According to Gartner, worldwide AI spending is forecasted to increase 44% by the end of 2026. Companies are investing in AI, and they are investing heavily.
But knowing where and how to invest isn’t easy, especially with what feels like a million different AI tools out there and a million more different ways to build your own. So how do you figure out what to build, what to buy, and which investments will help you move the needle for your business?
Riley Rogers: Hi, and welcome to the Win/Win Podcast. I’m your host, Riley Rogers. Join us as we dive into changing trends in the workplace and how to navigate them successfully. Here to discuss this topic is Cody Sims, head of commercial brand at Cox Communications.
Cody, thank you so much for joining us today. Super excited to hear your thoughts on this one. Before we dive into what is quite a loaded topic, could you tell us a little bit about yourself, your background, and your role?
Cody Sims: Yeah. So, hi, I’m Cody. I’m the head of commercial brand for Cox Communications, and it’s kind of crazy how I came into this role.
So I actually started my career when I was 15 and was installing phone systems for my dad’s phone company. And after that, I had actually had two parts of what I thought was what I wanted to go into, and that was either musical theater or physics, because those were two things I really had a passion about.
And when I got into college and had musical theater as my major and physics as my fallback, I realized that both of them left a part out of what I really enjoy. And so I ended up actually landing in marketing because it’s both analytical and creative, and that has served me really well over the years.
So, across Cox, I have done all kinds of things from product management to market development to pricing to competitive analysis, and now in the brand world. It’s given me kind of a 360 view of the entire business from a marketing lens. I would say that I’m pretty much a transformation leader. I really enjoy breaking things and building them up new again.
So, AI is happening right at the right time for me.
RR: I love that story, and I love that it’s taking you to a place that especially now is getting more and more technical, more and more analytical. I’m very excited to get into that transformation leader side of things.
But before we do, can you paint a little bit of a picture of your sales environment?
CS: Yeah. So when I first came to Cox, it was very similar to most of what you would call a CLEC, or a competitive local exchange carrier, which is primarily internet service, voice services, and obviously because it was Cox, some cable TV services that were the triple threat. That was kind of what in the early 2000s was kind of the way that they went to market.
But over time, the team at Cox realized that in order to stay competitive, they had to add to the portfolio to make sure that they were providing value to their customers, and I’m sure many would understand that and have gone through similar transformations. And so we had acquired several different other companies that added to our portfolio, and developed some of our own products, and over time that turned into a lot of products.
But it’s not just 70 products. It’s 70 products, it’s nine customer segments that we have from a segmentation perspective. It’s six distinct buyer personas, industry verticals, what’s serviceable at that address. So you take all of these different components and it’s almost like three-dimensional chess for the salesperson.
The way that I like to think about it is that the sellers, what they really need and what their challenge is, is that they’re not looking for specs. They’re looking for what are the business outcomes that my customer is trying to achieve, and then what do I have from my portfolio that will help them to achieve those results?
So it’s no longer a world where they can memorize everything and know every product in and out, and be the technical expert. They really do have to have tools and systems that help them to have the right knowledge at the right moment for the right person in the right place.
RR: Yeah, there comes a point when the human brain just can’t contain the context and the expertise that you need. So when you can’t ask for expertise, what you can do is provide, to your point, that just-in-time support.
And one of the things that you alluded to was that that’s where you kind of started some of that AI investment as a way to bridge that gap. You’ve given us a little bit of a taste, but what kind of motivated that early initiative?
CS: Well, I would say that, over time what we discovered was that we couldn’t keep track of all of our marketing materials, collateral, all of the pieces of information in just files, formats, and putting it online into a here’s-an-accessible-library. Because the library just becomes bigger and larger and more difficult to manage.
But I would say that we didn’t set out to do AI. We didn’t sit down and say, “Oh, hey, AI looks cool. Let’s make sure we’re doing it.” We needed to transform our go-to-market strategy so that we were more nimble, we were more competitive, and that we could deliver the kind of experience that our customers were asking for. And so AI was the mechanism that would help us to get there.
