The Agentic Stack — 2026 Predictions for the GTM Technology Landscape

Julia Nimchinski:
Thank you so much for the phenomenal fireside chat. Thank you, Eric, Kimberly, Angela. And our next panel is hosted by Michelle Buckley, VP of AI and GTM at Gartner. Welcome to the show, Michelle. What a treat!
Michele Buckley:
Thank you, Julia. I’m happy to be here. I hope you don’t mind me using my own background, but we’ve got an amazing panel coming up. I think we’ve got a quorum, maybe waiting for one more person, but shall I kick off with an introduction and hellos?
Julia Nimchinski:
For sure. One question before we do that. What’s your top AI and GTM prediction for 2026?
Michele Buckley:
My top prediction is that Agentic AI in go-to-market is mandatory and necessary. People should have started piloting in this year, and it will be… Affecting all roles in sales, marketing, and service in 2026, but a lot of people don’t know how to actually onboard agents, and that’s where the twist is going to be in the people and the process around the agents.
Julia Nimchinski:
Love it. The stage is yours.
Michele Buckley:
Thank you for asking me, because I’m going to be asking everyone the questions for the rest of this session, so… I think I will, jump into it. I see we’ve got A few folks here, so welcome, everyone, to the Agentic Stack 2026 Predictions for the go-to-market technology landscape. And I’m here as a monitor. I am a vice… a moderator, not monitor, call monitor. I am a vice president Analyst at Gartner, focused on AI go-to-market in my research, talking to hundreds of tech companies every year, so this is my jam. Yay. We have a wonderful depth and diversity of panelists that underscores how the go-to-market tech stack affects sales, marketing, and service. So, I will briefly introduce our esteemed panelists. I invite the audience members to look them up on LinkedIn to learn more about them, and I invite the panelists to, wave or do, like, a power pose, when I call out your name. So, we have with us today Brent Kremes, Chief Customer Officer at Gainsight, a customer… a leader in customer success platforms and customer-facing AI, and he’s on mute, so I hope it doesn’t stay that way for long. It won’t. Good to see you.
Brent Krempges:
Thank you.
Michele Buckley:
Okay, thanks. Oh, great, we’ve got Nathan Gotch also, who is the CEO of RankAbility, and he has deep, deep expertise in SEO, content systems, performance-driven growth. We also have with us, also on mute, not for long, Dr. Tuba Durace, CEO of Amoeba AI, a fast-moving innovator in agentic automation, enterprise-scale intelligence. And then, another esteemed colleague, Werner Schmidt, CEO of Lat Teve, or Late Teve. Focused on connecting sales strategy and plans to real-time execution, which I’m a fan as a former sales leader myself. And then, We, last but not least, of course, we have Santi, who’s there, standing up quite formally. He’s the CEO of Momentum. I am another fangirl of that platform, among the others. A platform transforming deal collaboration, sales workflows, and revenue execution. So, this is a big group, there’s 5 of us, and you’re all from different areas, but I think that’s what’s going to make this discussion really interesting, because we’re going to combine different perspectives during this panel. As a moderator, I’m here to pass the talking stick amongst you, and create an organic, informative, and fun session. So, to do that, you know, I’ve taken to heart one of my favorite sayings, which is, a fast game is a good game. So, I’ve created a plan. There’s gonna be 5 sections in this hour that we have ahead of us, short and long format types of things. So, we’re gonna start off with Section 1, the crystal ball, where you have… I give you an opportunity to give a quick opening prediction, then we’ll go into Part 2, more deep discussion of shifts in the HEPIC stack for 2026. Third part is where I put you in the hot seat. I’ll let you worry about that. In the meantime, we’re not working, followed by rapid fire, and then we’ll have a closing. So… Let’s have some fun, okay? So, better than a magic 8-ball, when I call out your name. please go for it and give me your headline prediction for a major change or shift in the go-to-market technology landscape in 2026. It may be in your area of expertise, it may not. This is what we create as we go along. I want to say, as… obviously, if someone’s talking, there may be someone on the panel who’s like, yeah, I agree, yes, thumbs up, double thumbs up. thumbs down is okay, too, or, hmm, not sure. You decide, but this should be interactive. So, we’re gonna start with you, Brent. You are on the main stage. What is your opening prediction for 2026?
Brent Krempges:
No thumbs up to the rest of the panelists coming out of the gates. But, so I’ll say, I think one is we’re gonna really shift from AI and agents informing workflows and truly running them. So that’s one… I’m going to kind of do two, though. I think the other big one that we’re already seeing start is more around value realization is going to be key, both from AI functionality as well as existing functionality. I think with that. outcome-based pricing is going to start to be much more of the norm. I think a lot are talking about it. Not a lot are executing there, so I think those are some of the big… transformations that I think are going to be more of an evolution than a big bang, in 2026.
Michele Buckley:
Hmm, so we certainly had a big bang of 2025, didn’t we? So, yeah, evolution is welcome by my perspective. Over to you next, Nathan.
Nathan Gotch:
Alright, yeah, so, you know, obviously for me, SEO has gone through quite the, transformative process here over the last couple of years, but the biggest prediction definitely for me is the full transition to answer engine optimization. I think that is definitely… what is going to be fully rolled out is going to be fully adopted.
Santi:
Shane.
Nathan Gotch:
2026. This is not to say that SEO doesn’t have utility. SEO is kind of that underlying you know, foundation for AEO, but Google has already made it quite clear that, AI mode is gonna be the… the, default in traditional search, so that’s gonna be… require a lot of adjustment for a lot of brands, and also, as we probably will talk about here, attribution is not gonna get any easier. So, that’s gonna be happening. Yep.
Michele Buckley:
A lot of people are gonna get headaches about that, but, thank you for that. Let’s move over to Dr. Tuba next, please.
Tooba Durraze:
Yeah, I have two. One might be a little bit more, kind of, far-reaching, but my first one is… Dashboards will actually die, and we’ll move to more, kind of, living computation organisms versus static things. spreadsheets, all of that, just the consumption of data will change. And my second one is, and I hope this happens, is… Folks are going to stop using the term AI synonymous with like, LLMs, because there are lots of different kinds of AI, lots of different kinds of agentic systems that are built on system… underlying technology that’s not just LLMs or transformers. So you’ll see a shift towards other kinds of AI, including my favorite one, I’m biased, neurosymbolic AI.
