The Spotlight with GitLab's Ashley Kramer

Welcome and Introductions

Loreal Lynch: Hi, everyone. Thank you for joining us today for another episode of the Spotlight conversations with marketing leaders about AI. And I'm super excited to have Ashley Kramer joining us today. She's the Chief Strategy and Marketing Officer at GitLab. Welcome, Ashley.

Ashley Kramer: Hello. Thanks for having me today.

Loreal Lynch: Yeah, awesome.

GitLab’s AI Journey and Initial Prioritization

Loreal Lynch: Well, I wanted to kick it off, Ashley, with just kind of asking you a lot of what we have been encountering and talking about about at Jasper, particularly as you know, I talk to more and more CMOs and marketing leaders, is it's not really a question when it comes to AI about why you need to adopt AI. I don't need to convince people that they need to use AI. They already know and they're trying to figure out how to do it.

And so I would love for you to talk a little bit about your journey at GitLab. I know you've been at GitLab about two and a half years and I don't think when you started. Well, correct me if I'm wrong, but when you started, was AI one of your top priorities and how did it become something that you needed to develop a strategy for and kind of prioritize?

Ashley Kramer: Yeah, it's a great question. And I started at a really interesting time where no, the answer is it wasn't a top priority. I mean, I think we all know AI has been around for years and years and years, but definitely wasn't a top priority. Where it really started from the GitLab perspective is ChatGPT came out, which then followed with things like Copilots, things in the world that help we sell to developers, help developers write code.

And we had a moment at GitLab where we realized we had to make AI a top priority, a higher priority in the product. So we started in the product, which of course you all know what that means. Then marketing has to kick in and make sure everybody knows it's in the product. It's going to evolve in the product. Our messaging has to get right. We're no longer in comprehensive DevSecOps platform. We are an AI powered DevSecOps platform.

And so from that perspective, we were really maniacally focused on the product platform side of AI. And I would say we weren't thinking enough about how AI can impact marketing our business. So I would say a good six to 12 months after that was when we had that wake up call of, oh, we could be gaining massive efficiencies, competitive advantages, better targeting if we were to figure out how to use AI within marketing.

And so I would say we were in General behind it is very, very obvious to us now that we need to continue every day thinking about this.

Loreal Lynch: Yeah, it's interesting. I mean I think it's a good thing and a bad thing. Like the positive of that is that you were thinking of your customers first. Right. Instead of, instead of your own team.

So I think that's positive. But yeah, I think that is what we hear from a lot of people, particularly in the technology space is it's really like first step is how do we deliver AI to our customers. And then it's almost a secondary conversation of what do we do internally to harness the power of AI for ourselves.

Empowering the Team and Driving AI Adoption

Loreal Lynch: And so moving into, now that you're kind of embarking on that second part of the journey of how do you empower your team and drive marketing and business success with AI? Tell us about a little bit about how you're approaching that.

Ashley Kramer: Yeah, so we right now started approaching areas where it seemed to be obvious that we could use AI places like figuring out what competition is doing. So we have a tool that we brought in that's telling us when a competitive website changes, when new content comes out, what we should be doing. We are using Jasper, of course on the content side and the way we use it is to help us, you know, not start with a blank slate. What I love is it, it will intake our brand voice and tone and it will help us get to a spot where then we can get the human in the loop and then we can start diving in around optimizations and targeting around customer propensity to buy.

But to me those are all areas that seem obvious. I run sales development as well as part of our marketing organization and there's a lot of efficiencies that can be gained. Again, we don't want to totally remove humans out of the loop, but we can use more intelligent chatbots to get people further in the process before we get that human involved.

And so for to date it's been kind of a tops down me in fact one of the reasons I started looking at Jasper was I know your chief strategy officer really well and he said to me over coffee, have you, have you ever looked at what we do? And long story short, I hadn't. And now we are a customer and so I've challenged my team.

Now let's stop that. Let's actually step back and have all of you in your roles figure out where we can be using more AI to get more efficiency and further ahead in marketing. And so my, my VP of marketing operations and analytics is now actually writing a Full strategy doc for marketing.

And of course the whole team will contribute because I think it's time we stop it being a tops down. Okay. I went and talked to l' Oreal and she's using this or I went to a CMO roundtable and I want the team thinking about this and I've challenged them with that going through the rest of the year.

Loreal Lynch: I think that's great. I think so. That's interesting.

And you alluded earlier to maybe you're a little bit behind on adoption of AI or kind of prioritizing it because you started with the product. But in terms of what you just said, of how it has been more of a top down initiative that's actually a little different from what we've seen from other customers. And in a lot of cases what we're seeing is it's more bottom up. It's individuals that start adopting the tools first.

