The Build–Buy Line for Agentic GTM — Fireside Chat with Kady Srinivasan & Jake Reni

Julia Nimchinski:
Thanks, and welcome to the show! Katie Srinvasan, CMO of You.com, World Class CMO, and Jake Rennie. Our good, old community favorites.
Jake Reni:
Not, not battle.
Julia Nimchinski:
Not at all. Co-founder and chief growth officer at Revenue Reimagined. Stoked to have you here. How have you been, and what’s your top GTM AI projection for 2026?
Jake Reni:
Caddy, you wanna go first?
Kady Srinivasan:
What… what prediction? AI prediction, or just in general?
Julia Nimchinski:
GTM and AI.
Kady Srinivasan:
My prediction is that… We will, we will see a few different kind of agents working, but we are still going to be in the process of figuring out how a majority of agents work.
Julia Nimchinski:
Jake, how about yourself?
Jake Reni:
Oh my gosh, I feel like my prediction for this changes on a weekly basis right now. Because of the pace in which things are moving so quickly. But if I went at a macro level. I would say my prediction is going to be, like, I don’t want to call it a reset. Because I don’t think this will be a reset, but I think that there will be… there will be a great understanding a better understanding of what has worked and what we are learning together as a community that’s learning at the same time, especially in go-to-market, what hasn’t been working, and I think we’re going to see a big step back on a lot of the aggressive initiatives around this to actually be more thoughtful, more intentional, and more strategic around the approaches to implementing AI in organizations, whether you’re driving efficiency, revenue, or cost savings.
Julia Nimchinski:
Very exciting times. Well… This stage is yours. Take it away, Jake!
Jake Reni:
Excellent. Well, thank you for having us. as you’ve already done the introductions, you know, I’d love to just go ahead and get started into the conversation, Caddy. And, you know, some of the things you and I have been talking about, I think, are super interesting. One thing I’d just love to just start as, like, a setting the stage here. Most people, obviously, who are going to be here listening and attending this session, are likely already, you know, bought into the concept of agentic AI. But let’s, let’s start kind of earlier on in this thought of, one thing I often find that isn’t clear, there’s not clear alignment on them in organizations, is who actually owns and should be leading AI initiatives within a company. What are your thoughts there?
Kady Srinivasan:
Yeah, thanks for having me here. By the way, Julie, I just wanted to mention that, and Jake, it’s super exciting to talk to you about it. I think we are going to have so much to discuss that we’ll probably run out of time. But, I mean, it’s a really astute observation that, I think the… the buyer of AI products and services right now is going… is so fragmented. You depend… it depends on… really on, what… when you think about the go-to-market tech stack, there are all of these individual users that are using tools, like, let’s say, the Gamas of the world, or the Descripts of the world, or the Heijians of the world. And there, what I’m seeing is a lot of the individual users, the marketers, the content marketers, the designers are using those tools. But then when it starts to try… become an enterprise kind of a motion, typically the CMO does get involved, but then, more interestingly enough, the CIO gets involved, because there’s a governance aspect of it, there’s a… what kind of data you’re uploading to all these systems, particularly ChatGPT, you know, kind of a thing. So, it’s so important, if you’re on the other side, you’re trying to sell an AI go-to-market tool to an organization, I think you need to really map out who are all your stakeholders, and who is going to benefit the most from this, and what are the objections that they’re going to have. And that said, let’s bring it back to this particular conversation. It’s about the CMO is going to be super interested in driving efficiency and growth from lead… a lead perspective. A CRO, obviously, is going to be interested in how do I drive more pipeline, close rates, conversion, and you, you probably can speak to it more than I can. So it’s, it’s, yeah, it, I think it’s, it’s a fascinating time to be in the AI go-to-market space.
Jake Reni:
You bring up an interesting point, and it’s not something we talked about, but I do… I do want to get your thoughts on this. You mentioned governance, and when you say governance, people tend to freak out, right? It’s a scary thought.
Kady Srinivasan:
Yep.
Jake Reni:
And it kind of feels like, oh, well, that’s where things can then go to die, but how does an organization start with AI initiatives while incorporating governance from the beginning? Where should they start with this in mind?