Michele Buckley:
I’m sorry, what was that last part?
Tooba Durraze:
Neurosymbolic AI?
Michele Buckley:
Oh, neurosymbolic AI. I’m learning so much already, and I thought I knew a lot. Okay, great, thank you. Alright, who are you moving on to next? Werner.
Werner Schmidt:
So, I think, I mean, a lot of change, like you said, Big Bang in 2025. For me, it’s… I think there’s still some ways to go through identifying the workflows of where that automation is going to happen, and where, you know, what are the outcomes going to be from that? Next is, of course, then deploying that agent, but I think then it’s measuring it, right? What’s the productivity that’s been driven from the agents versus what’s happening from the humans, and how’s that going to start to… Materialize into, you know, the outcomes that the company’s gonna be driving from a revenue standpoint to a margin standpoint. So I think that’s still… that’s got some time to work through. I don’t think that’s just going to suddenly change overnight. Too many pieces in these organizations, in most organizations, to truly understand what the impacts are. And I think, you know, to that, I think there’s going to be AI that’s come in that they realize, actually, this isn’t working, we need to switch it off and rethink things. Because I think a lot’s been put in very fast, and not really understanding what the impact is, not just on the organization, but the customers, and how that then actually translates into business at the end of the day.
Michele Buckley:
Hmm… I’m with you on that, thank you, and the… you know, I’m hearing in your comments, and thank you, everyone, I’m taking notes, and I think some of these themes will resurface as we progress through our time together. There is the technology, but Werner, as you mentioned, there’s also the people and the process around it, and So… let us move into our second section, which is for a little more discussion, a little more organic discussion. You don’t have to comment on every topic, but I’ve got four topics for here for us to discuss. In terms of the shifts that we’re going to see in agric technology in 2026. Let’s start at the very high level, like, and this is an open question to whoever wants to jump in.
Werner Schmidt:
Where do you see the, the…
Michele Buckley:
mindset that go-to-market leaders really should understand or take on board? What are some of the assumptions or ways that they’re looking at AI that they’re getting wrong already? Technology aside, what is the right mindset that You would recommend go-to-market leaders have as they move into 2026.
Santi:
I’ll jump right in. I skipped my prediction, but I’ll… don’t worry.
Michele Buckley:
Sorry, Santi, please do both! Oh, I’m terrible.
Santi:
To bring it back into the fold, actually, my… my mindset is very much informed by the prediction. The prediction was very much in line with Warner’s. I think, the… this is the year that we start organizing the Big Bang. Last year was a lot about scrambling and trying to throw things at the wall to see what sticks, what delivers value. And I think 2026 is the year we organize. And through that organization, you’ll start seeing a real chasm between the winners and the losers of this disruption. And I think the mindset is what’s going to separate one from the other. I think, now transition into the second question, what’s the real mindset? I think the winners are the ones that are going to see, this transition into leveraging AI. As a foundational re-architecture of their business, of their team, and of their tools. And the losers are the ones that are just gonna keep buying tools and agents without a real proper cultural change throughout the organization. That’s the core mindset. You gotta stop thinking about, oh, there’s cool tech, and I’m just gonna buy some of that and put it in my business, and it’s gonna deliver a real delta. You gotta start realizing, no, this is… pretty disruptive. The level of depth of alpha, of impact and disruption it can deliver is large enough to justify rethinking a lot of the building blocks of my organization, of my business, of my go-to-market.
Michele Buckley:
Yeah, I would agree with that. And that is really what’s behind that word, transformation. A lot of people, Santi, don’t… appreciate as much, and we’re starting to get immune to it, but the idea is, you know, if you’ve got a building, you know, you’ve got to shed out the back. This is not about, like, adding improvements to it. We’ve got to tear it down and rebuild it from scratch to really get the best outcome. And the like. Would anyone else like to comment on the mindset shifts you see?
Tooba Durraze:
Yeah, I’ll add a thought here. I think a lot of what I see happening right now is, we’re just trying to jump to efficiency, so everyone is jumping to, okay, what is the thing that’s going to solve my problem faster, essentially. There’s a lot of, like, automation and workflow optimization related to that. I think it would be really interesting for me to see how go-to-market leaders, like, take a complete step back and think about, are we even solving the right problems? And then, frankly, getting out of the way is another one, which is, you don’t really… get on a plane and ask the pilot, okay, how does the autopilot work? Or, like, I must understand every intricacy about how autopilot works, right? After a certain point in time, there’s enough trust that autopilots work. So I think the more, kind of, there’s a shift towards You know, let the machines compute, as long as there’s a very clear business definition and business goal, and you can kind of get out of the way of that as a go-to-market leader, the better your outcomes will be, versus trying to introduce a ton of points of intervention.
Michele Buckley:
How do you get out of the way? When you’ve got to rebuild things. So to speak, or transform it.
Tooba Durraze:
Yeah, I think the analogy I use is, right now, you’re in a… Tesla, there’s a self-driving module, but you’re at the wheel, right? Because you want to know, like, I want to be able to kind of step in it, because I feel like circumstances aren’t safe, or the software is not great. And then we want to be moving towards a Waymo, where you’re still, like, defining the end goal. Where are you trying to go for your business? The questions of that might be as simple as, we need our business to grow by 30-40%, as an example. And then you let the system then figure out what is the best way to get you there, within your parameters. I want a scenic route, or I want the fastest route, as an example. So to me, getting out of the way means, like, you’re in the vehicle, but you’re letting the system kind of drive. And then you’re in observance of telling the system where to go.