And so I actually think you're maybe a little bit ahead of the game in that sense. Is that one of the challenges that we actually see with customers is that because it's an individual first that's adopting it for their own kind of individual productivity, it's hard to think about how to scale that across the team and how to connect it to business results. And so like an individual, and this was in my last team, there might be an individual product marketer that would start playing around with AI because it helped them to summarize large documents into like a single slide.

So and that's like an individual use case. It's very hard to connect a use case like that to. We are using this at the top level and the CMO cares because we are driving more pipeline or we are, you know, driving more business impact.

So we think that there's a linear path from individual acceleration to team acceleration to business acceleration, but it's hard to get there without that top down approach. So I kind of think you need both like the bottom up and the top down, which is really interesting and kind of along those lines, my question for you is when you're thinking about kind of rolling this out to the team and you know, pushing them to think of use cases, think of adoption. How are you thinking about measurement and kind of like KPIs when it comes to like AI adoption?

Measuring Success and Efficiency with AI

Ashley Kramer: Yeah, for us it's, it's all around, you know, one of our core values at GitLab is efficiency. And, and so it's all about where are we gaining efficiencies. So again, I'm, I'm not of the mindset that AI is here to replace a lot of people in their jobs. There are certainly some organizations that it would impact more than others.

But when I think about marketing, it's. It's both art and science. You need to have the creative mind.

And sure, AI can help you not start with a blank slate and generate that. And then the science side, you can measure, you know, you can measure your ROI on. Okay, so we brought this in. Are we getting the blog written faster?

Because we now know more about our audience. Are. Are more people coming to it faster?

And so for me, it's all about am I gaining efficiencies? And depending on what we're using, that could be measured in human time spent, dollars spent, or the return on the investment in what we're doing.

Loreal Lynch: Yeah, absolutely. I think that's smart. And yeah, I will say I think that the customers that we see that are the most successful on their AI adoption journeys are leading with outcomes instead of leading with use cases.

But what are the outcomes that you're trying to drive and how do you back into the use cases to prioritize? But it's sometimes hard to think about that when there's such a rush to adopt. We just need to get some use cases stood up and shipped.

Ashley Kramer: I think most definitely on that one though. I mean, you probably get measured like me on what's your spend on headcount verse program. Right. And, and there's a way to think about it of when I go into next year, when we go into our next year planning, am I going to have to come in and ask for a ton more headcount or have I made the people on my team so efficient that we can do more with what we have put more into program?

And so I think there's definitely again, I love to keep saying to my team, I don't expect your jobs to be replaced, but maybe you're doing a lot more if you're learning how to leverage AI properly to your example earlier, then, then you're able to do more. And I want to give like one example of your bottoms up. Zoom started this new feature. I'm going to call it the wrong thing, but like you can turn AI on and it will capture notes from your meeting and give you awesome summary.

And so I get, we, we have assistants take notes throughout our entire conversation. Guess what? They don't have to do that anymore.

So they're even going to be more efficient, which is really cool. And it does capture. I don't know if you've been using it really funny conversations, by the way. Sometimes our banter at the beginning is captured and we're like, erase that from the, from this. Like, oh, they talked about their spring breaks like our kids on there. But, but yeah, I think that, that we, we.

Loreal Lynch: I'm glad that you mentioned that because that's another thing that I've been talking about is there's still, still is this question about like, is AI replacing our jobs and as marketers and I think in some, in some departments that might be more of like the goal is like you don't need to hire as many customer service agents if you can have AI bots and AI chat agents and things like that. In marketing, like you said, we are always under resourced. We never have the budget, we never have the heads that we need.

And so it's not like we're going to be like great AI. I'm going to cut a couple people like never. Like, absolutely not. Our headcount is very precious.

But I think it does become a force multiplier so that we can do more with the resources that we already have in place. And I think that's, that's a really good point.

Security, Privacy, and Legal Considerations in AI Adoption

Loreal Lynch: One of the next things I wanted to touch on with you specifically is the relationship between marketing and sort of like IT and legal when it comes to onboarding some of these new AI technologies and thinking about like how to evaluate them. Do you have any like thoughts on that?

Ashley Kramer: We have a deep process. So if you can imagine GitLab, at its core it's a platform, but it starts hosting customers code their private ip. So we have to be really, really careful with, with what we do and what we bring in.

