AI Adoption in Customer Success: What ServiceNow Is Actually Doing with Neil Sparkes

Speaker A: This podcast is brought to you by RevUp Academy. Hey CS Pros. I wanted to quickly jump in and let you know that if you're enjoying the podcast but wondering how you too can become a revenue focused customer success professional, then I have the answer for you. Some of you might not know this, but I run a cohort based coaching program that walks you through step by step on how to align customer success strategies with revenue. And it's called RevUp Academy. And RevUp Academy is my complete step, step by step coaching program that helps you elevate your skills and mindset to focus on driving revenue for both your customers and your company. And as this is a live coaching program, we actually only open doors for enrollment a few times a year and the next time we're opening doors will be in just a few weeks. We are going to open doors for the final time in 2026 in September. Which means this is your last chance for the year to be gain the confidence you need to drive renewals and upsells. And because you are a podcast listener, I am giving you all a discount for this being the very last cohort of 2026. So if you are interested in joining us and really leveling up your revenue skills within customer success, go ahead and join our VIP wait list for RevUp Academy. All you have to do is go to thecustomersuccesspro.com m forward/rev up and if you are on the VIP wait list you will be the first people notified when we open doors. But not only will you be the first person notified, we are actually giving a hundred dollar discount to my VIP wait listers for this very last cohort of 2026. So go ahead and sign up for our wait list at the Customer Success Pro RevUp. I hope to see you inside the Academy. Hello everyone, I'm your host Anika Zubair and welcome to the Customer Success Pro Podcast. Your go to space for real talk, expert advice and actionable insights in the world of customer success. I'm a CS executive leader, award winning strategist, CS coach and customer success fanatic. I help CSMs, UM and CS leaders build the skills and the confidence to become revenue driving pros and scale world class CS teams. So whether you're brand new to CS or a seasoned leader, this podcast is here to support your growth. Because customer success isn't a destination, it's a journey. And I'm here to be your guide and navigate every step of your journey. So join me every Wednesday where you'll get fresh CS Tips, tricks and strategies you can actually use. Some weeks I'll share my own insights and best practices from working in CS over the last 13 years and and once a month I'll bring on expert guests to dive into the most relevant and pressing topics in customer success today. So if you are ready to level up, hit, subscribe on Apple Podcasts, Spotify or wherever you tune in and let's make your CS journey a little bit easier together. Welcome back. Today I'm joined by Neil Sparks, the VP of Customer Success at ServiceNow. Neil and his team are right in the front edge of the biggest questions in our industry, which is how do you actually get customers to adopt your AI tools, not just talk about it? He has a global mandate to move agentic AI out of the experimental phase with his customers and into live production inside a large, heavily regulated enterprise businesses. And he has customer stories to prove it that it is actually working. And he has also just been interviewed by the Financial Times on this exact topic. So he is thinking about this from every single angle. Beyond the customer side, Neil has a strong opinion on what AI is doing to pave the way for customer success teams that really change the way they work day to day, which is where I want to take part in this conversation. So let's chat to Neil all about his experience with AI and getting customers to adopt AI as well. Welcome Neil to the podcast. It actually feels strange saying Neil, I should say welcome Sparky to the podcast. For everyone who doesn't know Neil. Neil goes by Sparky, but we'll say Neil. For the official purposes, I would love for you to give your own little intro into who you are for my listeners. Just what you're doing at ServiceNow and your role and the team and the remit you're running. Just give us a little bit of a inside scoop onto who is Neil.

Speaker B: Sure, thank you for having me. Um, so uh, I'm um, what they call the AVP of uh, Customer Excellence Group at ServiceNow for UK and Ireland. Long, long job title but essentially I look after everything post sale experience for our customers in the UK and Ireland. Business obviously ServiceNow, um, a big global multinational company. Um, but we look after around 700 enterprise customers and then um, some smaller commercial. Commercial customers in that segment. Uh, I hit my one year anniversary this week so survived my first year. Um, I guess I've got uh, it's going to make me feel old now talking about my career. 15 years I guess, uh, in SAS companies say um, Microsoft and the Access Group where I was uh, in similar customer success roles I guess at Access Group, Global customer success Lead, um, which we'll talk about today as well because there was some great uh, AI learning experiences and how you drive AI into customer success when I worked there, um, and Microsoft where we were building out the Azure business. So slightly different in there's uh, a cloud uh, um, platform as opposed to kind of SaaS, but um, still focused on driving adoption and value uh, in those customers. And then before that I was in financial services for 17 years so sat on the customer side buying all of these software products and uh, infrastructure, uh, and trying to get it working and delivering for internal and external customers. So on both sides of the uh, customer side and the sell t side.