But what really triggered this whole thing was what I mentioned earlier, was our segmentation. When we sat down and said, “Let’s rebuild the way that we look at our audience segments,” and we did that based off of what is the value to Cox of each of these customer profiles, and then what is the technology sophistication of that client, of that business.
And that intersection allowed us to create our nine different segments that we were working on. And so when we did the math, when we looked at all of that information and all of the things that we needed to be able to provide to those segments, we realized this was quickly going to turn into something that was far beyond any marketer’s ability to do.
But what we knew is that the Gartner information we were tracking said that personalization was having much higher returns on the way that people respond to information. And not only that, but if you do personalization and you get it wrong, if I call you and I say, instead of, “Hey, Riley,” and I say, “Hey, Jonah,” you’re like, “Hmm, nice try.”
So personalization is really important to being successful, but getting it right is even more important. So we realized that we needed to have some radical partnership between our marketing, AI, and sales teams, that we needed to make sure that this was not just an IT project, that we were going to go and pull a bunch of requirements together and everybody would be like, “Oh, hey, here’s this new tool. Everybody figure out how to use it.”
And it wasn’t necessarily about optimization. It was about transformation, the way that we go to market, the way we think about our customers, and the way we show up. So I would say that AI definitely was part of the solution set, but we had to look ourselves in the mirror and say, “It’s time for us to actually think about this in a completely different way.”
RR: That distinction comes through very well, and I think is very important because oftentimes when you’re in kind of the scramble to be keeping up with the market, keeping up with your competitors, there is this urge to just tack on AI because we have to.
But when it’s not strategic and it’s not built into the things that you’re actually doing, to your point, it’s we put together some specs, good luck using it. But instead, now it’s something that’s really built into the way that you work.
I would love to hear a little bit more about that specific use case, especially given the fact that a lot of teams are running into that question of how do we use AI and can we just build what we need ourselves? Given that you’ve done the math, answered the question, I’d love to hear how it worked and kind of where you landed.
CS: It’s very easy to fall into the trap of, “Hey, everybody, here’s AI. We put it on your computers, now go use it.” And so then everybody starts using AI to try to figure out, how does this help me in the job that I already do, in the role that I already do, in the processes that I already do?
And so then it really limits the impact that it can have on the business and the performance because either, A, you have a handful of people who are really smart, and they go crazy with it, and they create their own thing, or you have a bunch of people who are looking at it going, “Okay, came up with some ideas, but I still have to do my work.”
What you end up with is there’s no standard. There’s no flag running up the hill to say, “Everybody follow me. Let’s go do it this way.” So, it required both the yes, we had to make sure that the teams were bought into using AI, but we also had to have a standardized way of approaching how we deploy AI. And that brought us to the question of do we buy or do we build?
And because there were so many different parts of what kinds of functionality we needed, it wasn’t the same answer for every one of those needs. So in some cases, we have a tool, we have a partner, they already have AI integrated into their platform, let’s go see how we can use that.
In other cases, and I’ll give you an example, in the case of content generation, that is where we started with our AI journey about a year ago. We sat down and started interviewing and reviewing all of the different providers who can do content generation. Every one of them had a different approach to content development, content generation, which were all very good, and they attempt to make sure that they are covering as much of the marketplace as possible.
And so sometimes when you buy that, you end up with features maybe that you don’t need, and you also have to still go through the process of integrating those platforms into your security posture. So us being a connectivity provider for governments, for major corporations, enterprise carrier grade, we have a very, very strict and strong security policy, which means that when we bring new vendors on, it takes a lot of time and a lot of effort and a lot of back and forth.
And so what we found in certain cases, it was actually better for us and more beneficial for us to build the actual platforms that we needed for that particular use case. But like I said before, in other situations, we found that there was a partner who we had who already had AI integrated into their platform, and so they were already part of our security posture.
They were already inside of our ecosystem. So the question of build versus buy really had to do with time, had to do with return, and it had to do with the security measures that we had to put in place.