And of course we have all of confidential customer information as well. And so from that perspective, you know, on the IT side, IT security side, they're involved in anything we ever bring in, AI or not, just to make sure is it secure? You know, can we break into IT and get data? The legal side is the closer partnership when it comes to AI right now.

And I think that's kind of just because where we are in the world and so you think about it from three components and what they want to look at, which is one, in general, most of the AI we're bringing in under is underpinned by a large language model. And every large language model provider has a different approach to how they have trained their models. So the first thing that we in partnership with our legal team looks at is what data was used to train that model. That's the first one. The second One is what are we going to do with the data? What are what. Whatever is generated via that large language model, what are we planning to do with it?

And the third is equally as important as the other ones is what's it going to do with our data. So when we go and feed in. You mentioned a slide earlier. There was a really very public use case that was kind of a disaster for Samsung about a year and a half ago where just an innocent person at the company, it became a public article, put in private company notes to build a deck and didn't realize he or she was leaking private company data to be trained on models and shared with everybody else.

And so those three components and our legal team has really great guidelines that they go through. We also have a transparency center for our customers so they can understand how we're leveraging AI and how their data is staying privacy first and, and transparent in what we do. And so that partnership is the closest when it comes to how we're going to decide to leverage AI throughout not just marketing, but the entire company and in our.

Loreal Lynch: So yeah, that's interesting and it makes a lot of sense practically. Do you have some sort of counsel or regular touch point on how to kind of drive these decisions and like, who would you say is like the decision maker when it comes to this?

Ashley Kramer: We have, we have a, I would say we, we call it dri, Directly responsible individual that, that sits in legal. But yes, there's a team that would get together and depending on, you know, so like when we were bringing, when we were bringing Jasper in as an example, it was my VP of Product marketing in partnership with VP of operations on our side working with it, security first as always, then working with legal. And so we're a pretty async culture, Slack culture here at GitLab.

So we have that quick sort of evaluation process I guess you would call it, to figure out what will work and what won't.

Loreal Lynch: Yeah, that's great. That makes a lot of sense.

GitLab’s AI Proposition and Customer Education

Loreal Lynch: So I want to talk a little bit about GitLab's AI proposition because I know GitLab focuses on using AI to help enterprise companies not just create better code, but kind of speaking of security, it's also safer code. So would love to just hear about how you're thinking AI can actually be used to help protect your data.

Ashley Kramer: Well, we all deeply understand the importance of code checks, quality and security after what happened a few Fridays ago where a lot of systems were taken down. And this is where GitLab shines is we are an End to end platform for software developers, security and operations professionals to build, secure and deploy their code. But what I think too many companies in our space have focused on was the actual writing of the code, the generation of the code. Help me write code faster. That doesn't mean it's more secure, that doesn't mean that it's better and higher quality.

So where we're really focused in our platform is sure we can help you write code, that's great. But what we really want to do is infuse AI through the rest of the software development life cycle. So the code has been written, suggest to me who should review it so I can get that done really quickly. All right, run security scans and explain any vulnerabilities that are found and help me resolve them all the way through deployment and measurement of once it's in production, is it still, you know, is it still of quality?

And so our unique, this is a great time for us. We're actually just about to launch the rest of the features and capabilities that I explained. We have a lot of them in there today. This is, this is huge because it really doesn't just help people deliver software faster, which is our company tagline. It, it helps people deliver secure software, quality software.

So less companies have, you know, issues happen after code deployment like we've seen recently.

Loreal Lynch: And I wonder if with kind of this new evolution and with AI being really kind of part of the product and a very important value proposition, the role of sort of marketing and the sales process changes. Because what I mentioned earlier where, you know, it's always been why AI, why Jasper, why now? And now it's not so much why AI, it's more of the how.

And so I, I feel like our role as marketers has changed to, to being a lot more focused on education and needing to teach customers and prospective customers on not just about, not not talking to them about our technology, but talking to them about like what they should do with it. And you know, why this is different. And so I wonder if you're finding the same thing where the role and kind of the content and the messaging that you're producing is less focused on the specific value proposition and more focused on educating the market on this very new technology.

Ashley Kramer: It's, it's. That's exactly right. We do a DevSecOps survey where we, we survey 5,000 DevSecOps professionals across the world. We just got our results. That's usually one of our top rated of course next to the Gartner MQs as I'm sure you're wildly aware. It's, it's one of the best assets that we have because it does exactly that. It educates, it educates people that while writing code in generating code using AI is important, that's only 25% of a developer's actual day. The rest of the 75% are all those other things. There was another interesting one and I won't, my team's gonna be mad at me that I haven't memorized the numbers yet.

