What an AI-Native Turnaround Actually Looks Like

JaySo an AI system can't take responsibility for an outcome. it AI can't do the human side, right? It can summarize and explain detail, but it has no empathy. I do think SaaS still exists, it's just the interface changes.Monisha SaldanhaWelcome to Colorado Tech People, where we learn from the leader shaping technology in Colorado and beyond. I'm your host, Monisha Saldana. Turning around a SaaS company is usually described as a strategic problem. But what if the first thing a new CEO needs to do is simply listen? Today I'm joined by Jay Smith, co founder and CEO of WebScale, who spent more than a decade as the company's technical leader. Before stepping into the CEO role in 2025, we're going to talk about what Jay discovered when he took over, the difficult decisions involved in turning around a SaaS business, why employee insight became so important, and how technologies like Claude are changing what a relatively small team can build and operate. Welcome, Jay.JayHappy to be here.Monisha SaldanhaHappy to have you here. Okay, let's dive in. Let's start with the leadership transition. You were WebScale's co-founder and CTO for many years before becoming CEO in 2025. What led the board and leadership team to decide the company needed a change at the top? And why were you the right person to step in at that moment?JayThere's a number of ways to scale a company. You can go innovation, you can go, you know, standardization and achieve at scale and undercut on price or you can be intimate with your customers. So customer intimacy. I think for us, how we land is actually the intimacy part. We actually know our customers very well. many of them have been with us for a very long time. So that knowledge and expertise with specific customers, or actually all of them, and our willingness to pretty much do whatever it takes to achieve their goals, solve their problems, has led to us creating a set of unique features that are really specific to that customer set. I think my history with the customer is kind of the obvious choices to why me. Why the challenge or why the change, I think ultimately it came down to a transition that was obvious. So AI, everyone's talking about it. It's what's happening right now. there was an opportunity to apply AI within the context of e commerce, so specifically to our customer set and my background in machine learning, it's an obvious connection there. the transition to an AI native company, it only succeeds based on leadership. And so the belief was, well, having experience with AI puts you on the path to success with achieving that turnaround within the company.Monisha SaldanhaAnd when you became CEO and could suddenly see the entire business rather than primarily the technology organization, what were the biggest problems you realized WebScale needed to solve and how did you go about solving them?JayWell, it starts at the beginning, being able to see it. when I took over, we had about twenty independent systems that to really understand the health of the company at any given point in time, you needed to understand the data in all 20 of them. And they're all point solutions, none of them talk to each other. they essentially just had little bits of the problem. And so to get a comprehensive view of the problem required building a system that could integrate all 20 into a system. single working data set that allowed us to support every team's goals so account management being able to see contract details and who's renewing when they're paying and what their overages were and then helping account management do their job. I mean same thing in support, same thing within finance, being able to accelerate our time to close and and really getting an insight into in an automated way, get the busy work out of the way so that the creative people I have on staff could actually focus on the real problems to make the company better.Monisha SaldanhaI've heard that listening to employees played an important role in how you approached the turnaround. What did the people inside WebScale understand about the company's problems that leadership had perhaps not been hearing clearly enough?JayI think the biggest thing was that manual labor bit. the rote data entry and the dissatisfaction with their role and what they had to do in their role, that they just didn't feel supported. And listening to that, hearing that was obvious. I don't ever want anyone to describe their job as soul crushing, which I actually heard from some folks. that was understanding what was getting in the way and then doing the work to get it out of the way so that they were able to do creative work and help make the company better. versus do the same thing day in and day out.Monisha SaldanhaAnd when you listened across the organization, how did you separate symptoms, things employees were frustrated about, from the handful of underlying problems that actually needed to change?Jaythe common theme about data entry and the lack of automation, really it's not so much that they asked specifically for it, but they were saying, Well, I you know, I can't do that because they have to do all this other work. And then when I dug into, well, what's that other work? What are you doing? Where are you spending your time in support? It's you know, responding to alerts that you have no action. And so then realizing that there was a whole lot of busy work throughout the organization and that by eliminating the busy work, it put them in a position to be more creative and think about the success of the company and really get us aligned to goals. it's important to not tell people how to do their job but to give them interesting problems to go solve and trust that I have creative, curious folks on staff that will go attack the problem and solve it.Monisha SaldanhaVery cool. Turnarounds inevitably require saying no. What did you decide WebScale needed to stop doing, simplify or deprioritize so the company could focus on the things that really mattered?Jaythe previous administration was trying to push on functionality that didn't have clear customers, had big competition and we we definitely had to make decisions, hard decisions that involved shrinking the team in some aspects. So we just couldn't pursue certain avenues anymore. It's a challenge, right? 