RR: Thinking about in addition to those factors, when you’re evaluating these things that you outlined, time, potential cost, security, how are you kind of doing that ROI math to say one is going to be better than the other?
CS: There’s several different parts of that. And like I mentioned, we wanted to make sure that we were following our AI strategy foundation that said, we don’t want to introduce more and more vulnerable access points. And so it’s important for us to make sure that we are all coming together with everyone across the Cox leadership team according to who are the vendors that we feel the safest with, that we can go set up and make sure that we are pulling together the best of the breeds.
The assessment, like I mentioned before, is what is the value that we’re returning to the business in terms of revenue generation, new customers, cost savings in terms of not necessarily just reducing people’s time, but redeploying people to doing other important tasks. And then what are the things that we are doing that help us to keep the system all working together?
So, revenue generation, cost deferment, and then keeping a cohesive connection between all of the different platforms. So some of the things that we looked at from our comparing vendors versus doing DIY, is there a maintenance tail that goes in this? So if we build it, what does that look like in 18 months? How much more people do we have to have to support it?
Governance and observability, do we have the permissions, the versioning, the audit trail, all of the parts for discovering what is needed and then able to see it and observe it as we go? Interoperability, as I mentioned before, really important between different platforms that we have, that those APIs and MCPs all work together.
And then whose roadmap is this? Is this our roadmap? Is this the IT roadmap? Is this the vendor’s roadmap? If we know where we need to go, is there anything that’s getting in our way of being able to get there? And then of course, obviously the speed to value against the cost of being wrong.
RR: And so hearing you outline this very comprehensive list of considerations, you can start to understand why it starts to feel complicated and really hard to tackle. To your point, it’s been a year of figuring it out since you started developing that very first use case.
I’d like to go into a little bit of detail about the evaluation piece and deciding what vendors you felt safe with, that you were excited to partner with and continue to either use or build upon as you’re developing your AI strategy in alignment with your business transformation.
One of those that you landed on was using Highspot’s MCP server to support some of the workflows you wanted to spin up. How did you make that decision and why did that feel like the way to go?
CS: Well, as I had mentioned before, as we had gone through our history of, here’s a library of a whole bunch of stuff and everybody’s trying to find the right item, and it just was such a headache to make sure that we were always getting the right information to the right customers at the right time.
And not only that, but we had no real clear feedback about how it was performing. And so at that time, which I believe was in the 2015 to 2017 timeframe, is when we had first started our relationship with Highspot to help us better catalog the library, make it more searchable and usable and referenceable for the sellers to be able to share information and track the information, make sure that it was the most relevant and recent, and then help us to understand what’s working and not working.
So all of that was already in place before we even started the AI conversation. And so as we were doing our work around our go-to-market roadmap, we started with content because it was probably the easiest place for us to use AI to generate content, and that looked like a two-layered approach.
We had what we called a knowledge base, which is formally putting into AI rules that can be read by AI around all of our standards for brand, for legal, for segment definition, for product information, for pricing and promotion information, industries, verticals. All of that was put at this knowledge base foundation layer.
And then we built the content generation engine on top of that, where each of the agents within that tool would go find what it is that the marketer was asking to do, compare it against all the information in the knowledge base, the brand standards, all of those good things, and then produce the content piece that the marketer was asking for using that foundation layer.
However, once we got that moving and going, we realized that that level of personalization for marketing could be even more valuable and even more specific when used by a seller. But in order for that to work properly, the seller had to have access to a large range of information all at the same time, including any of the buying signals or online signals that we had through some of our lead generation partners, any of our information that we have within our own systems, like when was the last time they called into billing or when was the last time that they had an outage or what is their general sentiment that the customer has right now.
And then all of the information about their current services, their current products, all the things that are going on in their world. But then once we have all of that information, we have propensity to buy, propensity to churn, propensity all these modeling, now we need to be able to talk to them and provide a recommendation to the seller that says, “Here’s what we recommend you use, what you should say, how you should set it up.”
