From Workflows to Agents: How Enterprises Will Operationalize AI in 2026 — Fireside Chat with Timothy Young & Jason Napieralski

Julia Nimchinski:
And we are live! Welcome back to Day 3 of the Agentic Singularity Summit! Today, we have an amazing lineup, the CEO of Otter.ai, Jasper, OneMind, Gong, and many founders, AI native founders, SalesTech, Martech, GTM Tech, VCs from Insight Partners, Lightspeed, and senior CXOs from Atlassian, Adobe, and U.com. We’re kicking things off with Jason Navarolsky, GTM advisor and long-time operator across AWS, Google, and Oracle, and Timothy Young, CEO of Jasper. Welcome to the show, how are you doing?
Jason Napieralski:
Doing great.
Timothy Young:
I’m doing great, thank you.
Julia Nimchinski:
One question before… before we start. We ask this every session, all three days. your top GTM and AI prediction for 2026.
Jason Napieralski:
Chaos. It’s gonna be chaos.
Julia Nimchinski:
Love it.
Jason Napieralski:
be lots and lots of stuff. It’s going to be fast and furious, and there’s going to be lots and winners and losers, and I believe that all incumbents are at risk.
Julia Nimchinski:
How about yourself, Timothy?
Timothy Young:
I mean, I think next year’s gonna be all about integration. I think there’s been a lot of experimentation, a lot of focus, obviously, this year on experimenting with agents, but I think, as most organizations are finding out, integration is really one of the big keys to make them successful. So I think that’ll be top of mind for everyone next year, especially in GTM world.
Julia Nimchinski:
Awesome. Mark?
Mark Organ:
Yeah, I think it’s… it’s all about light, local, and specialized. Whereas I think, you know, to date, it’s been kind of large, huge spend, very general, and I think we’re gonna see… Just more hyper-specialized applications where there’s deep domain expertise. involved. And maybe where AI is being applied later, as opposed to, hey, I’ve got this AI thing, what can I do with it? It’s more people who really understand the domain deeply and then use AI to really turbocharge that, that process.
Julia Nimchinski:
Awesome. Mark is the original founding CEO of Eloqua and Influitive, stoked to have you here. And Jason, the stage is yours.

Jason Napieralski:
Sure, thanks. I think we’ll get right in to this. I think we’ve got a pretty heady topic here, and I think it’s going to be a little bit spicy. I just wanted to kind of set the stage here of, We want to talk about enterprise, and we’ve got some great people here to talk about Enterprise. You know, Timothy and Mark have obviously spent a lot of time in there, and Timothy is We had a short conversation, he’s on the ground trying to implement AI stuff in enterprises, and so he knows what’s going on. But I think, to set the stage, Mackenzie just did a report about a month ago, and I’ll share that with us. Let’s see… here we go. Does everybody see the screen? Yep. So, anyway, but the… this kind of encapsulates, and I think it’s pretty accurate from what I’ve experienced, but I want to get you guys’ take on it. The general gist of this is that everybody is trying AI, and very few are getting past the pilot stage to get into production of any kind. And when you see the left here, you see the average majority, we see some project ROI, and I put that sum in there myself. specifically because, the ROI calculations are not that we… our business got an ROI, it’s that we gave a developer a $30 license for Cursor, and then they were able to be more, 30% more efficient, so we call that ROI. So it’s… it’s a little bit of a red herring, that people are getting some benefit, but, you know, beyond kind of the ChatGPT frontier. We’re not seeing a ton in enterprises. So, just wanted to put this up to show, kind of, where the big businesses are thinking here, what the successes are, and then, you know, kind of talk to you, Timothy, and get your take on this. What do you think reads true, and what do you think is not right here?
