agent, and then the walkthrough that I’ll show you here. So what we’ve done now is, for this particular example, we’re going to use Danone, and that’s going to be our example. And once I’ve fed that into ChatGPT, this is the account and industry brief that we get back about Danone. So you can see it specifically looked about pickleball involvement, other sponsorships that they’ve been involved in, other experiential marketing, so you can see it’s a 10-page report that it came back with, and it spits this back in usually 10 minutes, maybe sometimes even less.

So that’s the output, so now we have that account and industry research done. Now we would move on to our second agent, which is our prospect research agent. So we would like to do this for not just, write a personalized message for that particular company, but even the individual that we’re reaching out to at the company, we want to write something personalized to them. So this agent is responsible for researching that prospect’s background, what interviews have they done, what is their social footprint like, anything else that we could again use for our personalization.

And again we’re using ChatGPT-4o, and in this case, in terms of the input, we’re going to have another prompt, we’re going to give it a company name, a contact name, a contact title, their LinkedIn, and then maybe even any past interactions, maybe your team has already previously spoken to that person or to that company, you could feed that all into this research agent at this point in time. So in terms of this live walkthrough that we’re doing, this target contact is going to be Ivy Chen, she is the senior manager of the yogurt category at the company, and we are going to then look at what our research agent came back with.

So again, we gave it some specific instructions on what to look for, and interestingly enough, in this individual’s LinkedIn profile it said that when they’re not in the office they like to go to great restaurants, play pickleball, and go road-tripping to national parks, and this was AI that uncovered this through their research process. It also looked at their sponsorship, other things in like health and wellness that she has maybe been involved in, it’s pulling quotes that they have, other experiential marketing, so you can see very granular information, and very specifically we’re asking it to look for certain things across their footprint here.

So now we have done the company-level research, we have done the prospect-level research, now we need to start bringing this all together. And we’ll talk about this a little bit more, but as you notice we’re giving very specific instructions, and the reason for that is, at least currently the way these AI platforms work, if you try to give it too much you’re not going to get as much back. Like, we could have asked the agent to research the company and then research the prospect, but what we’ve found is that you start to get not as detailed information, it starts to cut corners, it starts to maybe hallucinate at times, so as specific and niche as possible is the best.

But what ends up happening, though, is once you have all this research you need somebody to pull it all together. And again, just think about how it would work at an agency with humans, right, a human will be pulling all this different information, and then they need to combine it and synthesize it and come up with some strategy, and that’s where agent number three comes in, the strategist. So now this is someone that needs to really understand your services, your case studies, your thought leadership, so they need to be trained on you. And then they need to be also taking these account-level and industry-level and prospect-level researches to identify entry points and messaging strategies. So how are we going to, based on who we are and what we do and what we know about this prospect and this company, where’s our way in?

So now from a platform perspective we’re not doing any research per se, we’re synthesizing, so we could be back now on the normal ChatGPT-4o, not the deep research, and we’re going to be inputting a strategist research prompt, and then all those other researches that we’ve done, along with some additional information about the company, about your thought leadership, case studies, etc. So I’ll give you an idea here, again for our example we’re going to have the strategist agent prompt, which I’ll show you in a second, along with background on the company, which is in this case the APA, and then you’re going to feed all of that together into it.

So what we end up getting here is an example of what the company background would look like, so Amateur Pickleball Association overview, background, talks about their tournaments, so all the relevant information that you would really need to know who this company happens to be. And then we get into the strategy brief, so this is the output that it would get, it’s going to give you an executive summary, it’s going to talk about strategic angles, and as you recall, Ivy plays pickleball, so it looked at that and made that determination that that might be relevant. It’s talking about white space for Danone, talking about age, so it’s really pulling together all these key insights and finding some ways in.

It’s even coming up with some ideas on how a sponsorship activation could look like, it’s talking about audience fit, why and where do the audiences fit, it’s coming up with some different messaging insights, so you can see it’s really done an incredible job of pulling all of those documents together and coming up with a brief. And now what do we do with that brief, and that would then lead us to our fourth and final agent for this process, which is our email copywriting agent.