Speaker A: Love that. I also love that you just have such a rich enterprise whether like you said it was truly SaaS or services, cloud based products, all of this. It just, it's nice to hear how AI is being adopted in such rigorous environments or such. Yeah, uh, environments where you have high levels of compliance I should say. And I think it'll be a very interesting discussion because like you said you've come from financial services but even your time at Microsoft and now at ServiceNow for the last year, it's going to be a very interesting conversation because a lot of the people I do talk to about AI are probably in more nimble AI first organizations. Whereas you yourself find yourself in an organization where the shift is very real, not just for your organization but for the customers you serve. So it's going to be very, very interesting. But before we jump into all of that, I'd love to know, what does customer excellence mean? I know what it means because I've worked with you and your team. But for a lot of people listening they're what is customer excellence? What is the UKI customer excellence team actually do? So tell us a little bit more about your team. What do you guys held accountable to? What is customer excellence at ServiceNow?

Speaker B: More excellent than success. So um, we obviously had to come up with a better name, a new word.

Speaker A: Yes, everyone loves rebranding success.

Speaker B: It is a bigger encompassing role than just the customer success part. So um, like I said, I'm responsible for everything post sale and that's a mixture of um, our customer success offerings. So we do have CSMs, customer success executives, um, platform architects and support account managers in those offerings. So I guess that's the more traditional customer success side, but also um, Expert Services which is our professional services delivery arm, um, and the kind of uh, post sales consulting. So uh, some of our paid for services but also I'm on the hook for uh all of our customers adoptions. Whether you choose to sign up to one of our paid for plans, whether you use our professional services, whether it's a partner delivering the solution. So we've got a massive partner ecosystem, around 15,000 partners. Um and if the customer was also deciding to do it themselves which we obviously love when customers want to skill up their own workforce and deliver our solutions. Um but it's everything whether it's delivered themselves through partner um in partnership with ServiceNow. Um but the one I guess outcome remains the same. So we have an off star metric of mps um which has obviously been around for a while. So you love or hate the metric?

Speaker A: Yeah, I was like that's gonna a lot of controversy in this convers already. There's so it's such marmite that that

Speaker B: metric we could do a whole hour on nps.

Speaker A: Yeah everyone, everyone has a very strong opinion on that metric.

Speaker B: Um but more uh our customers adopting and seeing value. So some of the more recent trends we've seen in, in customer success. So tying ourselves back to why why a customer bought uh the solution in the first place. And then also are they realizing and recognizing that value which I think is maybe even becoming a bit of a um ah ah an old school metric in customer success. I know a lot of people are still talking about it as the thing.

Speaker A: Um, I'm not sure it's come full circle. I think like again I, I know you've had a longer career than mine but still even my early career it's come. I feel like we're doing this full360 right now in post sales organizations where what we used to be accountable for in the 2000 and/2010s is now coming back again. We're almost coming back to it all over again. Which is interesting to see because we are in this AI first world but we're come we're coming back to some sort of basics when it comes to value delivery, value achieved, outcomes achieved with your customers which feels again early salesforce days. But it feels like that same time is happening all over again but like 20, 30 years later.

Speaker B: And I think even, even in our customers it's difficult to get to the right stakeholder um in the business which is really recognizing the value that say the board are talking about or what is needed in the P and L or the uh shareholder value. So you have to tailor your conversation all the time in a value conversation.