'Cause you're basically placing a bet. you have the data that you have and you have to make the best decision about how to spend your resources 'cause we have limited resources. Now I think in our case, given that it's been a year or over a year, the data's showing that I think we were right, in the context that you know, we were positioned well to double down on our e commerce business. we jettisoned some infrastructure work that I think was maybe not aligned with where the core of the company was, where the revenue was certainly. And then being able to focus on our shopping assistant and ultimately the application of AI to help our e commerce businesses succeed. that's really borne out as, yeah, we have the skills, we have the people and we're able to do that.Monisha SaldanhaAnd how important was the arrival of genuinely capable generative AI to making this turnaround possible? Could you have implemented the same strategy three or four years earlier?Jayit would have been a lot harder, especially the integration and the conversion within the company, right? So you know, how you end up with twenty independent systems that don't really talk to each other is no insight really. Every team operating the way they need to and what I think we lose is collaboration, right? So it's impossible for different team members to see what's behind some wall. So you end up with silos within the organization and the silos don't play well together. And so generative AI had put us in a position to write the code quickly because all the APIs that are out there for these systems, they're common, it knew them, pretty straightforward to take once you had an idea of hey, I want to go build this integration and make bring the company together, straightforward to go execute with generative AI. Without it it's just a bigger project, more work. We probably don't have the funding to or the time, frankly, to get it done or to even know that it's done correctly.Monisha SaldanhaClaude has evolved extraordinarily quickly. Can you give us a concrete example of something WebScale can now build or automate with Claude that would have required considerably more people, time, or money only a year ago?JayYeah, definitely. So, my finances actually. So financials are automated now, derived directly from QuickBooks and Maxio and I get more timely access into the operation of the business. Something that, you know, in the past it was the account team within Excel producing an output that was very manual, labor intensive, took lots of work to get to the outcome. And now it's more I mean there's still work, but it's more qualification and verification as opposed to having to brute force manually build something. And that that's directly that's Claude writing the code to go do that.Monisha SaldanhaThere's a lot of discussion about putting AI into products, but you've also used AI to change how the company itself operates. Where has AI created the biggest internal productivity gain? Engineering, support, research, customer success, or somewhere unexpected.Jayso we we have a lot of success in support. That's definitely true. I think it's allowed us to run a fairly lean organization. it understands I mean, I think that the the transition in sport is probably an interesting one to consider that In support, we used to hire for very deep technical skill. we still have some very technical folks on on our support staff that understand in detail how everything within the company is built, for the customer, how the customer applications work. but it's changed the mindset. I think really it's focused more on people and networks of people is what I would say. That in the past you might ha hire someone who could be particularly gruff or it but is really technically skilled. That doesn't fly so much anymore because I have smart people on staff, they can learn. They're very good with people. So the people side matters a lot because often they're talking to a merchant who's experiencing some difficulty and you're trying to explain to them how their code impacted his performance. And now you you have AI telling you the detail. So you have to be capable enough to understand what it's telling you to say, Yeah, that's right. I get that. This data supports it, you can test it, you know that it works. But then more importantly, you have the people skills to actually communicate with Another person on the other end who really cares about the outcome and is stressed because they're experiencing a problem or some difficulty. And so It really puts more emphasis, on empathy and your ability to communicate complex concepts to somebody who's non technical and doesn't necessarily understand what you're telling them. And so you're acting as kind of this middleman and explaining, you know, middle person explaining this technical detail to somebody non technical to say, yeah, this is what we're doing and this is how and this why it happened. and I think that transition what what AI has done is it's made place because it AI can't do the human side, right? It can't pick up the phone and call somebody and say, Hey, we're seeing this within your system and explain to them what's happening, right? It can summarize and explain detail, but it has no empathy.Monisha SaldanhaAnd as a technical founder, how do you decide whether something should be handled by an AI agent, traditional software automation, or a human being?JaySo an AI system can't take responsibility for an outcome. So if I have like one of the support problems, right? The AI system could diagnose and it can recommend, but it can't make a change because it can't be responsible for the outcome of the change, right? If it says, Hey, I need to replace this system for whatever reason, right? I if that happens to break things worse, right, it it has no accountability for it it having done it the way. So there's a person in the loop at the place where there's revenue at risk, right? So it can summarize, understand, digest data and point you in the right direction to make the people that are responsible for that decision to make it it take action, right? So it can propose the information. But then a responsible party