So for this agent, what they’re doing is they are now taking the research, the strategy brief, and they are going to write hyper-personalized email sequences. But not only that, they’re going to apply best practices for subject lines, for initial outreach, for follow-up emails, so they are trained on how to be a great email copywriter. And again we can use ChatGPT-4o, not the deep research, and we’re going to be inputting the email sequencing prompt, which I’ll show in a second, along with that strategy report, and the output of that is going to be a multi-email sequence, and we’re actually going to do this live and see what it comes up with.

So again, we’re going to have that prompt, the strategy brief, and we’re doing this for Ivy Chen. So what we’re going to do now is, I’m going to move to, this is the brief that we have for our copywriter, so email number one, it’s going to tell you, hey, you’re a copywriter specializing in cold email, it’s talking about certain instructions that we want to have, certain subject lines, formatting requirements, etc. So I’m going to copy this into ChatGPT, and then I’m going to bring over that strategy brief as an attachment, and I’m going to press go, and we start to get our output here.

So what we’re going to do now is we’ll take this, and we can copy this subject, and we’re going to put it into a spreadsheet, and I’ll explain why, and then we have the message itself, we’ll paste that here. Let me try to change the view on this a bit. “Saw you’re a big pickleball fan, thought this might hit home,” Danone’s been a key player in the NFL, NHL, and youth sports, so again it’s pulling some past sponsorship information from the research, but they haven’t stepped on a pickleball court yet, so part of that research they uncovered that they’re not investing right now in pickleball sponsorship. The APA is growing, it’s aligned with your wellness mission, so again it’s done research into the mission of Danone, it talks a little bit now about the APA, their 200 tournaments, their number of players, etc, again from the company background information.

And then it starts talking about how we can drive retail sales, we could do yogurt stands, family lounges, etc, at these events, it’s coming up with some activation ideas, and then there’s a call to action at the end, and again this has all been pulled from the different, you can see how some of the different pieces of this research has now fit into this. And that would be email number one, and then for the second email in the sequence, we would just go ahead and we would copy the second prompt, and we’ve already given it now some instructions, we want this email to sound a little bit different, so let me copy this here and paste this into the same chatbot. And then we’re getting a new subject line, and we’re getting a new message.

So again, I won’t read this in general, but you can see it’s now trying to take a little bit of a different angle, it’s talking about sampling stations, lounges, so again it’s taking all of that research that has been done, some of those ideas, and putting it into a really relevant email format. And unlike some of those examples, the example I said in the beginning, this is very specific, really targeted to the prospect themselves, and not only the prospect, obviously the company as well.

So with that being said, we don’t just trust the AI, we do have a human element involved in this here as well. And as you saw, I copied the emails into Google Sheets, and the reason that we do Google Sheets is we will take that spreadsheet and we will actually upload it into our sales engagement platform, and then we could then send those emails in an automated fashion. You might have also noticed that there was some HTML formatting there, that’s really just because of the sales engagement platform, we write them in HTML format this way so we can use merge tags, so we don’t gotta get into a lot of those specifics, but just, if you’re wondering, that’s really why. But once we get those emails copied into the Google Sheets, a human, one of our account directors, will come in, they’ll do QA, they’ll do any edits if necessary, and then that gets uploaded, like I said, into our sales engagement platform, and then those hyper-personalized emails can be sent in an automated fashion.

So kind of zoom out and look at the entire process here, and back to the AIM model that we have. The first few steps are all around assemble, right, you need to start with all the different prompts, so it does take a little bit of time to create these prompts, you don’t want to rush through this, it’s sort of the old garbage in, garbage out, the better your prompt the better guided you are, the better output you’re going to get. So it starts with the blue, with all the different prompts, then you get into the different agents that we walk through, the account and industry research, the prospect, the strategist, the email copywriting, and then we get into the process of putting it into the Google sheet, QA, and upload to the platform. So assemble, integrate, and then message, with human intervention along the way here.

So with that being said, human versus automation, I think it’s good to kind of point out a few things. So we do still use humans as part of this process, we use offshore virtual assistants to run the prompts, so we will write the prompts, we will do some initial testing of it, and once we feel like we’re getting the right outputs we’ll hand it off to offshore VAs who are then running the research, running the prompts, copying them into the spreadsheets, and then our account managers will come along and they will do the QA and then upload it into the sales engagement platform.