Brent Krempges:
And I’ll just add. Both of these… I think both of these points are directly correlated to each other, and that so many right now are… trying to think of the quick analogy, but they’re buying the plane, but they don’t even know where they’re going yet, in terms of, like… but they’re assuming they’re going to get wherever they don’t know where they’re going really fast, and then all of a sudden they’re disappointed, because they didn’t get wherever they didn’t know where they were going fast enough. And the reality is, the only way you can get out of the way is to have the foundation in place first, which is, what are we trying to solve through an efficient path? And until you have that in place, like, you can’t get out of the way, and I think that’s where, in 2025, so many have overspent. and expected unrealistic expectations without even really understanding, like, what is the foundation that we’re gonna build this thing on. So, I think that’s really where organizations have struggled without, like, you know, it’s the old. slow down to speed up, and most have just tried to speed up even faster this year, because organizations are forcing them, if you’re not AI first, you’re going to fall behind, but there’s still foundational business components that need to be in place first, before you truly can go fast and get out of the way, so…
Michele Buckley:
That makes sense to me. I was gonna say, I can…
Nathan Gotch:
to that, because I have kind of a unique perspective here, because I work with agencies, and so every single day they’re talking to business owners, small business owners specifically. And there is this idea for some small business owners that ChatGPT is going to solve their problems. And the reality is that, you know, once… and this is… I’m much more tactical in my thinking here, because I actually, you know, use these systems extensively every day. But the more you use these AI systems, the more you realize that you need a human. And so… and the deeper you go, the more you need that human in the loop. And so, like, a small example is, like, you know, our company uses Replit extensively, and if you’ve used Replit or any vibe coding tool.
Santi:
It’s amazing the first time you use it. You’re like, this is unbelievable!
Nathan Gotch:
And, like, you build an app in, you know, 5 minutes, and it’s just the most mind-blowing thing. But then, when you start to really get into it, you’re like, oh, okay, well, there’s a bug there, and then a bug there, and then wait a second, I don’t even know how to code, so how am I gonna fix that? And then… and then you get into this weird environment, so there’s… so what happens is, like, what we’re realizing is, we’re able to automate so much of our processes from an SEO perspective and so many other things using AI and agents. But we always go back to having a subject matter expert lead those initiatives. And that’s the key that’s been really effective for us, is to have someone who actually knows what.
Santi:
what they’re doing.
Nathan Gotch:
controlling these robots. And that’s really what’s been the big difference, and, you know, trying not to have someone who’s a jack-of-all-trades try to figure all this out. So having a developer come in and actually them fix the bugs in Replit. Have them actually come in and do that, because As you guys know, there’s different skill sets, right? Like, someone who’s good at coding is maybe not very good at design, so, you know, having those different kind of, strengths, and kind of doubling down on that seems to be… has been very, very effective for us. So, yeah.
Tooba Durraze:
Let me… let me challenge that, because I think people use the term human in the loop, I… in a very expansive way, I would say right now. There’s a difference in, kind of, human-in-the-loop.
Santi:
Machine-wise, which is your…
Tooba Durraze:
have introduced points of intervention to allow the machines to get better, and then there’s the other side, which is, I think, what you’re alluding to, Nathan, which is always having a sense of observability or trust for verification on the system by subject matter experts to course-correct the system. Which, that piece of it, the trust by Verify, like, obviously needs to exist. I think your interaction with agents, systems, etc. would turn more into a sense of, like, you’re looking at the cockpit, or you’re observing, kind of, what’s happening, but not that things are, like, out of your control. But I do think that there is a world in which you will get in the way of the system, and we know that because we look at folks designing 30, 40-page prompts. to get a single thing out of the system. And these systems operate a lot better, were designed initially to be, like, a completely pompless architecture. So, I think there’ll be, like, some… maybe you’re right, maybe 2026 isn’t the right year for that, but I think there’s going to be some way where your expertise as a go-to-market professional or person who’s kind of running the business is in the strategy of the business and, like, where the business should go. and what things should happen, but you as a human, sitting on top of all that data, may be introducing the wrong points of intervention if you’re trying to kind of be in the loop on everything, in that sense. So I would say that it’s an uncomfortable situation for folks, because I think part of our safety is in control in that way, but I would say the systems operate a lot better if there’s A layer of governance, but not to the point where, you know, the system can’t learn from itself.
Santi:
I don’t think you two… I don’t think you two have to be necessarily at odds here. I think both could be right at the same time, and what makes a real… The vision here is what exact workflow, or what workload is being executed, and the expectation of quality and importance of the outcome. There’s plenty of workflows that can be run fully automatic, unsupervised today, and it’s perfectly fine, while others are perhaps a little more critical and require a little more supervision and a little more intervention. And I think it goes back to this idea that the right businesses will be the ones that have the Aptitude to delineate that, to put it all on the table, and to architect it. properly. So I think you both are right. And of course, in the future, as we go further and further into the evolution of this technology, hopefully we’re going to be moving more and more of these workflows from, like, still need a human to maybe don’t need a human anymore. But you gotta stay up to date and keep an eye on the tech as you assess that.
Werner Schmidt:
And I think to add to that, Santi.
Michele Buckley:
Warner, and then I’ll move on to the next topic. Go ahead, Warner.
Werner Schmidt:
Oh, perfectly, and it’s true, I think, I mean, if you look at it now, right, it’s identifying that workflow automation. And if we think about it, generally that’s been done by junior roles, where we can bring in more of this automation through these agents. But deploying the agents, one piece, but I fully agree. The quality of what comes out… I mean, we all use AI agents, I’m sure, in our calls, right? When you see what comes out, there’s no ways you can send that straight to a customer, and that is not at a level where it’s going to get better. I don’t think it is. There’s just this human piece that needs to be added, because otherwise you’re going to notice that that’s not a human that’s just sent me an email of someone reading it, or there was a mistake because the word was spelt Think of English between US English, UK English. You’re dealing with global companies. The AI agents are going to take a lot of time before they get to that point. So, yes, will they get there? Sure. Is that going to happen next year? I don’t think that’s going to happen. But, I mean, I view this a bit differently, right? Once you’ve got the workflows and you understand it, it’s important within organizations to understand what’s going to be done by the human. And what’s going to be done by the agent? And that’s… I fully agree with Santi on this, and that’s where companies are going to win and not win. But at the end of the day, I mean, the vision of this, where it goes right, there’s agents, there’s a marketplace, you go get the agent that you need to do what you need within your organization. And that’s how… what does that tech stack look like that you build up? So, lots of… I think we’re gonna all be on a journey on this, lots to, think through, of how it all comes together.