And all of that was inside of Highspot. And so we realized again, we could look at this and say, “Are we going to go buy a new platform? Are we going to use a platform we already have or are we going to go build something new?” And obviously when we looked at the Highspot platform, the MCP servers, and the way that it was laid out and set up already, we knew that that was the right path to go.
So what we had started with was the content engine, then we went into a sales enablement engine, and as part of that sales enablement engine, the only way for it to work properly was for us to bring in the Highspot MCP service.
RR: And how has that been working so far for your sellers? As you’ve rolled this out, how has it been used? Any anecdotal feedback you’ve heard?
CS: It’s pretty funny because we have done multiple rollouts of sales enablement platforms over the years, and as anyone who’s ever tried to roll out new sales items and new sales tools will say, it takes time, it takes consistency, messaging over and over.
But in this particular case, when we went out and did our roadshow with all of the sellers and sat down and showed them how the new tool worked, there were so many positive responses, and the adoption was much faster than most of our previous releases of other types of products.
And I think that the reason why is because it was bringing together all of those pieces of information that I mentioned before and bringing in the Highspot information that they were already very familiar with. And in our world, we call it the sales asset manager, SAM. And so they were very familiar with SAM and then this new tool with the AI capabilities built into it.
Now it’s specifically just telling them, “Here’s what you should do. Here’s the way to lay it out, and here’s all the content to talk to the customer about in what order.” And it took a lot of the burden off of them to research, go find a piece, start to build a story in their head, try to build a deck, and then think about what are they going to share with them in what order.
So it’s been a huge benefit to the sellers. They’ve loved it.
RR: Yeah, that’s such a strong signal when adoption doesn’t feel like a push and more of a grab. Curious if there are any other AI or agentic connectors that you’re pairing with Highspot in another AI application that you think would be interesting to share?
CS: We have basically six different programs or parts of our roadmap, and we’re calling them AI modules, and then they work together in different components for different functions that need to be done.
So as I mentioned, we have the knowledge base that is the base. Then we have the content creation tool, which we call CAMI. So it’s Content Automation Marketing Intelligence, and that has everything that is needed to produce and create new pieces of content, and then those content pieces are either generated in emails or things like that. A lot of them actually are put into the Highspot tool.
And then we have what we call SAMI, which is the Sales Automation Marketing Intelligence, and that is the tool that integrates directly with Highspot to make the recommendations to the seller based off of all of the other information, the 360 view of the customer.
We also have what’s called Livia, which is the Lead Validation and Enrichment. The tool uses all of these multiple different access points and different vendors to pull information about that particular contact to validate that it’s accurate, so that by the time it gets to the seller and they’re going to go do a pitch, they have a lot more confidence that who they’re talking to, the business, and it’s at the right address, and prevents them from wasting time.
And then, of course, Highspot is such a critical part of how that story all comes together because it’s capturing all the content that’s being created by CAMI, and then the AI that comes from Highspot is infusing into the SAMI tool that the sellers are using.
It’s an interesting thing because somebody might say, “Well, you’re not really using Highspot, you’re using SAMI.” And the reality is, well, yes, I am using Highspot because Highspot is feeding all of that into the SAMI tool. There’s a whole bunch of other stuff we add into that for flavoring, all of the information about the customer so that the seller has a 360 view, but that just sets it up. The what do you do next is what’s coming out of Highspot.
The next phase of this that we’re going to is a fully agentic approach to our marketing and sales engine. And what that means is that today, most of the work that’s being done is a marketer who is saying, “Here’s what I need to go get done. I’m going to use AI to help me go do it.” We’re going to flip that script, and we’re going to say, the agents that we create are going to do the work, and the marketers are going to instruct the agents on how to do that work properly and watch it and govern it. That will then accelerate for the sellers as well.
RR: We’ve heard a little bit about what’s been built in the last year, but it’s, again, to your point, crazy that that’s one year of building, thinking, strategizing, and it’s come to this point. When you look across all of that, what has changed for your sellers and for the business?
