Welcome to episode 92.
This episode is for you if you’re curious to know how an SEO strategy agency is using AI tools to reduce the amount of repetitive, tedious tasks and free up the team to work on more of the strategic thinking for their clients.
Dale Bertrand, Founder and President of Fire&Spark shares with me:
• How AI tools have reduced his team’s workload and improved efficiencies
• A step by step breakdown of his internal automated workflow
• Why creating predictive models might end up being your agency’s secret sauce
• What Dale now looks for when hiring new team members
I think you’re going to enjoy this chat and hopefully come away with some inspiration to look more closely at your internal processes to see where you can speed them up or eliminate them altogether.
I’m now sharing examples of relevant AI tools that account managers can use for those who come on my Account Accelerator training programme. The training is designed to help those in agency client service at every level to have a more systematic and predictable approach to client growth. The client-centric learning outcomes include being more consultative with clients, adding more value proactively and having sales conversations without feeling salesy.
If you or a member of your team is interested to find out more, all the details about the 9 week Account Accelerator programme as well as my shorter 1 week Account Booster programme are here on my website:
Account Accelerator
Account Booster
Transcript
Jenny Plant
So today I’m excited to welcome Dale Bertrand to the show. Now, Dale is the founder and president of Fire and Spark, an SEO strategy agency focused on conversion, based in Boston. Dale holds a BS and an MS in electrical engineering from Brown University with a focus on AI and computer engineering. So he applies his knowledge of AI to digital marketing. So not only has he worked as a digital strategist for fortune 500 companies and venture backed start-ups for over two decades, he’s also a trainer and speaker at marketing industry conferences on the topic of AI and digital marketing. So I’m really excited to get into some questions for Dale and his thoughts for where AI is going, particularly as it relates to what we do as agencies. Dale, first question, right? In addition to founding your own agency, you’re obviously speaking at all of these events and delivering AI training. So could you share with us the key topics you cover in your teachings and who you are primarily addressing when you are talking about these AI and digital marketing?
Dale Bertrand
Yes, there’s a lot there. It’s important to put it into context just with my agency journey, because I started doing freelancing and then building my agency in 2008, and around then was when we had a platform shift from the desktop web to mobile, and a lot of things changed. And we have been working on SEO because we are an SEO agency, we do SEO strategy. So we have been working on SEO for a long time, and that has changed many times over the last 15 years that I have been doing SEO. So I’m thinking of the AI revolution that we are in the middle of right now. In that context, it’s another platform change. If you work in digital marketing as an agency, your job is to keep up with the changes in the industry. So a what I’m talking about is my keynotes and workshop conferences and a lot of what I’m talking about at my keynotes and conference workshops really has to do with putting the technology into perspective.
Like, how should we approach it, making sure that marketers understand the capabilities, and then also agency owners and leaders at agencies understand how to introduce these technologies properly to their team so that the team doesn’t feel threatened that they are going to be replaced by AI. And then also the team feels empowered to apply the technology to automate or enhance in the right areas, in the areas that would help them to do better work and get more done. What I love about at my agency, how we have implemented the tools, is that we are automating tedious work. This is work that our staff came to us and said we would love to get this off our plate so that we can do the stuff that we love and that has really made more time for the creative work that my team enjoys. And that’s really the way we are thinking about it.
And then when you have these tools like automating, a lot of what you do with marketing automation sorry, when you have these tools automating a lot of your marketing ops, what ends up happening with the humans at your agency is that they are responsible for generating more insights, more ideas for campaigns, more things to try. And those things to try, some of them will actually work. So when you are operating in this mode where I can come up with an idea for a campaign, a piece of content, a cold email that we are going to send out for biz dev purposes, when I come up with an idea, we can implement it quickly. So I need more ideas to feed this machine. But that’s the fun part of the job. So we really want to make sure that we are applying these tools in the right way.
Jenny Plant
Fantastic. So there’s probably some SEO agencies listening in, thinking Dale’s managed to automate tasks to free his team up to do the thinking, creativity and insights. Can you give us some examples of what the tasks are that you have automated to give the listeners some ideas of how they could, if they haven’t already, start using AI?