Kady Srinivasan:
Yeah, I think it’s, you know, my… what I’ve seen in terms of successful AI, let’s call it deployments of tools is when the CIO is actually bought in right from the beginning, and then there is a very clear understanding of the transformation that particular tool can have over the rest of the organization. So we’ve brought in, you know, a couple of things, like LevinX, for instance. And, it’s… it’s so important that our CIO knows the kind of data we are uploading, the kind of information we are giving, and how that’s going to be sitting in a sandbox with a tool like that, and what we are trying to do with it. And then we, you know, kind of… you have to ensure that you’re also giving proper status updates to make sure that the stuff that you said was going to hopefully happen is going to happen. Like, is pipeline going up because we are using this tool, or should we be looking at something else, type of a thing? That’s, I think, that’s the best way to bring it together. I think that in the future, the CMO and the CIO are going to have to work super closely. Because there is so much, like, you know, porousness of the data in an organization. In order to make any agent work effectively, you have to provide it with a ton of context. And that ton of context is all the… The company private data, and the… but that company private data, before you go and upload it to a… an LLM, you have to make sure it’s cleared by a CIO who is much more, who’s more sophisticated than you are about governance and security and zero data retention and those kinds of policies.

Jake Reni:
That’s a good point. You know, something we talk to our clients often about is… is… You can start as early as having the conversation around just a readiness assessment in the organization. Are we ready to implement AI? What are the things we need to be thinking about from a governance, right? And so on from each department. And then from there, putting together an AI charter. And really getting clear about where we’re going with this. To start having those conversations you’re talking about, so that there is an element of readiness before we get ourselves into trouble, right? Now, let’s get to the conversation topic. The session description talks about GTM growing through threads, or multi-threads, founder-led signal mining, event nurture, composite loop versus isolated tactics. And I noticed this morning, actually today, you dropped an article, on Top Line about the rise of multitheater Marketer. What’s that about? Tell us more about that, like, and what led to you wanting to write that?
Kady Srinivasan:
Yeah, so let me, you know, take a step back for a second and just outline the context of what that is all about. So, you know, before I joined u.com, I was running a… I was a CMO of a large public company. I had a bunch of people, 180 people working under me. when I came here to u.com, I realized you… the number of people that I had working on individual things I don’t know if we need the same number of people, however, there’s different kinds of things that need to be done, so you need a significant shift in the way that you’re deploying talent. One of the things that I’ve noticed is growth has become extremely complicated, like, because There’s so much AI slop out there, so much content out there, so many tools out there. It’s extremely fragmented. So, in this world, the way that we used to go to market, like, you know, you create a white paper, you throw it out there on your blog, you hope that people come, you do some amount of SEO, then you do some paid and hope people come. That sort of linear process doesn’t work anymore. That’s the playbook.
Jake Reni:
That’s what everyone’s doing.
Kady Srinivasan:
Yeah, and that’s a SaaS playbook, and you literally have to, like, tear it up, throw it out, and really think about what is the new playbook. And the new playbook is… it’s not these linear set of activities, it’s what I call the flywheel. So how do you create a flywheel that, you know, compounds? And I’ll give you an example of what that is. The second thing is, what I also realized is, in doing that flywheel, you have to have a multi-threaded approach. So, you can’t just say, I’m going to just do this one thing, and then this other thing, and then this other thing. You have to say, I’m going to create a content that’s going to immediately go on the blog and the website, and also going to our founder social. and going to an email that I’m sending out to people, I’m going to look at all the responses. So you can imagine, a person who is doing that job now is a content marketer, a website designer, an email lifecycle marketer, all rolled into one.
Jake Reni:
Is it one person?
Kady Srinivasan:
It’s usually one person, or a person who’s, like, drawing on different skill sets, using agents, a person plus a bunch of agents, you know, so… it’s a multi-threaded approach to building a flywheel. So now you have a flywheel, and you have a multi-threaded approach to, like, the components of the flywheel, and that’s what this article talks about, is why that paradigm is going to be the future. And, I outline some specific examples of what has happened, and the thing… the reason I’m about to say why this works is we have a lot… I have a lot of confidence, because we 10x’d our MQLs over the course of a quarter by doing that, and our ACV went up, like, 86%. So I have a decent amount of confidence that this is the future of where this company… where the marketing org is going. So that’s the basis of the article. I also think it has big implications to how do marketers, and even to some extent, salespeople, think about their skills in the world of AI? You know, what’s going to keep them relevant, employed, you know, how do they need to up-train themselves, all that kind of stuff.