has to take that and then decide what action to actually ultimately go take. so there's there's kind of that, there's sort of, well, is there revenue at risk in that regard? I think the AI agent work timely, so if you're working at machine time, if it's you know, simple, repeatable issues, things that we understand well as a deterministic outcome, that that all bodes well for it. I think it's The easiest split to me is the responsibility piece where if there's revenue at risk, having a person in the loop to help make the final decision and take accountability for that is a big deal. if for the other pieces though, in terms of whether I build a control system that takes the the data and then automates some piece of it to automatic work. flow or something like that versus I have an LLM in the loop that's doing things on the fly, it really comes down to how repeatable what is it I'm trying to do. So so the the problem is what dictates that, right? So if it's fuzzy inputs, I get a ticket from somebody on the internet who is telling me there's something wrong or they want some change, an LLM might be more appropriate there because it's not a API call, it's not a structured system that you're talking to, it's a person. So you're translating what they're saying into, yeah, here's a set of actions that we think need to be taken. And really as to I hand that off to an agent or a person depends on the level of risk associated with the outcome.Monisha SaldanhaThat makes a lot of sense. WebScale is now positioning around agentic commerce. What fundamentally changes when e-commerce moves from people clicking and searching through pages to customers simply expressing intent to an AI agent?JaySo I think fundamentally it's the exact same. It just to me it appears to be a different channel. Right? So I may communicate to my agent, hey, I'm interested in this product, but I have these requirements. I want, you know, this price, or I have some need of how I want to buy. And then I'm handing the agent off to go buy it for me. So to me, that that's more it's still there's a buyer, it's just a different channel, and the channel looks different. So in my channel now, I'm asking a problem from my perspective of what it is I'm trying to solve. the LLM is then recommending products or solutions and that ultimately is going to turn into a buying decision. So if I'm asking about, you know, 10 providers for wet weather or whatever the the problem may be, I'm having that conversation with the LLM to understand my problem, then ultimately surface products that help me solve that problem. And then I'm going to wanna buy through that channel and now When I'm buying through that channel I'm not doing it directly, I'm handing it to an agent to actually go by on my behalf to basically find the product for the best price, deliver it at the right time and satisfy my requirement. So ultimately I see it as a new channel It's just another way for customers to buy. I don't think the agent it may make a decision about which vendor to buy from or how much, what's the best price, that kind of characteristics of the decision. But Ultimately there's a customer on the other end that's delegating that's saying, Hey, I wanna buy this product and go buy it. So to me it's still same thing, just the there's a little more automation in the flow and the characteristics of the channel are a bit different.Monisha SaldanhaYou went from founder and CTO to running the entire company. What was the hardest part of changing your own operating system, from being the person closest to the technology to being responsible for people, capital, customers, culture and strategy?JayLetting go. I don't scale. so the challenge when you know what's happening and you know in detail technically exactly what's going on, there's a tendency, especially if the customer is particularly important or strategic, to jump in and start doing things. but then that it kinda takes away from the team and it gets in the way more than it helps and you know, if I were to do that, then the team doesn't learn, the team doesn't get capable enough to solve those problems on their own. Most likely they are already and there's things I can't even see that because I have limited time, right? So, their time and capability, it's trust that I have the right team in the right place and just let go and let them do their thing. I'll jump in the background and ask questions and probe and and it's hard to to do that versus tell, say do this versus how's that working for you, you know, maybe there's another way to do this, you know.Monisha SaldanhaWebScale raised new strategic funding at the same time you became CEO. How did you convince investors that this wasn't simply a turnaround story, but that there was a meaningful new growth opportunity around AI and agentic commerce?JayI think we were at a unique place in time, just given sort of where we were and the existing customer base that we had in e commerce, right? It was an obvious like let's go build. We had the skills, we had the the timing, it seemed like right place, right time, with a customer set that was interested in what we were doing and that it made sense.Monisha SaldanhaWebScale also placed greater emphasis on partners as part of the new direction. What did you learn about the company's go-to-market approach and why did partners become such an important part of the strategy?Jaypartners historically had been very important to us. so the in the e commerce world you typically have a merchant, they outsource to an agency to go build their e commerce site, and there's a hosting provider in the middle that talks to both parties and helps deliver. We've succeeded historically by being a true partner to our agencies. we don't point fingers, we don't blame, we focus on helping the customer, the merchant, solve their problem in a constructive way collectively with the agency. and we have the expertise to go deep, but we go deep into their code for recommendations to them to help them address root cause and work through issues. That got lost somewhere in in the middle. I think our recognition that our partners are the strongest go to market for us is still true. we do a lot of work with our partners. We essentially just