Dale Bertrand
It’s so much because we run accounts, we work with clients, so
we want to be able to generate insights on a daily basis for different client accounts, looking at the data that’s generated from the campaigns and then also their website, their content performance. So we have automated the process, we have bots that run that will drop an email in my Drafts folder, my Account Manager’s Drafts folder, basically to a client about something that we notice and we have the opportunity, like I could delete it because it’s no good, or I could press send on it because this is something my client should know about. Or I can make some revisions and then press send. But that is the type of thing that we simply didn’t do for the most part before we had these tools. So a lot of times when we are thinking about these tools
we are thinking, I can get my current work done faster.
I can generate more articles for my blog than I had in the past. But I really want to think in terms of capabilities that are new. Like, what is it that you are not doing that you should be doing? And now, because it’s so much easier with these tools, you can do it. Another example is we are often investigating a change in organic traffic to a website. It might be a drop, it might be a shift from one section of the website to another. On very large websites, it can be hard to figure out, well, what is the pattern here? Why are they losing traffic? And we can use tools to basically look at data coming from Google Analytics, and then also Search Console, and then also SEMrush to understand why they are seeing a shift in traffic. And that boils down to something like, well, you are no longer authoritative for these types of topics, or this type of content on your website is no longer ranking.
And there’s so many things there. And that’s the type of thing that I used to do. It takes hours to figure it out, write up a report, and man, have we cut that time. So I do coaching for agency owners. One of the agencies that I work with, they are in London. I presented this exact use case and how I do it to their founder, and he said, Dale, you just reduced my workload by 20 hours. Like, you cut my workload in half by showing me how to use these tools. And I was like, Holy crap. But that’s what we are seeing too. Our goal is we could get a four X improvement in terms of productivity. And then how do we spend that extra time? We spend that extra time, folks at my agency, coming up with more creative ideas, coming up with more things to try, and then also doing things we didn’t used to do before.
I’ll give you another example. There was a time that we have an SEO methodology that we call Purpose Driven SEO. It’s all about us working with Purpose driven brands and then using their purpose to build their authority and their organic traffic. So what we did to promote that was we did interviews with marketers at Purpose Driven Brands, and it worked out great. It was a lot of fun, you know, we got eyeballs on known leads for the agency. So it was good. The problem was, it took so much time, Jenny, as you know, because you have to find the guests. You have to research the guests, figure out what the questions are going to be, do a little pre conversation, and then you produce the actual interview like you and I are doing right now, live recording and then afterwards, it’s like, all right, turn that into content. Send it out. It was so much work that we decided, we don’t think the ROI is there just because of so much work behind it, even though we really enjoyed it and were getting leads from it. So fast forward today, and we are seriously considering restarting that campaign because we can automate the research, inviting interviewees, like writing some questions, and then the back end when we are promoting all the content from it, that stuff is automatable now.
Jenny Plant
And so tell me, are these tools that you’re using available to anyone? Or have you somehow, with your knowledge of AI, got some kind of bespoke program that you’re using?
Dale Bertrand
So the way to think about it is that there are a number of platforms that allow you to do different things, like large language model like GPT4 there’s a lot you can do with that. And then a “Pinecone” database where you can give it text articles that you have written, or articles that our clients have written. And then it’ll pull the information you need when you’re prompting your LLM. And then there’s other tools, like ML tools, Auto-ML tools that allow you to build, to basically spin up prediction models. So
when I’m choosing keywords that I want to target, I want keywords that I know this particular client can rank for. So I’ll have a prediction model that predicts how easy it will be to rank based on data that we have gathered for that particular client or for clients in that same industry going after similar keywords.
So that’s a prediction model, machine learning prediction model. And then there are also, like mass personalization components. I’ll give you an example of that and this is a long winded answer to your question, but I’ll give you an example of that. Mass personalization as an example when I first started using it, I took all the emails that a friend of mine had sent me. He runs an agency, and what the database does is it takes those emails and it generates observations about that individual. So for this particular individual, my friend Justin, it said he’s a parent, he’s an agency owner, he’s a programmer, and then a whole long list of preferences for him, a whole long list of observations about Justin. So then what that allows me to do is when I’m sending him any correspondence, it might be an invite to one of my events or our newsletter or a solo email that I send out to my list.
We can personalize it based on those observations. So we would give some copy or a template to an LLM like GPT4 with the observations for this individual. And it’ll customize like two sentences at the top. And then it could also curate the actual links or content pieces that go in the email, if that’s the type of email that you’re sending based on the observations from that individual. And you can build those observations from any text correspondence. So it could be emails, it could be chat logs, that sort of thing. So then we have these and your question was about, well, how are we orchestrating all of this? So we have these base level components, the LLM, the prediction model, the preferences database, and then also this semantic similarity search database that we can put content in and then we can orchestrate it in a number of different ways.