You might be asking, well, why are we still using humans? Well, there’s a few different reasons, and I don’t want to get too technical here, but we really have found the deep research feature of ChatGPT to be a big unlock for making this work, because you get that level of detail. But unfortunately deep research is not currently available via ChatGPT’s API, so we can’t automate at this point in time the deep research, and again like I said we’ve tested this out on some others that do offer APIs, and we just haven’t found that the level of depth of research is the same, so that’s one reason.

Another reason, and this one’s interesting, so when you’re using the API you’re essentially starting from scratch every time, it doesn’t necessarily remember prior messages or any of the uploads you’ve done, so it doesn’t necessarily learn your tone, structure, and brand voice, versus when you’re using the web version of this, every email you write it’s taking into account memory. So you’ll see, even when you first start this off in the beginning, the emails are good, and then they go from good to great over time as it learns your style and your tone, but you don’t get that necessarily with the API right now. And we’ve experimented with doing some different prompting and we were able to get it pretty close, but we still have found that, at least for right now, the online web version is better.

Also, when you’re doing it manually you have the opportunity to make some real-time changes, you might notice that the subject lines are off or something might be happening, and you could refine it a lot easier, versus just automating the entire process and being left with a hundred emails at the end. And ultimately, a lot of what we’re doing is based on campaigns, and it’s account-based marketing, so we’re not necessarily sending tens of thousands of these emails, so we might only be doing batches of a dozen or 30 or 40 or maybe even less sometimes. So to build a whole automated process around that when we’re still tweaking things doesn’t always make the most sense. Now, I fully expect that this is probably going to change, but at least right now this is where we’ve landed.

So, why do these AI emails still suck, a lot of them? Well, I think, again, they’re trying to use one prompt to do too much, as you saw we have five, six different prompts that we’re using, each with a specialized role. And therefore there’s vague and general inputs that they’re giving them because the prompts aren’t very good, so it’s the garbage in, garbage out, like I said. I also think that the strategy phase, especially when you’re bringing a lot of research together, if we were to just take the research and give it to the email copywriter, it’s too much information, there are limits to how much it could process, so we’ve found having that strategist put the strategy doc together and then bringing in the email copywriter gives us much better results.

And again, the temptation for a lot of people here is to scale, and yes we want to personalize at scale, I know that was a pain point that we talked about, but it’s still a balancing act. This doesn’t mean let’s just load in 10,000 contacts a month and blast it out, we still want to have this be really a personalized approach, but enable us to do a level of research and a level of outreach that one human necessarily couldn’t do. So eventually what we think this will look like, not all too different than what we talked about, there still will be these research prompts that need to be done, but everything you see there in pink could be eventually automated. And eventually, I don’t mean too far down the path, it could be weeks or months before those steps could be potentially automated, so in theory we could have a list of companies in a spreadsheet, you load up all the different prompts and you press go, and then each one of these agents will run autonomously on their own, put it into the spreadsheet, and there could be another integration right into your sales engagement platform. And some of this is again available even right now, it’s just we have reasons on why we still like to have the humans involved in some of these steps.

So when you’re thinking about, should I be doing this manually or hybrid or using automation, a couple things to think about. If it’s a new campaign, you probably want to do it more manual and hybrid for now, because you’re still learning what works. If you’re moving into untested industries or markets, again maybe the manual hybrid makes more sense, or if it’s very strategic for you and these are very high value, maybe the manual hybrid makes more sense. On the other hand, if you have something that’s worked and it’s very repeatable, maybe automation is there, or if you have a super large data set, or you have this messaging that you know works well, then maybe you want to lean towards automation at that point.

So a couple side-by-side views here, you know, your kind of traditional manual process of doing this, like if someone was literally to sit down, do the research and write these themselves, this could take days or weeks to write this level of multiple emails in a sequence to this many prospects and companies, versus an AI workflow can do this in hours. Scalability is limited, versus the AI agent workflow which is high. Cost to have a human sit down and do all that level of research and writing, and the level of person that you’re going to need to do that in your organization, is very high, versus the AI agent workflow is much lower.

Consistency also varies, yes if you have somebody doing it and they’re a great strategist and a great copywriter it could be pretty consistent, but if you’re talking more junior BDRs, the traditional process is probably going to vary, versus the AI agent workflow, once you have your prompts ready, it’s very stable, especially if you’re doing QA. And then again from a personalization standpoint, the traditional process, yes you can get very high levels of personalization, but it’s slow, versus the AI agent workflow, it’s very high also but it can be done very fast. So there’s trade-offs with each one of these.