Michele Buckley:
Yeah, agreed, and I… let’s double-click on that great conversation that we’re having and get specific in terms of go-to-market stack. across sales, marketing, customer success, and service. And moved to a, getting specific on what type of workflows we think in 2026 can be fully delegated to Ageptic AI. To build upon what we were saying earlier, you mentioned these subject matter experts, so critically important. I think the big shift, the mindset shift for everybody is that That subject matter expert isn’t necessarily the programmer. it’s the business owner. I’ve had IT managers come to me going, oh, Michelle, tell me about this AI SDR. I just want to help the sales team. I’m like. I should not be talking to you right now. I should be talking to the sales leader. These ideas have to start in the business, because that subject matter expert is… there’s one on the technology, but there’s also one on the business outcomes as well, so… Let me turn it back to each of you, and since I forgot Santi, I’m gonna ask him to go first this time, and hopefully he’ll still forgive me. What go-to-market workflows do you see agents fully owning by the end of 2026. Doesn’t have to be a massive one, but, like, what kind of tasks or workflows today are you seeing? Yep, it’s ready, it’s doable, the ROI is there?
Santi:
Yeah, it’s a… it’s not a hot and sexy one, it’s not AI SDRs or AI sales reps. I think it is more practical, and… less exciting, and to me, it is the gathering of data from every single channel that sells and post-sales to engage your market with. Not because it is the sexiest, or perhaps the one that has the most time savings, but because it’s the most foundational. I think, ultimately. proper AI agents and workflows will benefit the most from having a swath of high-level, high-quality data available to them, so you have to start with extraction of data. And to have high-quality, reliable, high-quality extraction of data, you want to automate it. And AI happens to be a pretty good technology for that. So I would say note-taking, recording, email analysis, support ticket analysis, it should all be the workflow of 2026. Some leaders out there have been talking about foundation of first-party and third-party data. We all know about ZoomInfo and how you obtain third-party through play and the like. I think first-party is going to be the topic of next year.
Michele Buckley:
Interesting. Okay, I’ll go in reverse order last time. Werner, what do you… what’s your perspective on that, or whatever other workflow would be automated?
Werner Schmidt:
So, I think when you look at any workflow within a sales process, right from that lead of what needs to be done the way through, I think for me, I think what’ll happen is conversational intelligence. With AI on it, we’ll get a bit smarter on how it can populate parts of that Salesforce journey, or journey of the opportunity, where information can be populated in a more organized and structured way. So back to that point, Asanti’s right, you need the data, the data needs to be there. Without the data, I always think of AI, it’s garbage in, garbage out, right? Sitting AI on bad data is not going to give you anything, unless it’s right. So there needs to be a big effort in that, and that’s where AI absolutely can certainly help with that completion, which then ultimately will make things better. So I think that’s where, for me, is the main changes will be… well, I should say, really, the main effects we’ll see of this. I don’t think we’re at the point where AI is suddenly going to replace salespeople, or, I mean, SDRs to a point. Yes, the technology’s there, but I’ll talk to that a bit later on my view on it.
Michele Buckley:
Sure, just, just an example, lots of use cases out there. I suppose, to refine the question a little bit. I’m… I’m… I’m… I feel this nuance of there’s things that should be done, things that can be done, and things that will be done. And I love your comments around data and populating CRM. And I appreciate that is… important work, and it can be done, but… I don’t know, it’s not… you said, it’s not sexy, it’s kind of boring, it’s hard to get, like, a business case to pay for data quality. It has been for a while. So, I don’t know, I’m gonna throw it over to you, Brent, because you’ve been politely listening. What’s your view on what Let me clarify, what workflows will agents fully own by the end of 2026? Like, what is, like, yeah, this is happening, and if you’re not doing it, you gotta get on it.
Brent Krempges:
Yeah, I mean, so I’m… I’ve been thinking about it more as, like, tasks, not jobs. So I think there’s gonna be different internal tasks, and I think most of those tasks are what you foundationally already know what those are. So one… one, just to give specifics, I think when you think about, like, a renewal process, most organizations know, based off of the different segments, what are the certain things that should happen along a renewal process. So that should be, what is our first touchpoint? How do we engage with our CPQ in order to create a quote? etc. And I think most of these, you know, so in those types of examples, I think it’s… 2026 is going to be more all the internal things where you’re trying to drive scalability, and you already know what should be happening, but it’s challenging to scale. What I don’t see… It may be late in a year from now, but throughout the majority of 2026, I think most organizations are still hesitant to have agents actually interacting directly with their customers. And so the first phase is, where do you have all these internal workflows where you already know what you need to do, and you just need to put something on top of it?
Nathan Gotch:
to drive that scalability. Obviously, that data’s critical.
Brent Krempges:
But I think this first phase is gonna be… Where we know we have imperfect data, but we know we have process in place, where do we still have some level of human interaction to ensure that if things do go off the rails a little bit. it’s easily manageable, and it’s not completely disruptive to what our customers are seeing and hearing from an agent versus a human. And I think once you have that foundation in place, and organizationally you have the confidence of this is working. then you can start to open up, okay, now this agent can start to now directly interact with a customer, but other than support agents, where it’s more reactive, I think there’s still… like, I know personally, I’m still hesitant to have something proactively reaching out to my customers until we get some of these internal foundations in place first.
Michele Buckley:
I would agree. I speak to that point. I did mention AI SDRs. I do feel like we’re… not ready for that, per se. I am seeing some great use case, though, around pricing and quoting and order forms. So, salespeople, like, the customer wants to sign, I need a contract. I talked to one company, they were doing it… took them 5 days to get it around, turned around with different calculations of multiple currencies across different geos and the like, and they programmed an agent in a month. Now it doesn’t 15 minutes. And so, 15 minutes after going to a customer meeting, they’re like, oh, you asked about pricing. Here it is, just… Click if you’d like.
Brent Krempges:
The last thing I’ll say, Michelle, is I think SDR is less risky to most organizations. Obviously, I sit from within the CCO office, so I’m thinking about it more from a customer perspective. I think that’s where most are gonna that’s what’s gonna come last, is actually, like, the engagement at a customer level. There’s less risk if you screw up a prospect, like, every sales leader may roll their eyes when I say this, but, like, that’s less risky. We all know that it’s much more expensive to… create a new customer than maintain the current one, so I think there’s less risk appetite right now for how you’re engaging with your customers than your prospects, so I think We will foresee more of a transition for how you’re driving more scalability into the top of the funnel. And then everything in terms of the renewal and adoption agents are what’s going to follow.