Timothy Young:
Yeah, no, I mean, I think it’s a… it’s a great set of questions. I mean, you know, at Jasper. we work now with about 25% of the Fortune 500, so I see kind of a wide swath of organizations, and in total, we have about 1,000 enterprises as customers. You know, I would say that, you know, definitely customers are in a kind of a crawl, walk, run, but for most of the last couple years, they’ve been focusing on, you know, leveraging AI as an individual employee experience, which is, you know, potentially helping on productivity gains. But when you talk to leaders at organizations. They’re not seeing that ROI in the metrics of the business. They’re not seeing it hit the P&L. And I think what, you know, probably the most progressive, organizations are finding out is that it’s really not so much an employee transformation as it is a business transformation, right? So, what you really have to do is isolate business processes, workflows, you know, even workflows that you’ve never done before. and apply those to AI, thinking about AI as a new piece of foundational infrastructure, not as just an employee productivity application. But that, I think, is exposing a lot of organizations and providing a challenge in that many organizations do not have clear documentation on what their business processes are, or what contributes as inputs to a key performance indicator of the business. So a lot of organizations are now being challenged to kind of really explicitly detail out these processes so that they can apply them to AI to get that. ROI that’s going to end up showing in the P&L, and that’s a… that’s a process that most organizations are kind of just beginning to get into, now that they’ve seen a little bit of how it can work at an individual level, trying to take it to an organizational level.
Jason Napieralski:
Yeah, great answer. I want to go back to the point where you made that they don’t understand their processes, but first I want to get Mark’s kind of overall take on the study there.
Mark Organ:
Yeah, no, I’ve heard a lot of the same things. Lately, I’ve been really interested in the manufacturing space and robotics, and you know, it’s… remarkable there, just how these pilots are really not generating that success. I do echo that, I think for a lot of companies, you’re right, that they don’t really know their processes very well. They’ve not done a matrix on, like, which processes are the most broken, and And, and most important, that’s, And then, you know, using AI to, you know, to fix it. I think in many cases, I think companies would be well served by having a Skunk Works type of approach to it. I mean, this is… this is something that startups do really well, and big companies don’t do well. And so, like, what I’ve seen is that a number of these companies are kind of folding AI under, like, an innovation group, or they’re… it’s run out of the IT or out of the CIO’s office. And I think that’s wrong. I think what they should be doing is putting an office somewhere else, across town. I think they should be putting young people in it and challenging them to, you know, disrupt the current business if they can. you know, the classic screw-skunk’s approach, I mean, Disrupt your business before someone else does. And so, yeah, I think, I think enterprises could take a page from… from the startup playbook on this one.
Jason Napieralski:
Yeah, and Timothy, so you mentioned the processes, and you did too, Mark, but the process issues, is that a new problem for enterprises?
Timothy Young:
Well, I think it’s not necessarily a new problem, but I think, you know, humans, you know, and human memory have basically been the crutch for this, right? Where, if you think about, you know, how we’ve applied software, particularly in the last era with SaaS, everything was very deterministic. Right? I know if I put this data in, I run this process, I’m gonna get this repeated, consistent outcome every time. It’s very predictable. I think what enterprises are finding out is much like humans, AI is probabilistic. It’s not always rational, it’s not always repeatable, and so what you have to do with AI is you really have to apply not just the best model, where everyone’s been focused, I think, the last couple years, is you have to apply domain memory to that problem, and use that domain memory, which is all the context about your business, right? Project specifics, user preferences, learned information, feature lists, requirements. You need to pair that with the AI in order to get that repeated outcome. You gotta have the guardrails, you have to be able to have the scaffolding there. And most organizations are just, you know, straight applying models and basic prompts and kind of hoping that they’re going to get some organizational benefit. And that hasn’t proven to be. And so this, you know, the memory layer and the memory, you know, context challenge, is just starting to kind of hit the industry, and just starting to kind of, you know, hit enterprises as where they need to focus their efforts. And I… and I would agree with Mark, a lot of that is not going to be you know, in the IT department, right? That domain memory is in the employees’ heads, is in the… embedded in the organization, in the tacit knowledge they have, you know, in the marketing department, in the software development, in legal, in finance. And you know, so organizations are going to have to find ways to unlock that domain memory and partner with employees, and there’s also just a lot of human and psychology challenges in getting that information codified and ready for AI to leverage.