But another interesting one that came out this year, you know, a large portion, I think it was about 3/4 of development teams say, yeah, I'm responsible, I'm a developer and I am now responsible for making sure my code is secure. Which wasn't a thing in the past. But then you ask security professionals and they're like, yeah, developers aren't securing code. Like only one fourth thought they were doing it.

So you're starting to see this mismatch and that is where you can go educate the people we sell to on the value and the power of AI, not the what it is. And that, that happened two years ago. I had to go around the world and say like this is what it is and this is how we're doing it now. It's why do you need it?

Loreal Lynch: Totally, yeah. That's really interesting.

Q&A: Intellectual Property and AI Onboarding

Loreal Lynch: A question just came in which I think is relevant to this. So on the topic of code writing and AI, as Ashley points out, it's a great assistant, writes it faster but doesn't make it more secure. Well, GitLab helps with this. How do you address the proprietary element of code that's been generated by AI and used in what is essentially intellectual property, like who owns it?

Ashley Kramer: I love it. I love this question. Thank you to whomever asked it.

So we very specifically only partner with large language model companies that have the same approach as us, which is privacy first and full transparency. So basically what we do, it's, it's right now it's anthropic quad in Google's, Google's vertex models. And the reason is, is because the way they've trained their code is on open source code. So, so they've trained code on what people have already shared with the world.

And then they do not take RIP and use it. They don't take the data, the code that we're putting in to retrain their models. That's not the case for all large language models out there.

And so we make sure we sell to highly regulated industries, fin serve government. It is a non starter for us to go and share code to back to the large language model. Now in the future there will be other ways to do things like have offline models, clone models where you can also integrate your code just within your walls. That's coming down the road.

But to date that is our promise to customers and we won't partner with any large language models that don't have that same approach.

Loreal Lynch: That's awesome. Yeah, great question. Someone else put a question here in the chat which is this is a product marketer that says they've just purchased Jasper and will be implementing it and would love to know tips from you, Ashley, on just what to think about in terms of successful onboarding and adoption.

So you mentioned earlier you have someone on your team that's writing sort of like a strategy for AI, but I think adoption is a challenge that we hear people talk about a lot. So do you have any thoughts on what has worked or hasn't worked within your team?

Ashley Kramer: Yeah, the way we spearheaded it is my vice president of product marketing championed it. So then that's that top down support. Some of our director of content will then sort of educate the team on the value it can bring and it sort of proves itself immediately. You know, you start using it and you see, oh, it got me this far.

Then you take. So I would say with anything that you bring in, don't boil the ocean and try to just throw it out there and use it for every use case. And so start with like I mentioned before, giving it your brand voice and tone ideas, put in your message house and then let it start to generate just something like a blog or something and see where you can get it, train it, train it as you train yourselves and then go from there. Another thing l' Oreal that you and I talked about when we, when we were actually just talking about Jasper in another forum is I think you're working on this.

But it's super important when it comes to AI in anything you do. Everybody's trained to work a certain way and the most successful tools in the world I think infuse it right in your workflow. L' Oreal and I both came from the same company and analytics back in the day and that was a big thing back then is I tried to tell people nobody's going to stop what they're doing and then go log into a dashboard and you need it right there in your workflow.

And so I think Gloria, you could speak to this better. There's a lot of ways in Jasper too, where you're going to put it right where they are. Right where they're working in the workflow.

Loreal Lynch: Yeah, absolutely. And I think you're right. When in BI and analytics, which is where Ashley and I both worked before, it was exactly like you say, someone, you know, the people that are living in a BI system every day are analysts.

But when you're trying to enable your frontline teams, marketing, sales, customer service with data and analytics, you can't ask them to go log into a separate dashboard. And that. And it's very much the same with AI.

And when we talk about AI for marketers, which is really where Jasper is focused on, marketers are working in a variety of different systems every day, depending on their department. So if you are in Lifecycle marketing, you're in HubSpot. So we can't expect someone to log out of HubSpot and go and log into Jasper and do their work in Jasper.

And so we think it's important to be able to have those AI powered workflows in the flow of where you work so it can be within HubSpot that you're able to have that, that kind of AI power with you. So that is absolutely like a part of our strategy. And again, what we're seeing when we see customers that are able to drive successful adoption and kind of along those lines. Another question that I have for you, Ashley, is like, this is more of like future state with respect to how the marketing team will be structured. I think because we're so nascent in terms of adoption, we haven't really nailed like, you know, with the HubSpot example there was a whole new function. Like there's like a HubSpot admin or a Marketo admin or like an operational person that is now managing that type of software and AI. Like what we're trying to do in Jasper is again, empower everybody to be able to use it.