Michele Buckley:
Can you give an example of, like, a renewal task you see being automated internally?
Brent Krempges:
I think just the initial reach out. Like, what is your overall engagement in terms of sentiment? Like, usually you’ll have, let’s say, 120 days before the renewal, you do some level of engagement or reach out just to understand likelihood to renew.
Werner Schmidt:
And I mean, today, we already have some of that built within our own platform in terms of, like, automated emails, etc.
Brent Krempges:
Then you can take that a step further, where you actually have an agent that knows how to respond based off of how the email or phone call comes in. That could actually progress stages within your CRM. And then as a follow-up step, if you need to negotiate, then you can start to do some of your, again, within your CPQ, start to put the quote together. I think most mature organizations have some foundation in terms of how many days out from a renewal, do we start the engagement, what are certain, you know, discounts that we can put in place based off of the spin threshold? I think those are the types of things you could start to put in place now with agents, and then that starts to allow a renewal manager, or whatever that function is, to be, you know, let’s say, you know, later stage engagement within the renewal process, versus… You know, first step all the way through completion of the signature.
Michele Buckley:
Okay, good. So we talked about workflows. Perhaps we should, as part of this discussion, before we move on to some of the hot seat, huh, and the rapid fire. What pitfalls are going to… what is going to break in 2026. as companies are trying to adopt more autonomous agents. What do you see as being a common failure, or the place they face friction. Now, I totally hear you in your earlier comments, like, okay, they don’t have enough data, or they don’t have the right data, or they’ve got to prep their data, or… I think that is coming… More and more clear. What else do you see maybe breaking.
Tooba Durraze:
Add something there?
Michele Buckley:
Thanks!
Tooba Durraze:
We do not need perfect data to get started on AI systems. Like, it’s a controversial take, but I think that this idea of… this happened to… a lot of BI solutions, where it was, like, this idea of, like, I’m gonna bring in a new BI solution and have to kind of redo all of my data, is not necessarily true. You don’t need a ton of data, you have… you have AI techniques that allow you to simulate data, you don’t need the cleanest data, you just need auditability. If the data doesn’t exist, then that’s a whole other story, but just when folks think about that, it’s like, don’t think about data in terms of something that needs to be perfect for you to get started. Like, the systems can help.
Michele Buckley:
Fair enough.
Tooba Durraze:
Get better.
Santi:
Don’t worry.
Tooba Durraze:
that you would need.
Santi:
You need good data, but not perfect data.
Tooba Durraze:
Yeah, yeah, and let the system then… agents… a great use case for agents is auditability on your data, and data readiness, right, as an example. So I think… I think I always say that because folks have this, like, fear a long go-off in their heads when they’re like, oh, I don’t even have the right data, or I don’t even know if I have the right data in that sense.
Michele Buckley:
Fair call. So that’s a… that’s a good… let me turn that around a little bit to reframe my question, which is, you know, what… What do you see could be causing alarms? in 2026. Privacy? Yeah.
Werner Schmidt:
Yeah, I think… Well, I mean, that’s always going to be one, but I think organizations, assuming these huge efficiency gains, putting big numbers out there, and then it translates into efficiency gains, to drive the, you know, these big increases that are expected in productivity that AI is going to drive, because it takes time. Again, matching that… the human effort to what the AI’s doing on the process. And, you know, there’s… so I think there’s just a realization what is the productivity gains that we’re actually seeing from AI before putting some big numbers out there that organizations are going to tie themselves to. So, and I think… for me, that’s going to be one of the biggest pitfalls, and the companies that have been in the house on AI to suddenly hit, you know, 50-60% growth, or whatever it is, you know, just hold back a bit, and… and see this play out first, because there’s lots of… there’s lots of moving pieces. Will it get there? Yes. It just… I think it’s going to take a lot longer than you think.

Michele Buckley:
Alright, so I’m hearing…
Santi:
fan of the…
Michele Buckley:
that it’s, it’s more about heartbreak than systems… heartbreak of not meeting expectations. You know, Gartner does say on the hype cycle, we’re about to go into… well, actually, we are in the trough of disillusionment for… for AI and the like, and when everyone’s really like, oh, this isn’t as great. as we thought. Let me turn over to you, Nathan.
Nathan Gotch:
and how else?
Michele Buckley:
You’re gonna, yeah, you see stuff breaking, or… You know, tell us your perspective.
Nathan Gotch:
Yeah, well, I mean, it… I think it’s obviously industry-dependent, because, you know, everyone’s a little bit different, but at least in the industry that I work in, what I live and breathe in the search space, there’s already disruption happening every single day. And I’ll give you a small example. So, for many, many years, when we would run agency campaigns for clients, you know, whatever size business, SMBs, bigger businesses. You would need to build out a full team to run that campaign, because SEO touches so many different teams to properly execute. But I’ve been running an example over the last year or so where we decided to move, basically, our entire website all to Replit. And this is an extensive process, and the reason we were doing this is because we wanted to prove that we could actually run all technical SEO all on-site SEO, directly just using agents. And so what’s happened now is pretty unbelievable, because we basically have eliminated full technical SEO teams. We don’t even need them anymore, because the agents, because of the systems we’ve built within Replit, now can actually go and do very basic things, like Add internal links, improve site architecture, improve loading speed, even as far as making the site achieve accessibility standards, having it run compliance standards on the website, you can do all of that with the agents already right now. And of course, as was mentioned earlier, you know. does require trust that that is, working and is correct, and is actually meeting those compliance standards, but you can funnel all of those compliance standards into the agent and feed that knowledge base to actually get within a pretty high level of accuracy, so… The reason I’m bringing that up is because, like, recently in the US, there’s been… You probably don’t hear a lot about it, but there’s a lot of businesses that are getting sued because they don’t have proper accessibility on their websites. So there’s lawyers that will literally go and audit a bunch of small businesses, and if they don’t have compliance, they send them a notice and say, hey, you’re about to get sued. And so they’ll find someone with some sort of accessibility issue. And then… they will send these notices out. So, if you can get within a certain level of compliance, because they’re only going after the lowest hanging fruits, of course, the sites that really have a lot of accessibility issues, if you can get up to 90%, you’ve just eliminated some potential headaches in the business that you don’t, you know, no one wants to deal with. And so this stuff is already happening, and you can look across even the SEO process, like. you know, maybe my SEO friends are gonna like me saying this stuff, but, there’s content teams that are getting crushed right now. Getting crushed. I mean, AI has just disrupted that so much, from deep research to content production, even writing YouTube scripts has become something that’s almost fully automated if you do it the right way.