Jason Napieralski:
Yeah, yeah, hope is not a strategy, right? That’s kind of where they’re going. Mark, to that point, your IT point you made, which I think is an excellent one, the… do you think that there’s a fundamental misunderstanding of what a business process is, because it’s been muddied with all these tools and systems? So, for example, a sales process and a CRM process are very hard to disconnect from each other. They become the same thing. And do you think that that’s part of the issue here, with figuring out what exactly you’re trying to do with AI in the first place?
Mark Organ:
That’s a good question. You know, from my experience, I think companies do… they do understand what processes are? I mean, I just think of it as, like a flow chart. I don’t know if that’s… if that’s… if that’s muddied with, with, with tools. But I… I think that companies, don’t… Focus enough on the processes they have, and sort of inventorying them. And rating their… their success, and having a dedicated… a dedicated process for managing their processes. I know when I was running a company, that was my company, that was a big part of, kind of, what the COO’s job was. It was like, let’s see all the different processes in the different departments. And then, you know, every quarter, come up with some processes to fix. And, I use the broken importance matrix, very simple 2×2 matrix. Let’s look at the processes that are, really important to customers in terms of driving value and experience. And then… the ones that are already broken. A lot of these really go across departments. You’ve probably seen that. The processes that go, for example, across sales, marketing, and customer success, or across customer success and operations, those Tend to be the ones that are both important and broken. But yeah, I can’t… I can’t really tell you, you know, about Process adoption tools and kind of the, the, and the effect of it. What I… what I do know is that I think companies would be very well served to have more of a focus on the processes in their company, so that they’re constantly identifying the ones that ought to be worked on and fixed.
Jason Napieralski:
Yeah, and I come from the Jeff Bezos school of thought, where you anchor your everything about your company on your customer and work backwards from that. And I feel like… and I feel like, and I’m part of the problem, I worked in big tech selling SaaS software, and really what we wanted to do was sell software, and we wanted to make it as sticky as possible, so what we did, oftentimes, particularly in the The go-to-market phase, is that we would… try to get the company to adopt the processes that are built into the software, rather than adjust the software to conform to their process, therefore polluting the whole process chain and linking it with the software. And the reason I ask that is I think, and you guys may have relationships with Salesforce, so I will say this is my opinion alone, and this is not the… shared the opinion of Timothy and Mark, but I think Salesforce is the worst thing that’s ever happened.
Mark Organ:
What happened to software and technology and productivity in the world.
Jason Napieralski:
And the thing that I think is happening a lot, or I see is happening, and I don’t know if you guys are seeing this, is that these companies are focused on improving their usage of the tools. So Salesforce wants to put AI in there so Salesforce stuff is easier to do, or you get better insights from Salesforce data. But, what I think is a fundamental mistake here is that you are missing an opportunity to get rid of the whole CRM process in totality. and to figure out a new way to go to market to your customer, removing that bottleneck and horrible data source. You know, CRMs will rely on salespeople, which I think everybody here can agree that’s the worst people to enter data into a system of record. I believe that the systems in the enterprise, particularly, are polluted. by bad processes that were designed to sell and maintain software, and not designed, and they’ve lost the connection to the customer and the connection to the business, and they don’t know why they’re doing these things. So if you apply AI to that, you get to Timothy’s point, where they are hoping AI will fix this. their already broken system, process, and company structure, and then when it doesn’t, they get really disappointed really fast. Have you seen that, Timothy?