You can do it in a Google sheet and just make calls out to these things to grab the information you need to run a prompt and give you the response. Maybe the response is a bunch of article titles or something like that. Or the response out to your prediction engine might be grading keywords on their ranking potential, something like that. But we do that orchestration of all those components in either a spreadsheet or in Zapier. I prefer Make.com, very similar to Zapier. And also I have a programming background, so there are some automations that are written in Python and you can use AI tool to help you write that, which is crazy nowadays, but that’s how I’m thinking about it. The base model layer that we are using to implement these AI automation workflows and then the tool that sits on top of it that is orchestrating any one particular automation.
Jenny Plant
So how many other agencies are actually that sophisticated? Because what you have just described, those four key elements, you are obviously very experienced with AI, and I’m sure for you it’s a lot easier to sort of navigate to get those four components working together and then orchestrating this whole flow automatically. But in your experience of delivering these conference speeches, how many people, other agencies, for example, have got it so dialled in?
Dale Bertrand
I think I have met several that are headed in that direction. So I attended Marketing AI conference last week in Cleveland. Great conference for anybody listening who really wants to dive into the world of AI. And what I was surprised about is like 99% of the people there were just getting started. Like they were just playing around with Chat GPT. And I did meet a dozen folks who were pretty advanced, which is awesome. If there are any advanced folks listening to this, reach out to me because I’d love to network with more folks that are doing advanced stuff. Anything beyond just Chat GPT. I met one agency that really impressed me because they had recently hired a full time AI automation specialist and I thought my agency was the only one.
We have a full time person on my staff who’s responsible for AI automation.
So what we have her doing is talking to people on the team about what they do, what’s tedious that they would love to automate. I’m teaching her what I know about the base models and then also the orchestration layer that sits on top of it. Another thing that she’s doing, our in-house AI automation specialist is she’s looking at all of the training that we have. Like we have a ton of training for our staff in house and going through it and saying, okay, this step could be automated this way, this step could be automated that way, just so that we have an understanding of where the opportunities are. Once you understand what the capabilities are, there’s a switch that flips in your head because it just becomes so clear to you what the creative work is. The work that requires real human deep thought and inspiration, and then the work that is more just doing it’s either really tedious or it’s moving things around that a human just doesn’t need to do. And any time that you spend as a marketer or businessperson, entrepreneur on the doing side, the stuff that, you know, can be automated just really hurts.
Jenny Plant
What have you found have been the biggest barriers to people? Because I know what I’ve been hearing, but I’d love to hear it from you because you are absolutely right. This sounds like either something that agencies can do themselves. i e. Audit the tasks you are doing, find out where your team are spending time on tedious tasks, as you say, or you hire someone whose dedicated role is to do that, to go through the whole systems and processes of your agency, identify it, coordinate it, and then focus on automation, which seems to me I mean, I suppose I’ve got multiple questions. How difficult is it to find someone with that skill set currently?
Dale Bertrand
Well, they don’t really exist because this is all pretty new. So the woman on my team, I’m training her on everything that I know, and then I’m working with a few agencies, helping them do it. So that’s a great learning opportunity for them and learning opportunity for me and for my in-house AI specialist. Yes,
I think really what it comes down to is now, at this moment in time, you just really want to hire people who are predisposed to be curious about AI and what it can accomplish and shy away from people who, for whatever reason, are predisposed to avoiding the new technology. We saw the same thing when it came to digital. So if you were a marketer working in marketing around the time that digital became a thing so I’m thinking, like, early 2000s, there were some marketers that said, you know what? I’m not making the switch.
I’m going to do everything traditional marketing. I’m not going to worry about online UTM codes. That’s just not what I want to do as a marketer. And I’m going to ride this out until my career is over. And I think somewhere around 2010, 2015, those people retired. The transition to AI is so much deeper, so much broader, so much quicker, that it’s not like, you know AI is not going to put your agency out of business, but an AI first agency is going to put your agency out of business. So when I talk to agency owners, I would just say, like, hey, you got to figure out what does it mean to you to become an AI first agency and that means something different to every single agency. There’s some agencies out there that are doing brand strategy work and then others that are doing paid advertising online or SEO strategy like we do.