Jake Reni:
Okay, so you mentioned seeing 10x, results from campaigns. What… what were those… like, what were the architecture decisions that were… enabled those numbers? Were they build use cases, buy use cases, hybrid approaches? What were the applications there?
Kady Srinivasan:
it’s definitely a hybrid, so I… one specific example I’ll give you is, so we have this loop where, you know, our founder goes and posts some things on social media. In the past, we would just let it go, right? Now, what we have added on is, imagine, founder posts. Then we look at all the people who have engaged with that post, we scrape that list, we filter it for ICP, we lead enhance it, and lead-nurture it. Send it to the sales team. And then of the people that are qualified, we invite them to certain events, whether that’s founder-led or something else. That generates, in turn, more topics and thought leadership that then we use to do more founder posting.
Jake Reni:
Interesting, so you’re actually doing a value-add play here, and not immediately trying to sell them on a meeting, or sell them on a discovery call with a sales rep, but you’re using that to actually find higher intent where you can add more value.
Kady Srinivasan:
That’s it, yeah. Interesting. And the hypothesis there is, with a lot of these AI tools, or at least with what we sell, we sell problems and use cases, not just solutions. So we sell them… it’s the art of the possible. We sell them what they can be doing, so that takes a couple of beats, right? Couple of additional steps. So, that was one loop. As you can see, that’s a flywheel. Like, once you get that going, it just becomes a reinforcing, compounding type of a flywheel. And in doing that stuff, you can also see it’s not just one social media manager who’s going and posting, it brings together multiple disciplines, social media, field marketing, content generation. lead nurture, mo-ops, you know, that kind of stuff. So, that’s what I mean by the flywheel. And so, to your point about build versus buy. there is, I’ve actually hired, what I call prompt marketers, who… whose job is to build agents for us. And so we use a combination of our own build and, buying, buying, or, using from the outside, using, you know, like, workflows like Make and N8N and some of those tools.
Jake Reni:
It’s an interesting topic, you know, I’ve been in a lot of discussions recently where the idea even is going as far as saying, you know, we should be thinking about hiring cloud agents or AI employees as much as we think about deploying, you know, hiring a standard employee, or a human employee, right? But thinking about it from the perspective of that you should have individuals actually working specifically on you know, specific agent… agentic applications that are specific for tasks, as opposed to broad tasks, but getting very, very linear on the specialization of that agent, and then having an individual manage that. Would you agree with that approach? What would you say to that?
Kady Srinivasan:
Yeah, a thousand percent, because I think, again, look at how we need to make agents work. Agents are not going to be plug-and-play, like SaaS was. Agents need to be prompted with the right, not only the prompt and instructions, but also the right context. which is basically all your private RAG data, right? And that changes all the time, right? Like, in an organization, especially if you’re growing, the private repository of stuff changes. So you need a person who is basically managing, maintaining. The context of the agent to make it more effective. That’s one part of it. The second part of it is, again, you can’t just… I mean, we’ve tried this before, where we have AI SDRs and we let them loose. We don’t see much impact. We see some small amount of impact. then what we have to do is, like, really think about, okay, why is that not working? What can we go and do differently? So we’re… that ca… taking that feedback and creating that feedback loop, that today, no agent or no workflow, no company does that. So you have to do that manually.
Jake Reni:
Yeah.
Kady Srinivasan:
You have to hire people who can go in, look at the data, look at the thing, and then come up with another hypothesis of what might work, and then go change prompt instructions, change the data to make the agent work. And so, all of that stuff is… I just don’t see yet anything on the market that does everything autonomously like that.
Jake Reni:
Yep, yep. I totally agree. We do need the humans, and I think this is super interesting that you bring up this point, because, You know, you talk about this process, and one of the things we’ve seen in our business, a lot of clients wanting to deploy AI in some way, shape, or form or another, because they’re either getting pressure from their board, they’re getting pressure from their investors, they’re being told, what is your AI initiative or your plans around it, but they don’t know How to respond, and the question… honestly, the people asking the questions don’t even know what the question means. But they’re getting pressure, and so they have to do this, and so what we see often is deployments that are very brittle, and they break down because of lack of whether it’s this, you know, the rich context you’re talking about, or they’re trying to have these agents work too many steps in the process, and so managing that process, you know, eventually breaks down. What’s interesting to me is I’ve seen more failed attempts than I have seen successful, but I… when they’re successful, they tend to have high impact. I’m curious, from your… from your experience, where, you know, where have you seen these end-to-end agents have the most impact in deployments that you’ve done?