CS: Well, I would say the first thing is, is that sellers are now able to focus on what they’re really good at. What I mean by that is their confidence is shifted to focus on outcomes and value. They are now able to build trust and provide value, which is honestly what all of our customers, especially our business owners and decision makers are looking for.
And then for the marketers, it’s no longer about building a queue, trying to figure out what is the message that’s going to hit the most people with the most response. This idea of efficiency for media or efficiency for marketing materials. It’s like, what is the one message I can send to a million people and have the most response? Well, now you actually can flip that on its ear and say, “I’m going to personalize it at scale.”
So that is super exciting. And then the last thing that I would say is that consistency became structural. The same knowledge base, the same rules across every surface, making sure that our content is clean, correct, built on the same policies and rules, but is personalized. Doing those two things at the same time is very tricky, and being able to do it with AI is the only way we could get there.
RR: Curious if you’ve seen any sort of measurable returns.
CS: Our lead accuracy, like I mentioned before, moving from that 13 to 18% all the way up to the 95th percentile. We have campaign speed to market of improvement of 55%, meaning the amount of time that it takes us to get to market is cut in half. The marketing content teams are 40% more productive, which means they’ve been able to redeploy their time for 40% of the time that they spend at work on other projects, which is amazing.
Our conversion rates are up, our driving net new revenue is up, and we have seen material improvement in click-through rates and conversion rates when we are more specific and personalized to the audience. So Gartner was right. Yay. So that’s been really good.
And I would say that part of the reason why I think that, at least for part of what we did, doing it as a build ourselves was wise, is because we learned so much by going through the process of just banging our shins on the corners and running into cabinet doors that were open, and we’re just like, “Oh, wow, that was, I did not see that.”
So it’s been a huge learning process, a very, very intense learning process, but we’ve all had a really good sense of humor and amusement and just, we are having a ton of fun.
RR: And I think that’s one of the more encouraging things to hear. Is that nobody starts perfect, and you just have to build your way up to good. And once you get there, you start to see again, like those measurable improvements. But it is a process. So I guess the message there is stick with it.
Which I think kind of feeds into that last question I have for you, which is for anybody who is running into this question, hitting their shins on all of these problems, how would you recommend they approach the question of building, buying, blending some things together when they’re thinking about their AI investments?
CS: Well, I would say the first thing is you have to look in the mirror and be real with yourself and say, “Is my processes and workflows working? If I blew up my entire go-to-market, I blew up all my processes, what would it look like?” And don’t start with a tool. Start from a place of what would serve me best.
The other part of it that I would say that Highspot did really well is because of the MCP product, I was able to look at it as how am I using this from a plumbing perspective, not just a judgment perspective. And what that means is that it worked well with the strategy and the AI strict rules that we had built for ourselves.
Highspot, kudos to Highspot, built a platform that is trusted and that works well with all of the other components that we had flying around, whether it was Salesforce or AWS or even our Accenture development team being able to use the components and pieces to connect to the whole ecosystem.
Then the other thing I would say is that even though we’ve been doing this for a year, a year is like eons in AI’s time. It was every other week there was something that changed, something new, something shifted. So you have to go into it with this idea of you have to prepare yourself that this is how I set it up now, but I might have to change it tomorrow, and just be okay with that.
So my answer for build or buy, my answer is both. Build the things that make sense for you and where you have the resources and when it’s the right fit. But definitely buy when you are in a partnership or when you have someone that you already know that you can trust.
RR: Very pragmatic. That’s kind of the only way to do it. One thing I’ll say, I know I am walking away inspired, and I can imagine our audience is going to as well.
So Cody, thank you for the time. I really, really appreciate it. It’s been so wonderful to hear a little bit more about what you’re building.
CS: No, I love it. And the reason why this is great for me is that it forces me to think back on this journey that we’ve been on for the last year and really consider what is it that has brought us to where we are, what are the things we’ve learned, and then, maybe how are my bruises doing?
RR: Well, thank you for the time again. And to our audience, thank you for listening to this episode of the Win/Win Podcast. Be sure to tune in next time for more insights on how you can maximize go-to-market success with Highspot.