But do you think that's realistic in the long term? Like do you think there will be specialized roles that will be created as like an AI ops person? Or do you think it's realistic to expect that everybody on the marketing team, you know, is, is adept and skilled at being able to use and leverage these tools themselves?

Future of Marketing Teams and AI Roles

Ashley Kramer: I think it'll be a journey. So I like to compare this. Back in the day I had a title at a company called Head of Cloud and there were Chief Cloud Officers and that was when the rise of cloud computing was happening. Those titles are kind of laughable. In fact, when I interviewed at GitLab my CEO laughed at me for that title. He's like, what is that?

And I said there was a day where nobody, everybody kept hearing it's important to move to the cloud. And nobody knew what it meant. So that was literally my job to go teach them what it meant, how to do it, the benefits, that it was secure. I think that's where we are with AI right now. You will see some companies with a chief AI officer. You will see some companies centralized like you said, like an AI, you know it or security or legal team. I think that's a short term title approach. Yeah. To really, really be successful. I don't want somebody in marketing with a title marketing AI or marketing AI operations. I want every single, not even one of my leaders, but every single person trying to figure out how they can leverage AI to become more efficient, to drive better business results, to serve customers better.

And I think if you silo that then it's, it's not going to have that. It's somebody else's job then. All right, it's not my job. I'll wait for the.

And the point of the AI strategy that we're writing is so everyone can contribute to it, generate thoughts and start to think how can this really change the mundane in my work life and the things that I shouldn't be spending my time on so I can use my brain power to do more creative and powerful things.

Loreal Lynch: Yeah, yeah, I think that's definitely. I share that perspective. I think and that's, that's the strategy that we're trying to enable as well is allowing every marketer to be able to use this and have it be as easy as possible. We don't want people to be prompt engineering in their day to day jobs. We don't want people to have to like think about this. We want this to be like we said, embedded where they work and just easy to use and to get, you know, the fast time to value I think is very, very important to us.

Q&A: Sustainability, ICP, and Closing

Loreal Lynch: Let's see, I have one more question that just came in and we'll, I think we'll end here because we're almost at time market, market analysts are split on whether or not the cost of energy consumption to run generative AI models versus the revenue results will burst the bubble and end in broad divestment. Are there hesitations on going all in on baking generative AI into workflows? Given the questions about long term sustainability of the technology?

Ashley Kramer: I think go ahead, you go for it. I think this again is going to be a journey. I had the privilege to speak at the United nations actually about a month ago.

And there was sitting next to me a chief, Chief AI Ethics officer. And she had a lot of thoughts on this. And so from the perspective of where we're going, I think there's, there's a lot of pieces that go into it. How, how do we approach first the cost, then the sustainability side. Sorry, that's hard to say.

And the, and how do we thoughtfully figure out how to infuse AI for good when necessary without causing huge economic impact? And so I don't have the. I'm certainly not a Chief AI Ethics officer. That role is real and there are a lot of people carefully and thoughtfully thinking about it.

And so I, I think it's going to be a journey to see where we end up. Laureate, I don't know if you have. Sorry, I think I spoke over you a little bit there, if you have any thoughts on this.

Loreal Lynch: No, I think, yeah, because it's talking about the consumption versus the revenue results. And I think, yeah, I think right now we're still like, we're still so early and I think there's not a lot of, there's not a lot of proof yet on where kind of where this is going and how it will net out. So I think, I think we need to just continue to work with customers and build those models basically. Like what is, what does success look like and what results should we be looking at in terms of what is, what are the trade offs like? I almost feel like it's too early to see, but yeah, I agree with you.

And okay, another totally separate question just came in and I don't know if we have, I think we have time maybe for one more, which is. And I don't know if you can speak to this. It would probably be a better question for your VP of Product marketing. How much is AI helping to define the ICP targeting? Are you guys using it in this way at all?

Ashley Kramer: We are using it for some Persona work. So we, you know, we are very close to our customers, which is the best way to start to figure that out. But we are using it for some Persona and icp. I would say refining. Not. I don't. I think your product marketing team is not close enough to customers if they needed to actually create the icp.

But to refine and then to figure out how to better target and customer propensity to buy type models, those are the places where we can extend off of it. But I would say yes to refine.

Loreal Lynch: Awesome. Great. Well, great questions I know we're at time. Thank you so much, Ashley. I've appreciated this conversation and your insights. If you have more questions, feel free to drop them in or, or email directly and we can, we can definitely get back to you.