Speaker A: Yeah it's an ROI conversation now it's no longer just, I mean yes, it's value, but it's like you just said to the board, to the senior stakeholders, to the CFO or whoever's in charge of the P and L. It's now becoming much more relevant to be a financial metric as much as it is a value metric at this point. So it's very interesting to see, I

Speaker B: think throw AI in there and the new way of measuring uh, AI adoption and value and uh, the cost of

Speaker A: AI, which is, yes, that's another line item to add into this whole mess of how we're measuring value. Um, but listen, you came from financial services. I love talking to all the leaders that I, I do on this podcast and everyone has such a unique beginnings I guess of their career. What inspired you to start working in customer excellence or customer success or let's just say the post sales side of software businesses.

Speaker B: You're definitely going to make me feel old in this.

Speaker A: It's unintentional. You just have so much wisdom to share. That's what we're going to frame it as, is wisdom to share.

Speaker B: I'm an engineer by background so I did an engineering degree. Uh, I love hands on like doing stuff, I think, uh, bucket list kind of career. You know, people say if you weren't doing what you did, what would you do? Probably a builder or, or something practical. But um, I guess that's what made me.

Speaker A: Although that has to be the most AI proof job these days.

Speaker B: Exactly.

Speaker A: I feel like if you're a plumber or a builder, is AI going to really replace your job?

Speaker B: Exactly. Consider a career switch maybe in the next few years. Um, uh, I love, yeah I love building things and being hands on, say as an engineer. So I got into IT infrastructure because back in those days it was building data centers, trading floors, you know, kind of wiring and like the practical side of things. And I love problem solving, um, and kind of uh, finding solutions to things. So that's, I guess that's what makes me, me and my DNA.

Speaker A: Yeah. And so I think that makes someone great in any customer facing role. I think like the ability to find solutions and solve problems. Be curious. Uh, yeah. Whether that's actually building or in theory giving them an idea of what they can build. I still think that that's a core of what we do every single day with our customers and I guess that's

Speaker B: the side I love about m. My job still is the problem solving and finding ways and solutions to get around things. Um, but I guess I evolved through it, uh, technology, uh, repeating Projects. So obviously new technology was coming out. So virtualization. When that came out we moved from single servers into multi servers on one piece of hardware. Then cloud started to come out. Um, and I was heavily into exchange and messaging. I love the human side of technology and how it impacts people. Um, so I was always into email and messaging. Kind of teams started to come out and how you do video calls that really excited me. Um, I got to a point in some heavily regulated uh, places that I worked where I wanted to do more cloud stuff. I wanted to use Office 365 uh, Teams. But there was a, ah, massive reticence to bringing it in because of the security and regulatory side of the business. So um, I decided the best place that I was going to get more immersed in cloud was to join Microsoft. And I built my whole career on Microsoft technology. I knew the teams there. Uh, so I jumped over um, in my first year I decided to take a customer facing role um, and learn what it meant to be uh, on the streets if you like, uh, kind of in a, what a, what a

Speaker A: sweet from being in the engineering building side of things to your first role

Speaker B: at Microsoft being customer facing, huge teams as well. Like you know up to 300 people in, in some of my roles. So going back to being an individual contributor I really struggled with it on the, I missed the leadership side massively in my first year. So uh, I managed to find this uh, incubation business in Microsoft that was called uh, Azure. Um and I think we had like eight people in the team and we were out trying to get cloud adopted.

Speaker A: You make Azure sound so small within the world I guess maybe at the time but I'm just thinking you're like oh, this incubator. So one of Microsoft's biggest products.

Speaker B: We had targets in our first year that were less than they were in the millions, like single digit millions. And um, I did that for uh, around eight years. We grew that business and it was in the almost double digit billions when I left. So um, yeah it was amazing to be on that, that kind of transformation but I just wanted to do cloud new technology and uh, honestly back then, and I think this is important for now because I left uh, my financial services career and joined Microsoft in 2013 so 13 years ago, um, and I was concerned that my job was going to be eliminated as a technology infrastructure leader. Like what's the future of no servers, everyone's going to migrate the cloud. Um, and they have and it's been a massive adoption but that, that role's evolved and it, it's changed into something else. And you know, all my colleagues still happily work in that industry and they've, they've adapted and learned new skills. I think that's important because cloud was going to disrupt us and I took the bold decision in that my job was going to disappear. I think AI in the same transformation.