Kady Srinivasan:
I think it’s, for now, a lot in content. It really unlocks massive growth, in content if you can do it in the right way. So, I… we… the way that we use, agents to unlock all of this is, like, we really… we have a sort of a supply chain of content creation, if you will, where we come up with ideas, we then use AI to build the, basic structure, we use AI to check for accuracy and validate stuff, and then we close it out with a human just kind of editing and writing in a human voice, and then we upload it. What I would have… what I would love to do as a next step is then bring in a… an agent that analyzes how we are doing against all.
Jake Reni:
qualifying.
Kady Srinivasan:
put out there, and then generate ideas back to what we should be creating. That loop hasn’t closed yet, so… but that’s a big portion. Me, personally, as a CMO, I used to spend so much time on making decks, because that’s part of the whole thing, right? Like, you have to get.
Jake Reni:
We love it. Yeah, we love making decks, right?
Kady Srinivasan:
Oh, God.
Jake Reni:
It’s spreadsheets.
Kady Srinivasan:
And I started in the Microsoft PowerPoint days, where you could literally, like, change things manually. Now I use a combination of an LLM and Gamma.
Jake Reni:
Love gamma.
Kady Srinivasan:
Love it. It’s like 5 minutes, and I have a real deck that I can stand behind.
Jake Reni:
So it’s… so that’s… I don’t know if we would call gamma an agent, but it is a definite…
Kady Srinivasan:
massive improvement in, how we… how I communicate internally with my stakeholders.
Jake Reni:
Oh, it’s incredible. I mean, even as a quick side note, I saw the power of it is so strong that… I was telling you, I have a 17-year-old, and he saw me using it, and he said, Dad, can I use this for my big presentation for one of his AP classes? And I said, look, here’s the deal. If you show me you know how to take manual notes, and you bring home detailed notes from your lectures, I will show you how to turn those detailed notes into an actual presentation on Gamma. So. I’m not gonna let him skip the work. Yes. But I’m gonna show them how to do it.
Kady Srinivasan:
Yeah.
Jake Reni:
like, blew his mind, and we gotta be teaching this generation how to do it, so that’s.
Kady Srinivasan:
That’s… I love it, and I love the fact that you’re focusing on getting him to really understand the concept and take the notes, because that. Obviously, that is the key. Otherwise, you get… create slop.
Jake Reni:
Exactly, exactly. I know we don’t have a lot of time left with you, so I’m curious, just as we kind of go back to the build versus buy topic today, For GTM specifically, do you have an opinion on whether it’s better to always build or buy, or do a hybrid of both? It should be thinking about building internally so we have a defensible advantage? Is there an advantage to buying? What are your thoughts?
Kady Srinivasan:
I think it is a hybrid for now. I think if you are really short on time, and you want to get to market super quickly, I think it makes a ton of sense to buy and customize, even though you’re going to have to spend, I think, a lot of cycles in customizing, because even things like Gamma, I love it so much, but I don’t think I can use it in an enterprise context just yet, because they have some ways to go. You know, so there’s… Buy if you don’t have time, but my… my suggestion is build, because every company is going to be so unique in its operating context, and the way that I would want to create workflows internally is going to be so different from some other company. So, if I can hire, like I said, like, you know, I think I mentioned this, this prompt marketer that I’ve hired, if I can bring someone like her on board and have her let loose on these flywheels that I’m creating. That becomes my competitive advantage. Right? Because nobody else is going to replicate it quite to the same degree. So… And then that… you can imagine that you do more and more of that, and you become… you have more and more richer, more sophisticated flywheels that really then become sticky, within the organization. It’s very hard to do. First, you have to find the right people, you gotta train them up in the right way, you have to have the right systems, you have to have the right data, so it’s a lot of work in the beginning, but I think the That’ll pay off in the long run. So I.
Jake Reni:
Agree.