So we believe, why does our system work? Again, the AI agent is focused on one specific task, which will give you better output, humans are still involved in the process to ensure quality and nuance, automation is going to come but only when we know it’s going to work, and again this is not necessarily some plug-and-play AI tool, this is really thinking about it from a strategic system level.

So I do want to share some results and takeaways that we’ve experienced as we started this process. So when we’re looking at conversion rate, and by conversion I’m meaning companies we’ve reached out to that we’ve landed a meeting, for this particular client, the Amateur Pickleball Association, we’ve been working with them for a number of years, I wanted to use this as an example to show you more historicals. So over the year and a half or so that we worked with them before the AI agents, when we were doing this manually, we were converting about 6.53% of companies that we reached out to, we got a meeting with. Now, when we did the AI agents, we were able to increase that so far now to 10.3%, so a 57% lift in conversion rate using this AI agent approach.

Some lessons that we learned along the way here. Number one, prompt over platform, success really is about smart prompt design more than what tool you’re using, as I’ve talked about. One role, one agent, make sure that they are very good and trained on that one specific task, that’s going to give you dramatically better results. This is important, I haven’t talked about this yet, but start with the end in mind, before you start building out all these different agents and building out the prompts, think about what the message is that you ultimately want to send, what are the pain points, what particularly do you want it to look for. Because once you start thinking about what you want that message to sound like in the end and what a great message will be, you can reverse-engineer and make sure that you’re creating the prompts that are going to find that particular information.

In the pickleball example, I knew that I wanted it to first and foremost look for opportunities that the brand or the prospect was involved in pickleball, so I specifically gave the researcher that remit to find those particular insights, had I not thought about that up front it might not have actually even looked on their profile whether or not they play pickleball. So think about the end in mind first, and then, right input, right results, if you give too little context or too vague you’re not going to, or too much, you’re going to overwhelm the model, so make sure you’re really zeroing in with precision. And then again, systems beat shortcuts, there are tools that are out there, we’ve experimented with them, and the outputs that we got are very much similar to what you saw in that crappy email that I shared, and it’s because these tools are not necessarily built at a campaign level with multiple agents like what we have done here. Can you get it done faster maybe with them, yes, but you’re not going to get the same level that I feel this system is providing for us and our clients.

So at the end of the day, build everything around that AIM model, assemble, integrate, message. Now, I know I focused a lot of this around email, and if I had more time I would be talking about all these other things as well, but just to kind of give you a picture, I wanted to focus on one channel particularly. However, the same exact system can power other ABM channels at scale, it can be used for not just email copywriting but ad copywriting, that same research and that same strategy brief can be handed over to a trained AI copywriting agent that writes LinkedIn ads, display retargeting, etc, but you would just train that agent on what makes a great LinkedIn ad. Same thing from landing page, you could hand this off to a landing page agent that’s trained on how to write great landing pages, or general social media posts or newsletters, so really any of this can be done, you could do the whole process, have it write the emails, have it write ads, have it write landing pages, all with really just adding an additional brief and taking the same research that you’ve already done.

So I’m hoping that this is not too overwhelming, I’m sure we have some people on the call that are very much involved with AI and this is easy breezy, and then I’m sure we have others that this might be a little bit overwhelming to. So if you’re feeling maybe a little bit overwhelmed, I wanted to offer at least a starting point for you, and the idea here is don’t necessarily start with the full system that I had outlined, start a little bit smaller. I would recommend that you first start with that account and industry research prompt, so using that deep research feature to analyze, just start with one company, one industry, challenge and positioning, just start there, get a prompt that you feel good about and get comfortable using the deep research feature.

Then skip the prospect level, skip the strategist for now, and just create a copywriting prompt, so feed it then what you feel, again it could be email copywriting, could be an ad, whatever you want to start with, but you are going to ask it to write something specific to that company. And then it’ll include, you’re going to want to include a one-page description of your company so it understands who it’s writing on behalf of, and send it and see what happens. I think just starting with that, then you could start to add additional agents along the way, and you’re going to feel much better about the process, so I think starting small and expanding out from there is a good place to go.