It’s just different for every agency. I’ve got a clear idea what it looks like for our agency to be AI first and I have to say we have an absolutely incredible team. We are 25 people here in Boston. How long ago was this? This was the beginning of the year.
In January, I did a presentation to everybody. I said, here’s the capabilities of AI. We are really going to dive into this stuff. We are going to try to make ourselves two times as productive over the course of 2023 and we are on our way. And I said, “who wants to join the AI task force?” and everybody raised their hand. And so we have had a systematic exploration of different use cases. Like how do we use it for keyword research? How do we use it to identify traffic drop or diagnose a traffic drop? How do we use it to help us build proposals and respond to RFPs? We have got a list of 100 different tasks and use cases and we prioritize them by impact. So if we were to automate this or make it better, because when I say automate, people think, oh, this thing will now take less time, it’s so much more than that. If we were twice as productive, we are still going to have the same headcount, but my people are going to be doing more awesome work.
Like, I would love to free up my account managers, they are freaking smart because otherwise we wouldn’t have hired them. And then also they are doing excellent work for some global brands. So what I would love to do is free them up to help us generate more content. On the content side of our business, a lot of people talk about everybody writes until you got a full load of clients and it’s like, all right, how do I write?
So that’s something that I have been thinking about. I’d love to make the folks on our team more productive and then have them help out on the marketing side of the business. That would just be a win for everybody.
Jenny Plant
It sounds totally doable and I could see from your enthusiasm, I’m sure that is translated to the team as well. So that’s why they are like, “I want to be on the task force”. So I’ve got two key questions because in my experience, having recently run a webinar for agencies called AI For Account Managers, one of the big themes that was coming through from people listening in was what about privacy? Like currently, particularly the likes of Chat GBT, whilst you can put it in privacy mode, do we trust that’s not going to then share information? So on the one hand, I totally understand if things are out in the public domain and it’s published data and you are looking at either summarizing pieces or utilizing the data that’s already been out there. Fine, but what about the inputs from anything that you know about clients information? I mean, what’s your view on maintaining privacy so you don’t get into trouble?
Dale Bertrand
The important thing there is that you are going to be using – at least at Fire and Spark, we are using a number of different components. So LLMs like GPT and GPT 4, the database I told you about, prediction models, personalization engine. And each of those components guarantees a different level of confidentiality. So if you are using the free version of Chat GPT on day one when they launched it, the reason why they launched it was to help them improve their model. So read into that they are using the information you are giving it, they are using the information that you are giving it to train their model. And you have to be careful about that. That’s why people are concerned, that’s why you are asking the question. But now that there’s so many more models out there, it’s really a matter of choosing components that offer the level of confidentiality and security that you need for the work that you are doing and your contractual obligations to your clients. But there are versions of all of this stuff that is confidential.
Jenny Plant
And would you say that because I know what it’s like going through terms and conditions, they go on forever and it’s very small type. I mean, is the average layperson that’s working in an agency really able to judge that? Or should they seek legal advice because the stakes are high, aren’t they, at the moment? Because we are early into this and as yet lawsuits for example, we have already seen some book authors who are trying to take the likes of OpenAI to court around they have used their books as training material for their models. So it’s kind of up in the air a little bit, Dale, isn’t it? Do you agree?
Dale Bertrand
Right now, it feels like the wild west in terms of there are things we are doing today that we don’t know if we will be able to do tomorrow, we don’t know if they will be legal to do tomorrow, and we are not sure if they are legal today. So we are in this situation where we know these tools are incredibly powerful. We don’t want to be left behind. So we need to be learning them and understanding their capabilities. But at the same time, the legal frameworks around these models have yet to be fleshed out and probably take another year to really flesh that out. What I would recommend is making sure that you, – so agency owners need to think about what is your policy? And this could be five bullet points around how we protect client information, how we honour contracts and obligations around confidentiality to clients that we are working with. And I think the best way to go about that would be you could go online and search for basically AI guidelines that a number of governments and institutions -. I know the Mayo Clinic came out with one. The city of New York City has one for its employees. There’s another one. The city of Boston has one and it’s just basically one page of guidelines on how to use these tools. And for agencies, right now, what I would recommend is just staying away from using tools that are going to end up training their models on your client’s data. Like that’s the big thing that you want to stay away from. I’m not a lawyer, so check with your lawyer. But there’s less of a legal risk that anything you produce with AI is going to open up legal liability. I couldn’t imagine that happening. What’s going to happen is there’s going to be a legal fight between the tech companies that are making the models and then publishers. And I’m sure the tech companies will end up paying the publishers something. But also the other thing that I want to point out is that there’s an assumption buried in your question that what we are using is generative AI tools, but there’s predictive AI tools and there’s also AI tools that allow for personalization and there’s AI tools that allow for automation. So the legal complications there will be different. So it’s not like when we are using prediction or automation use cases, we are not going to get sued because we published some content that was generated by a model that was trained on somebody else’s content. It’s really a different use case. And I just want to point out that a lot of the way I’m using these tools, even when I’m generating text and content from it, is I’m pointing the tools to something that we have already written and we are repurposing and we are telling the large language model, don’t use any information that’s not in this e-book that I wrote. Something like that. And there’s still the possibility of some legal complications there. So everybody should kind of think through your own exposure around what you are doing. But there’s less likelihood that we would release something that’s plagiarized. And then obviously you want to run a plagiarism checker on everything, which we have always done when we work with writers.