Kady Srinivasan:
That’s the… yeah, that’s my hypothesis. Now, there are a lot of companies who are trying to make this easier. They have forward-deployed engineers who can come and do this for you. If that’s an option, sure, those are expensive. So you have to be a bit careful about what you.
Jake Reni:
But it’s a good… it’s a good option of… taking the build step, if you want to, without… Owning all the risk internally, when you might not be set up properly, or have the expertise in-house to do it, and it’s a good stepping stone to building out that capability. Interesting. Okay, so we only have a few minutes left here, so let’s do rapid-fire last couple questions and see if Julia has any questions from the audience. But, okay, one, what’s one of the, like, what’s… The one mistake you see companies make repeatedly when implementing Agentic and GTM.
Kady Srinivasan:
the biggest mistake is what I call this context engineering mistake. And just so… I should have explained this a little more. What I mean by context engineering is, agents need not only the instructions, but also all of the data and the context.
Jake Reni:
behind it.
Kady Srinivasan:
how they need to operate. And when you don’t give enough thought to getting the right context into the agent, then it almost always yields an output that is not quite what you wanted. So you spend a lot of time just tweaking it. So that’s the one thing that I see happen over and over, is They… the way to take context and scalably introduce that to agents is missing, and people don’t spend enough time and effort doing that. So let’s make…
Jake Reni:
that the second question, because I was going to ask you what’s one step, or one, you know, thing that people can do tomorrow, but let’s go to that. If that’s the biggest mistake you’re seeing, then, you know, for those watching, not wanting to make the same mistake, what’s the first step they should take to do this properly with context when building an initiative internally?
Kady Srinivasan:
So, I would say the first one, and thank you for… yeah, I think you pushed on the right things. The first one is even figuring out where you want to introduce agents. Because, again, like, it’s not like a real light bulb that you just put in somewhere and switch on, right? Like, you really have to be very deliberate about, do I want to take away this particular workflow and introduce an agent here? Is that what is going to accelerate my thing, or do I need to think about something differently? So there’s… in my… the article that you kindly talked about, I have this… this… this thing of pick the places that you want to introduce what I call threads, like, pick the threads, because that becomes super important. If you’re going to spend the next 3 months of your life Making an agent work. Then pick the place where you think it’s going to work.
Jake Reni:
Totally agree, and I think that’s why earlier we talked about not just an AI readiness assessment, but then what is your company’s AI and, like, what was the word I used? You know, your, your… your charter. Start there. That forces this conversation. In fact, we have a free one that we can share if anybody wants that. Julia, I’m happy to share that with you if anyone asks, but that is definitely somewhere where people need to start. Alright. Julia, I don’t know if you have any questions from the audience you’d like to share with Caddy or myself, but… I’d love to hear from you.
Julia Nimchinski:
Sure, thank you for an amazing session. Yes, we do have a couple of questions. One of them is about the hidden maintenance cost of in-house.
Jake Reni:
That’s good.
Kady Srinivasan:
That’s good, yeah. Yeah, I mean, like, I would say it’s not even hidden, it’s quite out there in the open. Not just the agents that you build in-house, but also the ones that you bring on board or buy. You’re going to have to maintain them in some way, shape, or form, because, you know, the context changes of the company, the data changes, the… all the stuff that you thought might work doesn’t… won’t work. Right off the bat. I think it comes down to… again, like, do you have the right people who are managing the humans and the agents, and can you put the right firepower behind it? Do you have the right DNA of skill sets to do that? What have you seen, Jake?
Jake Reni:
I would agree, actually, that’s where I would start, is you’re almost building in technical debt if you don’t have the expertise in-house. Because then someone has to maintain. And if you’re not planning to, or willing to, or have the capability to, that’s a big problem. That also introduced a lot of security risk as well. I think we’re gonna see a lot of people dealing with just data security issues as they start trying to implement here, so I think those are some of the built-in, like. Risks, or what is… what was the word, like… like, efficiency issues or something like that, but that’s the challenges I’m seeing.
Kady Srinivasan:
Yep.
Jake Reni:
Yeah.
Julia Nimchinski:
Totally agree. One more question, where do vendors outperform internal build? And what are your thoughts,
Kady Srinivasan:
Yeah.
Julia Nimchinski:
since 2026.