return to our roots and said, Hey, here's how we work with our partners to achieve success and when we win we all win and it's better for the merchant.Monisha SaldanhaOne of the dangers in a turnaround is changing so many things that employees experience constant whiplash. How did you communicate the new direction clearly enough that people understood not only what was changing but why?JayRepetition. we were consistent. We knew what we wanted to do, and then we said the same thing over and over again to the team. I think even with that, the whiplash is still there, right? They see change, they see people come and go. So what we needed in support changed where we need this empathy and we need this like human focus versus you know deep tech like the the old ways weren't good. We didn't need senior Linux sys admins so much. We needed people that understood products and could communicate to customers and understood the technology but they were communicators first. that inevitable change that came with the personnel on the team, is still it was a challenge to manage. I mean it constant communication, I'm just communicating the success and helping the whole company see what we were doing, the outcome, right? And as we started to see the the outcome and the turnaround take shape, the everybody felt it.Monisha SaldanhaWe often hear that AI will allow companies to operate with smaller teams. What does that actually mean for employees at WebScale? Does AI primarily eliminate work or does it allow people to spend more time on doing higher value work?Jayit's the latter. It helps people do higher value work. I don't necessarily believe in the smaller team thing. The smaller team thing suggests that more people do more work individually like you have more outcome per individual with AI and I don't think that's necessarily true. I think you get better outcomes. So collaboratively, it's like having, all the world's knowledge at your fingertips, and then that means you have better insight into what's going on. It changes the nature of work, I suppose. People feel more satisfied with the work that they do, more confident that they're right because they have the support of the system to help them understand what's going on. We've been able to scale you know, in the turnaround the early bits, the scaling starting to happen without the need to we're hiring at a different pace and we're hiring in different roles. And we certainly are automating some frontline things like in support they feel it probably the most, where a lot of the busy work is taken off their plate. Now that doesn't shrink the team. It means that they have more time to be thoughtful in their response, thoughtful in their investigation. They do better work, I would say, as opposed to less.Monisha SaldanhaThere's an interesting paradox in your turnaround story. You've used more technology and automation, but you've also placed greater emphasis on listening to humans. Why do you think that these two things belong together rather than being in conflict?Jaythe technology is eliminating the the rote, right? So in software engineering there's rote in that somebody's implemented the connection to this API before or, you know, authentication through Google or, you know, whatever you're gonna do. it's been done and so why waste time doing it again? just because? I mean it essentially eliminates that work that's not terribly interesting, but has to be done and it's able to do it and so you can put systems in place that automate those pieces so that it eliminates the busy work that's associated with work. And what it leaves behind is the thing that it can't actually do, which is connecting to other people. being creative and solving problems in, you know, your unique way. I think having so then it places more emphasis on communication, your ability to talk to others, play nice with others, have creativity in your work. So you want people that have hobbies outside of work that are completely different that causes them to show up with a brand new perspective that others on the team don't have and then we get better as a result because the heterogeneity of the team increases and then, you know, as a collection we we do better. what's AI good at? It's good at doing all these rote menial tasks that we're not really interested in doing, but we had to do to get the work done, and now we don't have to do them anymore. So then what do we do? Well we spend more time thinking deeply about the problems in front of us and we spend more time talking to other people about those problems and helping them understand our solutions and how it works.Monisha SaldanhaFor WebScales e-commerce customers, what is the outcome you ultimately care about most from AI? Lower operating costs, higher conversion, better customer experience, or something else?JayThere's the core products and it's you know increased availability, better performance, those kinds of things. From our AI solution, it's matching customer intent to what's on offer, right? So so often so when we go into an an engagement with a customer, we have a conversation with them about what their success criteria is. What problem is it that we're really focused on and trying to solve? Sometimes it's a customer service problem, sometimes it's discoverability or a challenge with product search on the website. to be more flexible, communicate with the person more directly and how they communicate. Understanding what their requirements are and what they're looking to solve and then matching it to the solution that we offer So in all those cases, it's not so much you know one thing to say, yeah, we're gonna help you sell more. Obviously everybody is interested in that, even the customers that say, Yeah, I have this customer service problem, and I want you to help me fix that. So take the basic customer service problems off my my CSR team's plate and help them focus on the complex. That's a straightforward thing to do. but the next step is, and by the way, because you do all these other things and you understand the product set, help me sell more. So when that customer is showing up, you know, looking for a particular product, help them answer that question. and find match them to the product they were looking