Jenny Plant
Okay, that’s really good advice. And for those listening that are looking for someone to steer them, I had Sharon Toerek on the podcast last week and she’s someone who’s out there speaking about AI and the legal implications. So I’ll put some details in the show notes. Dale, let’s go back to your orchestration and this automation predictability model that you have created for Fire and Spark. If there’s an agency owner listening, thinking that this is super smart, I need to be doing that, will Dale help me? Is that something that you can go and train other agencies on or is that proprietary to what you are doing in your agency?
Dale Bertrand
Oh, I’d be happy to talk to agency owners and at least point them in the right direction. There’s a number of tools that we are using that I’m happy to share and show people how we have done the high level orchestration and the results. I mean, in one of my workshops, we show like, personalized cold outreach emails and how one campaign, we were sending a cold outreach email that was kind of a template mail merge style template with a 1% positive response rate. And then when we started personalizing them, 19% positive response rate, which is like through the roof. I mean, that’s insane. So yeah, I’d be happy to talk to folks and just it’s fun for me to understand what type of AI implementation would be most impactful for your particular agency, for your team, for the work that you do, for your clients, and then point you in the right direction in terms of tools that you should look at.
Jenny Plant
Fantastic. Well, that’s great. We will put your details in the show notes at the end. What can you foresee happening over the next five years, particularly as it relates to AI changing and the SEO game?
Dale Bertrand
Stuff is going to get real, because there’s so many variables. We don’t know the extent to which these tools are going to get even better. I mean, oh my God, they have got so much better just over the last year. Imagine five years from now and the truth is we don’t know. We know they are going to get better, but are we going to have super smart AGI, which is artificial general intelligence that starts telling us what to do? That’s an actual possibility. I think it’s unlikely, but it’s an actual possibility that we could end up there in the next five years. When it comes to agencies, there are always going to be companies with money that need advice, need people to run their campaigns, that need experts that understand what they are doing, understand these channels, understand these tools, and can apply their learnings to their business. So that is always going to be a thing. But what I see or what I envision is that the line between agency and client is going to change. And we don’t really know yet. But you think about the world of SEO strategy where there are more and more things that the clients can do on their own. They are still going to need the highest level strategy from us. And we have always focused on strategy rather than implementation, but the type of advice we are giving could definitely change. And I always want to make sure that I’m highlighting that there will be new capabilities and to give you a good example here, it’s possible for agencies like mine and the agencies listening to this presentation, it’s possible to create predictive models that are your secret sauce. Let’s say, for example, one of the agencies I’m working with, they focus on the education vertical, so they have the opportunity using AI tools to use the data that they are generating as a part of their campaigns, all targeted in this one vertical to create a prediction model.
A model that predicts the right topics, that will perform, how to write content, that will convert, what offers to send out to students and parents in the education space that are working. That’s a closed loop prediction model. And because that agency is focused on working with clients in that vertical, they have access to data that’s proprietary, and they can go to the market when they are doing their business development and say, we have built this predictive model. Here’s our case study showing how it accelerates enrolments or engagement with content, whatever it is. So a lot of agencies should be looking at building that type of predictive model as a secret sauce for their agency. I’m aware of three that I could think of that are doing it and we are experimenting with a number of different ways to do it with our agency. But I think that is how agencies are going to build a mode around their business because it’s access to that type of data.