Timothy Young:
Yeah, I mean, for sure. I mean, I think… I think this is the big challenge for… you know, for companies who have deployed, you know, a ton of SaaS, and then SaaS companies themselves, which is, you know, if you think about what SaaS was, it was essentially you’re selling access for humans to a database through a user experience layer, because most humans cannot write SQL queries, right? So, you sold them this UX with C, you know, licenses, but what happens when agents and AI come along, right? They don’t need to use a user interface, they can speak SQL, they can talk to the data, and so what we’re seeing customers starting to deploy is you know, less usage on SaaS, and they’re having agents doing structured work, routine work, but they’re just going right into the data warehouse layer, right into the CDP. Right? And they’re, you know, accessing the data there, analyzing it, manipulating it, leveraging it, and they’re bypassing that whole SaaS software stack that was really a user experience built for humans. Right? And, I think that is something, that is exciting, because we want to shift the routine work to agents and AI, so that my salespeople can focus on strategic activities, so they can focus on relationship building. Right? So they can build skills. I don’t want them to be doing data entry, and manually adjusting their forecasts, and all of that routine work is just wasted time, right? I want them actually out in the field, working with customers, and let agents handle all of that. And then, when you do that. you find that there’s a huge reduction on your usage of SaaS. Because, I mean, I work with large enterprises, and when you talk to them. hey, and you know, you talk to a CMO, what’s the average number of SaaS apps they have, just in the MarTech space? It’s on the order of 150 to 200, you know, and so if you think about the challenge in managing all of that. It’s immense, and they all want to reduce that, refine that, and, you know, we want to get humans out of you know, software data entry and manipulation, and push that to AI, and allow people to do more human work.
Jason Napieralski:
Now, you mentioned the people aspect, and that’s, I think, this is a tough one, because it’s… the answers that I’m getting don’t make it feel very good for those of us who are employed in these companies, but… The… one of the theories that I have, and what I’ve seen, is that, in smaller companies, there’s more generalists. And generalists are what you need in order to actually use AI in the way that it is in its current state, in the non-deterministic state. And that you need to… generalists need to understand enough about the business and the things that are… that you’re asking the AI, so when it gives you an answer, you can give it a smell test. You can tell if it’s hallucinating. If you’re a domain expert in marketing. Versus an IT person who doesn’t know anything about the marketing, but is managing the marketing system. You know, there’s a very big disconnect there if you are managing the AI systems in there, because you don’t know what’s right or what’s wrong, and hallucinations are real, as well as the context problem. My… my, I guess, more spicy opinion is… is that in… as you go up the chain in enterprises, that more specialization occurs at every level and every position. And that is directly, directly proportional, inversely proportional to success with AI, in my opinion. And that each of these people are very specialized. And that there’s a challenge in whether or not we can retrain these people to be generalists, to be AI pilots, and how many of them can make the shift. And developers are some of the worst, and I don’t know if you know this, or if you see this, Mark, but developers are some of the worst at adopting AI, there’s definitely a big, a big… on the better developers are like, you’re not gonna out… outdo me, kid, kind of attitude towards the AI tools.
Mark Organ:
Yeah. Yeah, no, I have seen that, and I mean, I think it’s not just true in developers, and maybe it’s a couple of things. I mean, one is it’s a change management problem. People are used to doing things a certain way. I mean, even me, I struggle with this sometimes. And I think AI, to do it well, sometimes needs a complete rethink. But I also think that there’s… there’s an emotional issue. If you’re a senior developer, and you… you consider yourself a craftsman. You know, it’s probably not that different from, you know, the… the… the weavers that smashed all the looms way back, way, way back, the Luddites, that’s where we get the Luddites from. Because, you know, and it’s… as it turned out, you really needed a whole different class of weavers. That enthusiastically adopted the new technology. And I think we’ve seen that in disruptive innovation after disruptive innovation. So, it’s interesting, you know, I’ve seen this prediction that the, You know, that… that you… you’re… there’s this absence of these junior people. Like, for example, if AI can do almost anything a junior lawyer can do. will we be training any more lawyers? Because the… the step ladder is no longer Is no longer working. you know, it’s an interesting question, but yeah, I think that there’s, I think that where companies should really… they should really be building a process To disrupt itself. And looking for the people inside the company that are ready to take that next step and to say, hey, this… to your point about building around the customer, here’s the way we drive value or experience for the customer today. Find a 10 times better, faster way To do that. And you can use any technology you want, you don’t have to use AI. There may be a lot of times where, frankly, we’re sprinkling AI on something and it doesn’t need to be there. Use AI if you can, but I’m gonna give you a small budget, so with a small budget, you’ll probably need to try to use AI in some way. But that’s where I’m seeing big… the biggest innovation I’m seeing using AI are companies that are desperate. That are behind the eight ball, they’re the number 6 player in an industry, they’re not able to raise capital, they’re not able to hire the best people, and out of that desperation comes amazing innovation. And that’s where I’m seeing mind-blowing stuff with AI. And so I think companies that don’t have this emergency, that have lots of money and lots of talent, should create an artificial emergency. And I think that’s actually where we’re going to see great stuff coming in AI.