Michele Buckley:
I could go down the list, and this is, like… Yeah, but same with translation, you know, transcription.
Nathan Gotch:
Right, like, this isn’t, this isn’t…
Michele Buckley:
disruptions.
Nathan Gotch:
But, like, yeah, it’s not theoretical, like, this is happening every single day, and there’s people that are pushing the boundaries, better, for better or for worse, as far as automation, but, like, on YouTube, there’s some people that don’t even know that some of the videos are actually AI. So, it’s already getting to that point. So yeah, just my, just my opinion.
Michele Buckley:
Well, it’s… sounds like you’re saying that… there are… some things that have already… are already broken, and have been for a while, and you haven’t noticed, like accessibility, like compliance, and AI is now enabling Things to be found faster. And reveal more of those cracks that were already there more quickly, and also fix them more quickly as well.
Nathan Gotch:
And it also used to cost a significant amount to actually hire a team to do that for you, and, you know, I did it in Replit for 25 cents.
Michele Buckley:
Yes.
Nathan Gotch:
So when you’re thinking about the profitability and the speed of some of these tools, it becomes very hard to justify a human in some cases. And…
Santi:
The math just doesn’t work.
Nathan Gotch:
And that’s… that’s what’s happening, at least in my world, in my world, specifically.
Michele Buckley:
Hmm, I would say in a lot of worlds, the most… basic administrative, just like when we, 10, 20 years ago, we went through outsourcing, you get this done cheaper, in a different country, now it can be done cheaper from AI. So, so that’s exciting. I’d like to conclude the discussion part now, because, to move on to our other sections, as we have 20 minutes left in our session, if that’s alright. And I’m going to… Look back to some of the predictions you came, and put you in the… Hot seat. I know, it’s a weird thing. Where’s it gonna come from? Maybe it makes you, like, sweat a little bit, but I’m gonna go around each person and ask you an individual question that… for you to answer, to provide, your insights. Alone, and let’s say, you know, in just 2 or 3 sentences. And you’re in the hot seat because you kind of may or may not know what I’m going to ask you. So, okay, I’m gonna… I’m gonna go to you first, Brent. You mentioned in your opening call, you… you… I took my notes correctly, your prediction is about outcome-based pricing. What would you say… who is most nervous about outcome-based pricing? And who shouldn’t be?
Brent Krempges:
Hmm… You know, actually, the irony of this, I think AI companies that are positioning these crazy ROI, but then they’re not actually selling based off of those outcomes, are probably who are going to be the most nervous. I mean, being the customer success company, I predict that we’re going to get a lot more AI companies, because they’re going to realize they have a retention problem. Pretty quickly. I think those that aren’t nervous, or shouldn’t be worried, are those that Already have, like, value realization as a core part of how they interact, engage, and sell their product.
Michele Buckley:
Interesting. Okay, great. Thanks. I’ll go to you next, Tuba. What is the… You mentioned earlier in your prediction, dashboards will die. I think a lot of us would be very excited for that to happen. You know, how is that actually going to happen?
Tooba Durraze:
I think when we stop looking at AI and assigning it values like AI is our teammate, or, like, our analyst, or anything of that sort, and we start looking at the capabilities of AI as, like, a living organism, you realize the job of this living organism is to point out to you what’s important and when and what to do against it, right? Those are kind of the three values. As a person who’s running a business. Again, fundamentally, there’s only one question. What’s going to make my business grow? Or what puts my business growth at risk? And the answers to that cannot live in a static dashboard, because you’re then biased. You’re pointing your question towards a specific set of data and saying, these are the things I should abide by in terms of how I’m governing my business. So I think it changes with Folks realizing how much you’re leaving on the table without, like, outside of the dashboard, and realizing that those things have a huge impact on your business as well. Like, finding, like, millions of dollars in anomalies that you wouldn’t see if you were just looking at specific dashboard, and then eventually graduating towards like I said, like, AI, the brain of your business. And then, you know, if it’s the brain of your business, its job is to reduce your cognitive load, so not you always seeking, but then it pushing to you and telling you, hey, you need to pay attention to this. this is what’s happening, or this might put your business at risk. So, I don’t know if we’re psychologically there, I’ll say that, honestly, because people are very attached to visuals and dashboards in that sense, but I think More and more, as people realize, there are things hidden in your data that you wouldn’t have been able to find because you didn’t ask the right question, or you didn’t point it at the right… like, set of data, the more people will realize, you know, I have all this opportunity to make my business better.
Michele Buckley:
I think the company ramp is a great example of that, the fast growth they’ve been seeing as well, seeing things that couldn’t be seen before. So, thank you for that hot seat answer, Tuba. I’m gonna go to Werner next. You were talking earlier in your predictions about How measurement and attribution is going to change. what measurement or KPIs or vanity metrics in 2026 do you think we really need to stop? Using or pretending to believe.
Werner Schmidt:
I think it’s… there’s, you know, with AI, certainly, I mean, if you think about the AI insights. you know, it is super powerful of what it can discover very fast. I mean, if we think about it now, you know, technology and how AI can run the sales plan, and identify things that you would have never have thought that are there because of the models that exist to be able to compare it against, and the speed at which it can do that, certainly gives you a whole new perspective of how you can look at the business. So, those vanity metrics that are helping guide you to certain, or got you a certain place, are being replaced to some degree of far more meaningful metrics that are actually going to tell you if you’re going to hit the number or not. So, but that takes time. I mean, I think all metrics are needed at some point for some reason, but I do think that the level of insight that you can now gather from these models of what the AI can do is really elevating the discussions of what… what happens of, to then help set the priorities of what you’re going to focus on to get to, to get to those numbers. So I… I think the, you know, the standard way we think about the go-to-market function as a whole And the metrics we’re using outside the main SaaS ones, of course, that we know and love, are going to start to shift and change to more metrics that are going to tell you if you’re going to hit the number or not.