And the more specialized you are, either horizontally- or, the more specialized you are, which would be either like horizontally for us, that’s SEO strategy, or if you’re vertically positioned in a particular industry. The more specialized you are, the more you can generate that niche proprietary content that you can use, and the more you could generate that niche proprietary data that you can use to train up a prediction model that would be proprietary to your business and really set you apart.
Jenny Plant
And if someone’s listening to this and thinking, well, that’s me, I am vertically positioned, and I could do that, are the tools available right now? And if so, where’s the starting point?
Dale Bertrand
Well, no, they are the starting point. What I hesitate to do is send people straight to a tool, because the answer to this question isn’t really a tool, because there are many of them. This stuff is easy now. It used to be hard. I spent a summer when I was a student building out an ML model for the US. Government, for the Department of Defence. That took a summer. Now you can import a spreadsheet into a tool, and it’s going to use Auto-ML to find the right algorithm, build the model, and then you have access to it. So it’s just so much easier now. But what you want to do is you really want to think about what prediction problem is, what data you have, how you want to present that data, and then what exactly it is that you’re going to end up predicting.
And I’m happy to have that conversation with people who want to move in that direction. But the work is really in understanding the capabilities, making sure you have enough data, making sure that you have the right data, which is basically the right columns in a spreadsheet. And then you feed it to one of many tools that can help you do this.
Jenny Plant
But that data, presumably, again, is it data that you have collected from working with clients? And therefore, is that a little bit risky? Or is it the data that you are using that’s been externally generated, that’s already kind of out there in the public domain, like, just for a layperson like me?
Dale Bertrand
It is all of the above.
Jenny Plant
Okay.
Dale Bertrand
Data is data. So you could imagine trying to figure out, like a use case would be imagine you are working in a particular vertical and you are running Google Ads campaigns, and you are interested in understanding, well, what types of keywords could I bid on that would convert? If you have a lot of data, for example, the search terms report in Google Ads that will have a search terms column, which is a bunch of keywords, and it’ll have another column around the number of conversions that those keywords have generated on your website. You could build a prediction model from that. And what that prediction model would do is grade additional keywords. So anytime you’re thinking about targeting a keyword in one of your Google Ads campaigns, you could run it through that prediction model, and they’ll say, well, based on its similarity to other keywords that either converted really well or didn’t convert at all, here’s how well the prediction model expects that these keywords would convert, but that’s conversion.
And staying on this example of keywords, there are other attributes of keywords we care about, which would be how well they might rank when it comes to SEO. So if you have the right data, that could be data that’s coming from search console accounts. That’s where you would want to make sure that in your client contracts, people understand what you’re doing. Or it could be data that is coming from more publicly available, from like, SEMrush or something like that, where you would have less of a concern contractually.
Jenny Plant
Got you. Okay.
Dale Bertrand
But when you start to think of it that way, there are thousands and thousands of different data sets. I talked to a client yesterday. They are in the real estate niche, and we talked about some real estate, like, publicly available government data that they can use to build a prediction model. So there’s just so much out there that are possibilities and then there’s another client, so I’m advising a start-up in the insurance space. And insurance is highly regulated, so there’s a lot of data out there, so there’s a possibility there. And then yesterday I talked to an agency, they are a brand strategy agency and what they wanted to do was when they start working with a client, look at all of the deals that the client has won or lost over the last two years, and then grab data on every single one. Of those businesses and you grab that data from LinkedIn, put that into a spreadsheet, and the spreadsheet will have basically deals we won, deals we lost and here’s all of the information about those companies and then use AI to find correlations, like correlations around who was the contact at the company, how big was the deal size, what did they want, what industry were they in, how big was the company, all that kind of stuff and it’s just easier now to use AI tools that can find those correlations for you and identify patterns and features in the data that you may or may not be able to find as a human.
Jenny Plant
Very interesting. That was fascinating. And I’m sure a lot of people are taking note of this. Building your own proprietary predictive model that specializes in your niche. What other tips do you have for agencies to prepare them for what’s coming?
Dale Bertrand
Yes, I mean, I’m always thinking about new services that you can offer that just weren’t possible before this. And the prediction models is kind of part of that. Another one is services around mass personalization. It depends on what your agency does. But if you are an email agency, you need to figure out how you are doing mass personalization, because that’s like the first and easiest way. But I talked to an agency that was doing text message, so they were doing SMS marketing and they are just getting started with mass personalization for SMS messages. There’s a lot there.