Jason Napieralski:
Yeah, and to your point, Mark Andreessen, just said again, and he said it a number of times, but he said that he is basically short on all incumbents right now, because AI is going to eat their lunch. And, you know, let’s… we’ve got a few minutes left. Let’s kind of focus this last kind of point here on what… if you are a big company, that you… and you… and you’re not hurting yet. How do you make yourself hurt? How do you… how do you understand… how do you get religion on the fact what Mark is saying, is that there is somebody with 10 people that is going to do what you’ve done with 8,000? And how do you get their spine straight and to be creative and desperate to actually try to get this stuff to work? Timothy, what do you think?
Timothy Young:
You know, it’s interesting, we do a lot of workshops, you know, with customers, and, you know, we started out doing them, you know, 18 months ago, really focusing on, hey, let’s map out existing processes that are important to the business. The challenge there is when you talk to any individual, they usually only know their part. you know, their step of a broader process, so it takes a long time to figure out. And then I think, to Mark’s point, you know, if you start thinking about approaching those processes with AI, you run into a whole bunch of organizational psychology dynamics that are very challenging. So we’ve tried in, you know, in the last 6 months, I think, to a lot of success, in really, you know, kind of inverting that, and sitting down with organizations and saying, hey, what has been impossible until now? Right? Imagine now that you have AI, you have kind of infinite junior labor, infinite… You know, junior intelligence, it’s always getting at a foundational level smarter. But what are the impossible things that you’ve always wanted to do in your business, but you’ve never been able to do because of resource constraints, or timing, or past leadership? And then let’s start from there. And you co-create it with teams, and you create brand new processes coming out of it, and you kind of… Look at it that way, and then you take that group, and you begin to ring fence them, kind of, you know, internal compete, internal challengers, to old models. And old, you know, older ways of working. And what we’ve seen there is you build kind of a groundswell of excitement in the company where it’s not about replacement, or doing better, but it’s actually, you know, completely net new innovation. And where we’ve seen companies push that way, I think they’re starting to see a lot more success and a lot more, you know, groundswell adoption and buy-in from employees.
Jason Napieralski:
Yeah, great point. Same question to you, Mark.