Michele Buckley:
Okay. Let’s put you in the hot seat, Santi. Agentic AI. Buy or build? We built agents out there, there’s a lot of agent platforms out there, and obviously you are very close to this space, and trust in your thoughts.
Santi:
I think it’s a mix of both. Unfortunately, the answer is complicated. I will say, you should decide. You know, it’s always been a tough choice you had to make between buy and build, and you always had the choice to do both, and… usually one of the two was always the right choice, depending on the use case. I would say the easy things that are somewhat trivial and pretty well-built already by a vendor that are not keeping you up at night, just buy and deploy. The hard, pivotal changes you need to have, I think you should build, at the very least, at the very first incarnation, so you really understand that problem. If you cannot afford to mess it up. build. It’s in line with some of the core thoughts here. Companies are getting more technical, they’re gonna hire more builders, that could be architects, it could be engineers, but some of the supporting organizations in go-to-market, like sales enablement, and of course, RevOps. they’re getting bigger, and they’re getting more technical, and they’re getting more expensive, but the yield that they will provide with the things they buy and they build are going to be orders of magnitude of what they were before. So, I guess my hot seat answer is, both.
Michele Buckley:
Oh, no, well said. You’re… I’m hearing in your comments Right? If it’s easy, automate it with something pre-built, like, you know, a light bulb, an Alexa light bulb that follows instructions, but if you’re rebuilding, a building, you’ve got to really get an architect in, is like what you’re saying. That’s right.
Santi:
note-taking, buy your note-taking, but build your forecasting. How about that?
Michele Buckley:
Hmm… ugh.
Tooba Durraze:
You will build your forecasting, really?
Santi:
I think so. I think in 2026, maybe in 2030 I wouldn’t, but in 2026, I… I want to put my job on the line on somebody’s own little five-coded, you know, agent solution.
Tooba Durraze:
Imagine there are solutions out there that are not, like, put it, Asian solutions, but I feel like there’s a world there where you let like, you can’t afford a neurosymbolic architecture if that’s sort of, like, not what the inception of your company is, and those do a better job of, like, forecasting than a handmade model would. So, maybe we could pick this up later, but I’d be very curious on why you would, built that.
Michele Buckley:
Large enterprises have a lot of technical debt that is holding them back. And difficult to pivot, the big cruise ship, so to speak. Yeah, but Thank you for that, Tuba. I love… I’m happy to have more than one person in the hot seat. Last to go? Who has to go is Nathan. What… what type of shifts are you seeing in terms of, like, investment? Say, marketing investment seems to be, an area that you’re focusing the most. In 2026, what shifts are you seeing there? or predict.
Nathan Gotch:
Well, I… yeah, I mean, I think… you know, since ChatGPT launched, SEO’s been kind of on the hot seat itself. And I think that’s been, you know, it’s taken some education, it’s taken some time to understand kind of what this is going to evolve into, and agencies have been kind of hit hard over the last couple of years with investment from businesses. And there’s, you know. Maybe this false sense that, you know, internally, you know, systems can be built, and teams can be built to… let’s say not hire agencies anymore, and I think that is possible. I think it can be done. But what we’re finding is that what has happened recently is a lot of the businesses that have attended to do that are back hiring agencies again. Because it was so incredibly difficult, and they’re starting to realize that, wait a second, this is much harder than we actually thought, to actually do SEO full-time when we have all this other stuff to do. Right? So, I would say that’s kind of the biggest thing, is, you know, agencies are getting much better at understanding the, the current landscape. And the landscape is, you know, it’s… as I mentioned in the beginning of this, has evolved, you know, basically Totally to the point where every business needs to be thinking about how they get the AI to actually recommend them in the AI answers. And so, if you’re not visible for these commercial queries in ChatGBT, Google’s AI Mode, Perplexity, Grok, you name it, any of these major platforms. you’re gonna be in trouble. Because… and there’s plenty of ways to influence these AI answers. They’re not, They’re not set in stone, let’s just say that much. And, I can give a very small example. At our company, we’re building an AI tracking tool. Where you can run commercial queries, and then we create synthetic prompts, and it will actually show what brands are popping up, and then we’ll extract and say. Nike shows up, their share of voice is 25% for this bucket of queries. Well, our tool hasn’t even launched yet. And so, but it… before, over the last, 3 to 6 months, we’ve been creating listicles saying that our tool is the best tool for this. And so now what happens is every time that someone runs these queries, like, what’s the best AI search tracking tool? Rankability pops up in these AI answers. And it’s only because we’ve influence those answers. And so it was more of an experiment to show that, you know, because no one has access to the product. We’re clearly not the best objectively, like, other than us saying it. And so, but these… these answers can be, for a lack of a better word, manipulated. And I’m not talking specifically about the static corpus, the actual, you know, raw training. You can’t influence that. But you can influence it when retrieval is used. And so that’s where a lot of SEOs, they really start to shine, because retrieval, one of the primary sources of retrieval is search. So, so anyway. That’s one of the biggest ones going on, and I think there’s been… I think it’s finally, like, establishing that AEO, whatever acronym you want to use, that’s one that I think most people are kind of settling on, which is Answer Engine Optimization. That’s kinda gonna be the thing, I think, going forward. So, yeah.
Michele Buckley:
That dynamic I have seen in lots of different sectors where AI is deceptively easy And so they’re like, we’re gonna do this in-house, and then they try and go, alright, we’re gonna try to build our own agent, we’re gonna build our own workflow, it’s a little bit harder. So…
Nathan Gotch:
I wish it was that easy, honestly, I really do.
Michele Buckley:
Well, so now it’s time, we’ve got about 6 minutes left, so we’re going to do a rapid-fire round, alright? This is a fourth section where we need fast answers, short, punchy, fast answers. If it’s a whole sentence, it’s too long. So, let’s go with some adjectives and nouns and the like. on… Let’s talk a little bit at the height as we go into the trough of disillusionment.
Santi:
What do you think is the most.
Michele Buckley:
overhyped use for agent AI, where it’s just, like. it’s just… it’s overdone, and it’s under-delivered. All right? I’m gonna… we’ll do a quick round now. So, Brent, you’re up first.
Brent Krempges:
Whew, start me with a tough one. What comes to mind is, like, where there’s ambiguity. Ambiguity in whatever it’s trying to solve.