for. So I think invariably increasing conversions is the that's probably the number one driver. we actually get better when everyone gets better. So when the merchant's making more money, obviously you want to use our service.Monisha SaldanhaTake us five years forward. Does the traditional SaaS model, humans logging into applications and clicking through workflows, still exist in anything like its current form, or do agents increasingly become the interface to software?JayYeah, I think I think the kind of work is gonna dictate that. So you know, obviously I don't think we're ever gonna be clicking through somebody's interface and setting up configurations and things like that. I think we're gonna be talking to a conversational interface, describing what we wanna have accomplished and then a change being proposed and verified and saying, Yeah, go do it. I don't think we're gonna be clicking through workflows to to do things. That that doesn't make any sense. but I I do think SaaS still exists, it's just the interface changes. I wanna buy is through a conversation and I'm Curious about a topic and I'm investigating and learning, and then I want to say, great, how do I get that? How do I get started? What do I need? Where do I go? And then have the agent go procure the things that I need to get started. And that to me feels like agentic commerce versus and it's the same kind of thing in SAS, right? Is saying I come to the interface, say I have a problem, and this piece of software is responsible for that outcome. What's make misconfigured, and I'm talking to the system to get it corrected, versus you know, having to click through some UIs and find logs and understand what's there and make a change. I'm actually just having a conversation with the system to understand what it's doing, but the service that it's provided providing is still being provided, right? So I'm still getting benefit from some service, and a company is providing to that to me. So it seems like you know, Some folks say, Well, the SaaS- pocalypse. Well, nonsense, right? The interface is changing that works today given where I'm at and the position the business is in, but in the future, does it? If I'm ten times the size, am I gonna want to be doing that? Right, maybe not, right? Maybe we're actually evolving to where now I want to SaaS system to accomplish my next goals or outsource that integration work to somebody else because I'd rather focus on my core business as opposed to you know providing better feature function to my customer as opposed to building these things myself. It's just the work nature of work changes. So I think SaaS exists and the interface changes to be one that's more conversational. And most likely I have my front end and this MCP is interesting. I know it it's getting some. somewhat of bad rep, but being able to integrate or expose my SaaS service to the agent that I prefer to talk to, I think it's yet to be seen. Is it I bring my own and then I connect to an API via MCP to be able to execute the workload that I want to work or like answer questions, ask things, do things for me? Or does the SaaS provider give me a chat interface that I go talk to? It it kind of feels like the former, but I think both might might exist in the future.Monisha SaldanhaAnd last question, what is one book you recommend every builder read and why?JayWell I don't know about the one. there's the one I'm reading now. So shameless plug for Seth Godin's new book. It's actually pretty good. It's called The Knot. I'm I'm finding it fascinating.Monisha SaldanhaWhat's it about?Jaysolving problems. Ultimately, so there's a most people have an idea. He has a Jackson I think it's I think it's in there that he has a Jackson Pollock concept, right? Where Jackson Pollock, right, he in his lifetime is one of the few masters that becomes world renowned and recognized in his own lifetime. and he you know he creates the big things that you see a moment all that in so now so here's Jackson Ball he does his thing and then he goes off and then he has an epiphany and so he has this desire to go follow his artistic passion like go do the thing. I hope I get the story right. But essentially he has this thing he wants to continue to do his edge pushing, avant garde kind of his art that no one's ever done before. But at the same time, he also really enjoys the success that came in his lifetime because he's Jackson Pollock, you know. And so he said this, but I also want to have that, right? And so then one view of the outcome of his life is to say, well, he did follow his passion, but then he struggled because the world didn't come. The notoriety and fame didn't come along for the ride. Because they they the world wanted him to be Jackson Pollock, you know, the the big splatter art thing. They wanted that, not what he was doing now. So so the the notoriety. And fame didn't come along for the ride. And so you could say, well, yeah, he struggled with that. And the the it's kind of getting clear about what you actually want and how to how to solve that and be unburdened by the butt part of it. So get clear on it. Is it do you want to be famous or do you want to be this great artist? And so now you have this choice.Monisha SaldanhaWow, fascinating. Yeah, really good. Add it to my list.JayIt's a sugary.Monisha SaldanhaGreat. Well thank you so much, Jay.JayYeah, thanks, Manucia. It's been a pleasure.Monisha SaldanhaWhat stands out from Jay's story is that turning around a technology company isn't necessarily about arriving with all the answers. It can begin with asking better questions and listening to the people closest to the problems. WebScale's transformation also shows how quickly the economics of building a SaaS company are changing, when small teams can use increasingly capable AI to accomplish work that previously required far more resources. At the same time, the technology only becomes useful when it is connected to a clear customer problem, reliable data, and people who understand where human judgment still matters. Jay, thank you for giving us such a candid look at what it takes to rethink a company while you're still running it.JayYeah, happy to.Monisha SaldanhaAnd thanks to all our listeners. We'll see you next time.Lis