Mark Organ:
Yeah, I’m a big fan of the Skunk Works approach. I mean, look, this is a problem that has been solved before, and including by big companies. So, some, you know, Gillette is famous for disrupting its own business. 3M, has a challenge where they basically try to create a new business every 3 years that is you know. like, a half of their current kind of business. And, you know, and it works. The original Skunk Works came from Lockheed Martin that had to come up with a 10 times better, faster, cheaper way to build a fighter plane, you know, in World War II, and so they actually had a factory that was literally called the Skunk Works. and put a bunch of, you know, the 1940s version of nose-ring, purple-haired people, you know, in there, and gave them a crazy huge challenge. I did that in my last company, where I found that my engineering organization was not really, adopting mobile enough was not… was not mobile first. That was actually a common problem in… in the industry. And so I basically took, You know, a bunch of really young people in their early 20s, and they went to a whole different city. And with the only mandate is build a 10 times better, faster, cheaper way to get people to advocate for companies that they love. Didn’t specify a technology. But they, of course, gravitated towards mobile, and they… they also… what’s also kind of cool is they innovated in a whole number of other areas that ended up challenging the corporate center. You know, corporate center is just like, well, how are these guys five times faster than us in these different areas? maybe we should adopt their… their processes. So, I don’t think this is done enough in companies. I think people are afraid, afraid of things like this. Other things I love, too, I love hackathons, are also great. The best features that my last company ever produced never came out of any product roadmap. It’s unbelievable. So you have these product managers, they’re speaking to customers all the time. Their job is to go and create create new stuff. And all the best things came out of this super hacky, pirate-like approach, where people who are front and center, who are touching customers, like customer success people and salespeople, were pitching ideas, developers would build them, roll them out to customers within 24 hours. And then, wow, you know, those would really take off. So, I think big companies can do this sort of thing, too. It’s not… it’s not just startups. Should think of… and… Think about the people in the company that they could use to mobilize to do these things.
Jason Napieralski:
And it’s not a should, right, Mark? Don’t you think? At this point in this phase, if you listen to Mark Andreessen, isn’t this a have-to? Isn’t this, like.
Mark Organ:
I will… I…
Jason Napieralski:
You have to do this, or you’re going to be on the, on the, on the graveyard.
Mark Organ:
I agree. I think AI-first companies are going to disrupt pretty much every industry. I actually agree with them, and matter of fact, I put my money I’ve done in my investing, I do a lot of that, like, shorting incumbents, and then looking for the, you know, young up-and-coming, you know, companies. So no, I think that’s exactly right. I think every company should be, doing some… I just call it a pirate process, whether it’s a Skunk Works, whether it’s a hackathon, some way to hack the system and get way faster learning and experimentation. And you don’t even have to mention the word AI. You know, if you have the right people and you have a really exciting mandate. there’s no option. Like, what else are you gonna do? You have to use it. Yeah. Yeah.
Jason Napieralski:
Yeah, so that’s, Julia, any burning questions from the, audience that we need to answer here in the last…
Julia Nimchinski:
We only have one minute left, Jason, and the burning question is, what’s the best way to support you?
Jason Napieralski:
To support me, it’s, I’ve got an app that I’m releasing called AI Chef, so if you like to cook, I need some beta testers too, so if anyone knows, reach out to me and they want to test this thing. It’s an AI-generated recipe thing that organizes things. Basically what ChatGPT does, except for it’s not your chat anymore, it’s in your inbox, so that’s what I’ve built.
Mark Organ:
I love to see that. Actually, it’s amazing how much AI has transformed the way I cook. Yeah. It’s no longer about recipes. It’s like, what do I have that’s fresh? And then, you know, make me something that’s Thai-inspired, and blah blah blah, whatever.
Jason Napieralski:
That’s exactly what it does, yeah, except for you can save it and organize it, unlike with ChatGPT.
Mark Organ:
I’m in, man.
Jason Napieralski:
Yeah.
Mark Organ:
I can’t.
Jason Napieralski:
Alright.
Mark Organ:
Pull that up, that’s awesome.
Julia Nimchinski:
Deputy, how about yourself? What’s the latest and greatest with Jasper?
Timothy Young:
I mean, we are… we are growing extremely fast. You know, we’re a company that started out kind of in the prosumer space, right? We launched 9 months before ChatGPT. The last couple years, we’ve been, you know, completely focused on shifting the business to enterprise. You know, this year, in terms of, like, headcount, the company will double. So, if you’re ever interested in working more closely in AI, you know, especially around MarTech, we’ve got opportunities everywhere, and then, you know, if you’re in a marketing organization at a small, medium, large enterprise, and need help. We’re here to share best practices, stories, and you can start using the platform just free out of the box. And really accelerate all your marketing workflows.
Julia Nimchinski:
Love it. Thank you so much again, and we are transitioning to our next panel.