So, I know I’ve shared a lot of information here, and if this is something that you would like to bring to your agency, or if you’re already doing it and you want some help on how to implement it, I’d be happy to have that conversation with you, my email address is there, my website Christianbanach.com, you could reach out and our team can set up a conversation. Or if you’re not ready for that right now, we do have an email newsletter that has 40,000 different agencies on it, where we announce CMOs, RFP opportunities, we share other growth strategies like what we’ve shared here today, and you could subscribe, it’s totally free at Christianbanach.com. But that being said, I’d like to thank everybody, and I think we might have some questions, and I can stick around to hopefully answer them to my best ability.

Brent Weaver: Yeah, very nice, thank you Christian, appreciate the amazing talk. Let’s look at chat here, let’s give Christian some appreciation, some love, let’s see some of those emojis fire off here, we had lots of good comments coming through, chat got thanks from Fatima, very nice, I’m sure more folks will jump in there here in a second. Let me see if we have any questions, sort by latest here, I’m not sure I’ve seen any Q&A come in just yet, Christian, but that doesn’t mean it won’t come in. Folks, if you have specific questions for Christian on how to leverage those ABM strategies, just looking at the stuff you’re going through with the deep research on chat, I don’t know man, the times are changing, Christian, just how quick you can grab information on companies or profiles is absolutely incredible.

This comes in from Sebastian, after you write a super personalized email with this method, how is the first call, do you say something, do you acknowledge, like, hey, I just had my AI friend stalk you for like 30 seconds, do you break the ice in that way?

Christian Banach: No, I mean, I don’t think that, it’s never come up where somebody, I don’t think the emails that we’ve been able to produce recently come across as AI, and I don’t feel that’s even necessary. I mean, it’s still being sent by a human, no different in my opinion, we are using spellcheck and Grammarly and all these other tools, and we don’t make excuses for why Grammarly helped us write a better email, or why spellcheck helped us. So no, I don’t think that that comes across, these are still emails being sent by humans, from out of a human’s email box, and they’re fielding the responses, we’re not using AI to respond to those emails. That might be talk for next year that I could give, I’m sure that’ll be there at that point, but no, that’s not something that we do. I mean, it’s 2025, we’ve got AI, we’ve got automation, and we’re all still tethered to our inboxes, maybe there’s a future where that isn’t the case.

Brent Weaver: This question comes from Nicholas Williams, what tools are you guys using for content writing, so I know you kind of featured ChatGPT quite a bit in today’s talk, any other tools you’re using for writing for these types of outreach?

Christian Banach: Yeah, I mean, we’ve experimented with quite a few, and every other week it feels like somebody comes up with some new feature, just to keep things simple we’re just doing everything in ChatGPT right now. I think there are times where Perplexity might be a little bit stronger in certain types of writing, but through at least this process we’ve been able to get what we feel is really good, strong writing, especially from an email perspective, out of ChatGPT. But we’re always experimenting with others, I would encourage you guys to experiment as well as different tools come on the market and different features become available.

Brent Weaver: Any tips on how to use AI to avoid the email getting sent to spam, thanks Angie?

Christian Banach: Not necessarily. I mean, what we find though is when you’re loading up and sending the same message to hundreds or thousands of prospects, the email providers recognize that it’s the same email going out over and over again, and that’ll hurt your deliverability. But we have seen this approach improve deliverability, because these are all custom emails, custom subject lines, just like again a human would be sending, you’re not sending the same email over and over again to prospects. So I don’t necessarily say that there’s some AI hack that we’re doing, it’s just by nature of sending not the same message over and over again, you’re landing in the inbox more.

Brent Weaver: Very cool. Well, Christian, we’ve got a fun session coming up here in just a second, I want to make sure our audience has plenty of time to participate, but just want to say thank you for this session, majorly valuable, we have it recorded, it’s going to be published as part of our two-day event here at Agency Advantage. Let’s give Christian a round of applause here folks, give him some love here for his talk, Sebastian says super great talk, agreed Sebastian, thank you. Christian, thanks again man, we will hopefully see you back here again soon.

Christian Banach: Sounds good, thanks everybody, take care, Brent.

Brent Weaver: All right folks, we’ll come right back with our Future of Agencies Illustrated by AI challenge, so we’ll be right back.