Jenny Plant
Okay, I’m going to go back one step because SEO agencies are based on ranking on Google usually, isn’t it? I mean, I know there are other search engines, but Google is the big one. You can tell I’m a layperson talking like this, but in the future do you see that Google is going to be the first point of search? I mean, obviously it dominates at the moment, it’s become a verb, “Google it”. But in the future with all of this AI technology where for example, you are on Chat GBT 4, you ask it a question and it will give you some links to sites.
Dale Bertrand
Yes.
Jenny Plant
Do you see in the future that people are going to be initiating search on Google as much as we have been used to?
Dale Bertrand
So there’s so many pieces to that question. The first thing is that Google is a trillion dollar company. They will do everything within their power to maintain their market dominance. And that’s so that they can maintain the cash volcano that is their search advertising business. So they will do what they need to do to maintain that business. But what we don’t know is what does organic search look like when they do what they need to do to maintain that business? I believe that they will roll out something very similar to this generative search experience that all of us have probably seen. That’s what you’re describing, Jenny, but I don’t think they are going to make it default. That’s the big thing that we are trying to figure out. Is it going to be a tool people can use if they go out and look for it, or is it going to be the default experience when you type Google.com, that’s what we don’t know. But what I believe is they are not going to want to make it default unless they have to. And I don’t think they have to. I think they want to maintain the status quo as much as possible. That’s why they were behind OpenAI and Microsoft and Bing for a short period of time. It’s going to take years for that to play out, and I mean like two or three years for that to play out, not months. So we have got this weird disconnect where the technology is moving very quickly. Google invented the technology, by the way, but actually rolling it out to consumers, especially in the search engine, and then changing consumer search behaviour, that’s going to be very sticky and very slow. So the way I’m looking at SEO going forward, I mean, I run an SEO agency, I expect to be doing SEO ten years from now.
So that’s super important. We look at SEO as basically generating conversions, sales and revenue from high intent searches. If you’re paying for its paid search. If you’re not paying for its, organic search. As search engines evolve, there will be different techniques, and we are staying ahead of it by identifying those strategies to basically win customers, win sales from organic search. But the thing we don’t want to forget is the reason why organic search has been so powerful for so long is that there’s intent behind those searches. So if I’m going to Google as my primary way regardless of what Google looks like in the future, if I’m going to Google as my primary way to find vendors for whatever service that I’m looking for, or products, that I’m looking to buy that is super valuable. And whoever can stay on top of the SEO game, even as the search engines evolve, is going to have valuable services to offer. Or it could be a brand that’s figured it out. So we are committed to figuring it out as those changes happen. And I don’t expect that it’s going to change as radically as it could and it’s going to take a while.
Jenny Plant
Okay, fair enough. Thank you so much for answering that question, Dale, this has been super valuable, are there any parting words of advice for agency owners and also the account managers that are working in the agencies?
Dale Bertrand
I think the AI tools that we talked about for agencies can be super powerful and just committing yourself. What I would recommend agency owners do is commit to becoming an AI first agency, whatever that means for you. It doesn’t mean you need to go to grad school. It doesn’t mean you need to stop doing the work that you love and you enjoy and somehow automate it. What it means is just trying to figure out what your relationship with this technology is going to be and then embracing the parts of it that do make sense for the use cases that do make sense for you. And that can be so powerful in so many ways. Building out proprietary services, automating work that you’d rather not do, coming up with creative ideas, testing more of your creative ideas, like going live and shipping more of your creative ideas.
That’s an exciting future. That’s not a scary future when you really understand the capabilities of the tools. And anybody who’s interested, like I said before, I love having these conversations, so definitely reach out to me. Happy to talk.
Jenny Plant
And where is the best place to people to reach you? And who particularly would you like to hear from?
Dale Bertrand
Well, so anyone who’s working, agency leaders, agency owners who are thinking about how to apply these tools to their particular situation. That’s a really fun conversation because I can point people in the right direction and feel free to email me directly. So I’m dale@fireandspark.com. All spelled out, so feel free to email me directly or connect to me, Dale Bertrand on LinkedIn. Happy to have those conversations.
Jenny Plant
Amazing. Dale, thank you so much for your time. We’ll make sure to put all those links in the show notes. Thank you so much for coming.
Dale Bertrand
Awesome. Thank you for having me.