Kady Srinivasan:
I would say the… some of the places in go-to-market that I’ve seen where vendors outperform is, things like inbound phone… or, like, conversational AI agents, anything related to voice agents, they’re… those, it seems to me, I’m not… no longer an engineer, but it seems to me those are really hard to build. And so where there are vendors who are already doing it, they definitely have much more of an advantage, and so it’s better to buy those. And I’m not talking about the CS, customer success agents, I’m talking about for go-to-market, like inbound, SDR, voice agents type of a thing. That’s definitely one thing. Content is so easy to build in-house. Anything… I kind of separate the world of agents into two, right? One is the, sort of, like, the rules-based agents, and one is the action agents. The action agents, where an agent has to take an action, like send an email, or do X, Y, and Z, it’s… Sometimes easier to look at a vendor, because they’ve spent a lot of time engineering that.
Jake Reni:
Yeah, I would agree with you. Yeah, I mean… specifically for, like, very deep, vertical, specialized applications is really… I think vendors outperform their… Katie mentioned this earlier, but, like, context is the most critical part to this being successful, and, like, the instruction layer, and, like, the rag layer. If you’ve got a vendor dedicated to you know, giving that context or building that context into, you know, their agent, I promise you, they’re gonna have more time, resources, and funding to put into that than you are internally. So, like, learn on their dime, probably not on yours.
Kady Srinivasan:
Great point.
Julia Nimchinski:
We have to know, what’s in your Gentec GTM stack? And what do you plan to try in 2026?
Kady Srinivasan:
Gosh, I have… I want to try so many things, but right now, I’d say the… the ones that I really like are, 11X, Connect the Dots, Sumble, Relevance AI is another one where we use that to build certain agents, so those are some top of the mind for me.
Julia Nimchinski:
just had their CEO and another funnel.
Kady Srinivasan:
Oh, amazing.
Julia Nimchinski:
Okay, that’s.
Kady Srinivasan:
Great, that’s great.
Julia Nimchinski:
How about yourself, Jake?
Jake Reni:
Just too many. I don’t even want to look at my credit card statement right now, because, like, we just play with everything. So, I mean, from… from each LLM, I think we pay for all of them, to, you know, to Lovable, to Cursor, to, like… like, I would say more… my answer probably would really just be, like, if you’re not dabbling or playing or learning, you’re missing out, because you’re just gonna be a step behind. So, as opposed to me telling you to try something specific other than gamma, that we talked about earlier, like, just get your hands dirty, do it now.
Kady Srinivasan:
Yeah. You need an agent to manage your credit card bills.
Jake Reni:
I know, there’s no room for that.
Julia Nimchinski:
Do you think agents will be buying from agents in 2026?
Jake Reni:
Jeez, not… I don’t want to say no, but I… would love to see the cluster that turns into, so let’s hope for it, right? Got it.
Kady Srinivasan:
Yeah, I agree with Jake. I don’t want to come out and say impossible, but it’s also… it feels right now to be so far away from reality that I’m wondering if one year is going to be enough to bridge that.
Julia Nimchinski:
Thank you again for the fantastic session. Super excited, always, to see you, feature you, and promote your work. Catty?
Kady Srinivasan:
Jake.
Julia Nimchinski:
how can we support you? Do you have a substack? LinkedIn? Where do our people go?
Kady Srinivasan:
I do have a LinkedIn and substack. Please, look for my LinkedIn. My substack is called Multithreaded Marketer, for those who are thinking about the… My whole philosophy is I want to help marketers get ready for the next generation of AI, so that has a lot of detail about what kind of jobs, skills, etc, so please check that out and let me know what you think.
Julia Nimchinski:
Thank you again.
Jake Reni:
And I’m on LinkedIn. You can check out our website at revenue-reimagine.com. And Julie, if it comes up, I’m also happy to share that readiness assessment or that AI charter template with the community, so I’m happy to share that out with others, too.
Julia Nimchinski:
Definitely. Thank you so much again, and that’s a wrap for Day 3 of the Agentech Singularity Summit.
Kady Srinivasan:
Thank you for having us.
Julia Nimchinski:
Thank you, Catty.
Kady Srinivasan:
Bye. Thank you, bye.
Julia Nimchinski:
Thank you to all speakers, thank you for all of you watching, thank you for the community and partners. We will be back in January, January 15, I believe, and February. So mark your calendar, lots of exciting events in 2026, and Happy holidays to you and yours!