Michele Buckley:
Okay, good. response, and very quick, thank you. So it’s overhyped, which is, yeah, you can do anything, no. If there’s, like, gray area, no, you can’t be done. Alright, over to you, Nathan.
Nathan Gotch:
AI is not very good at actually producing good video content.
Michele Buckley:
I like that. Hyped video. Yeah, I do. It’s really obvious, isn’t it? A lot of times. Okay, cool, next, over to you, Tuba.
Tooba Durraze:
Prompt-driven reporting. Don’t use AI as a calculator. It could do it, but where is that getting you? I don’t get it.
Michele Buckley:
No. It’s an optimistic calculator as well, so that’s so good. Werner, your thoughts?
Werner Schmidt:
For me, overhype replacing salespeople and SDRs. So, I don’t bet. I think it’s…
Michele Buckley:
choice.
Werner Schmidt:
Yeah, I can’t dispute that.
Michele Buckley:
At all, but, yeah, alright. Last we have, Santi.
Santi:
Werner did have my first answer, but I’ll redo it to go with something more.
Werner Schmidt:
Original.
Santi:
I’ll say, initial sexy, compelling demos that don’t actually end up delivering in the end. I don’t know if you all remember the days in which it was actually kind of hard to put together a demo for solving a hard problem. These days, you get a 100% hit rate on that initial demo. Anybody can do it. The problem is delivering in the end.
Michele Buckley:
Yes, fabricating it. Yeah, I have been hearing from a lot of clients that a lot of them come to us and saying, hey, and you know what I’m finding? The vendor’s saying they could do it, but then we bought it and tried it, and it didn’t do it. So that is, Hmm, that is something to be very careful of. Okay, let, we just got a couple more minutes, so I’ll skip ahead. What Agentic use case have you seen that’s having the greatest business impact, personally, let’s say, in the last year, in the upcoming year? I’m going to go to you, Brent. What use case is having the greatest business impact so far? Yeah. Right now, real life?
Brent Krempges:
Can this be more than one word?
Michele Buckley:
Yeah, you can freewood.
Brent Krempges:
Yeah, so for us, it’s mainly been around, like, manual risk identification, and I will… I’m gonna elaborate just a bit, so… For us, it’s more around organizational sentiment. Now that we have so much data flowing in from transcripts, emails, etc, we’re able to quickly identify sentiment risk, where in the past, it was more an objective-type risk, and so it’s allowed us to get much further ahead in terms of identifying and predicting risk. Within some of our own, workflows.
Michele Buckley:
Okay, good. sentiment leading to risk. Got it. That’s how you would do it in three words. Yes, good job, thank you. Yes. Okay. All right, over you. Nathan, what’s the greatest business impact you’ve seen from Agentic AI?
Nathan Gotch:
Well, definitely as a software founder, being able to create prototypes, and actually see the vision of what I really want, and then handing that off to the developers to actually execute that has been just so monumental. Because before, it would be like, have this thing that we want to build, and it’s like… here’s the project brief, I hope this works out, and then you get something back, and you get a V1, you’re like, this isn’t anywhere what I wanted to see.
Michele Buckley:
I know, you’re trying to draw it on a piece of paper.
Nathan Gotch:
Exactly, so now you can actually have the AI produce these beautiful prototypes, and then let the actual experts handle the rest, which has been very helpful.
Michele Buckley:
Got it. I’m gonna cut you off there. We’ve got 2 more minutes and 3 more people. Tuba, what’s you seeing as a big business impact?
Tooba Durraze:
Yeah, I would say, to Brent’s earlier point, like, unified intelligence, but extracting value out of your unstructured data. I think that’s… that’s awesome, where you see that going.
Michele Buckley:
Hmm. All right, great, excellent. Okay, Werner?
Werner Schmidt:
Well, with us being a software, you know, building software, which AI sits on, to be able to just surface insights that you just wouldn’t see before, and then being able to make decisions off that, it’s just, you know, very hard to go find it manually. Now it’s just been surfaced to you, and I think that’s… That’s been, you know, huge to then help organizations.
Michele Buckley:
Hmm. That would be the… Not the gray area we talked about earlier, but… The blind spots that you never even thought about before.
Werner Schmidt:
Exactly, those risks and opportunities, identifying the risks much earlier on and being able to course-correct when you still have time, and then if there’s opportunities, doubling down on it. So that now means, you know, you can surface that and see that much sooner.
Michele Buckley:
Excellent. Okay, hard to believe we are on time, so Santi is going to have a minute to provide us a last word on what areas you’re seeing have a great business impact, and Steve looks like a nice guy, so you might be able to, like, have 10.
Santi:
I’ll take a minute. So, I will second a lot of the insights here. I do think that extracting insights from datasets that are inherently unstructured and massive and have not been exploited yet is indeed the best, but I’ll take advantage of the fact that you all already said it, and I will bring up one out of the left field, and I’ll say. The delegation of tasks. Where you can now thoroughly explain exactly what you want, and have an AI model help you write it down in depth. disambiguation and proper delegation is actually a really hard thing that is a waste of time for companies left and right, so we’ve embraced that pretty deeply inside our company, and, you know, it’s not probably the highest yield, but it’s one I did not expect.
Michele Buckley:
Excellent. Thank you for sharing that, Santi, and my esteemed panelists. Werner Schmidt, Nathan Gottsch, Tuba Durace, and Brent Kempredge, thank you so much for being part of this and let, having fun and a laugh here and there as well. This brings us to the end of this panel. I see Steve lurking, I’m not sure what happens next, so do I hand over you? Just say thank you to everyone.
Julia Nimchinski:
Thank you so much, Michelle. Phenomenal panel. Thank you, thank you, thank you, thank you, our esteemed panelists, once again. And we shared all of your LinkedIn profiles with our community. They asked, but other than that, Michelle, what’s next for Gartner? What are you excited about? How can we best support you?
Michele Buckley:
Well, thanks for asking. I’m loving being part of this community, such great ideas. This is, you know, the people who are actually doing it can contribute. We’ll have another community event in March at Gartner, Product Leadership Conference, that I’ll be presenting at. Also, I’ll be at our Chief Sales Officer event In May, in Las Vegas, so I’d love to… if anyone’s there, please say hello, let’s… Chat more.
Julia Nimchinski:
We’re gonna amplify it. Thank you so much.