Speaker A: We're, we're going to get into a transformation right now actually, which I think we're very heavily in the middle of while recording. We're in the middle of 2026. Yeah, it's no longer about even trialing or testing. It's about true AI transformation within organizations. And you personally and your organization have a very aggressive target this year and a big transformation. And like you were just saying you were an Azure moving cloud transformation, now it's agentic transformation really. And like how are you getting your customers to actually start using the agentic side of your software? And I'd love to kind of just touch upon what it is you, you guys are doing. What does that look like from your customer space? Like what does that look like from a ServiceNow customer today? Is it chatbots? Is it adoption of like truly running agents within their business? What is, what does that look like and why does it matter now? Within, within ServiceNow?

Speaker B: Yeah, I think like, like a lot of um, software vendors we've built AI into our product offering. Um, a big play for ServiceNow is obviously enabling our ah, customers to leverage AI with the enterprise controls and frameworks that we would want um, them adopting to manage that. Uh, I think the challenge is nobody's really done it. So um, obviously people are adopting AI in many forms like uh, Copilot and it's appearing in everything we use, isn't it?

Speaker A: Even Spotify everything Spotify playlist recommendations, aidj. I mean every holiday is planned by Claude and ChatGPT at this point too. And I just think there's no part of my life that isn't infiltrated with AI at this point anymore. But it's interesting to see how major enterprises are trying to, to make it a part of like you said, the everyday use of your customer. Like co pilot is used and I

Speaker B: think you know, generative AI which is kind of, let's face it, it's like search on steroids. So I'm the same as you. I use it to plan my holiday destinations. What I'm going to do when I get there, like you know, kind of be, be my virtual tourist agent. Like that's, it's almost like uh, advisory and guidance like it's able to assimilate loads of information you would spend ages looking for and ask it natural language questions. I think the journey that we're on goes back to the value conversation. Like how do you actually make AI do stuff for you and actually like get it from being into doing so how do you stop it just kind of giving you advice and like assimilating that search, um, content into making decisions, being able to take workload, uh, take some of the um, uh, repeatable tasks and actually like add that decision making layer into it. So we've had automation around us for decades in various forms. I remember when I was in my infrastructure jobs, you know, automating stuff with scripts and. But it wasn't able to make decisions based on bigger context and bigger information. Like the bringing together of those two is what adds the value. Um, and I guess that's the journey we're trying to take our customers on. How do you confidently go into letting an agent, uh, loosen, um, in your enterprise environment and making decisions on your behalf and executing things without uh, the nervousness that it's going to go awry or do something that you're not aware of or you know, kind of operate outside of the control framework that you're.

Speaker A: Yeah. Just a quick pause from the podcast before we get back into our conversation with Neil. When renewal season ends up landing, what actually happens to your customer success team? Well, most CSMs are super smart and brilliant at the relationship part. They have the calls down, pack the support, they help keep people happy. But when it comes to sit across from an economic buyer and justify spend, they reach for a usage recap instead of a value story. And when nobody in the room can actually prove roi, the renewal turns into a discount negotiation and expansion is never, ever going to happen. Now I want to picture something different for your team. Your customer success team ends up walking into every renewal and every quarterly business review, able to connect what your product actually did to the outcomes that the buyer actually cares about and in their language so that the numbers in front of a CFO actually makes sense and the CFO is happy to sign off on that spend renewals end up stopping feeling like a fight and expansion becomes the obvious next step. Instead of uh, that awkward ask that your CSMs are currently doing and your GRR holds steady and your NRR ends up climbing and forecasting renewal gets a whole lot less stressful. And that's the exact work that I do. And it's the exact reason why Neil brought me in to train his father. Full UKI team at, uh, ServiceNow on ROI and value storytelling. So if you want the same for your team and you want your team to get better at ROI and value storytelling, then head over to the customersuccesspro.com team event. The link is down in the show notes. All right, let's get back into the episode. And I want to jump into agents for sure to second. But before we even get to like, how you build out zero human touch agent workflows for your customers, how did you even start to measure the value of using agentic? Because I think that's where probably a lot of companies are right now, where they're thinking, okay, AI saving us some time or AI is making us a little bit more efficient or better at, ah, what we're doing. But it's hard to really measure that and to really come up with a strong value story around how AI is making your product more efficient, faster, better. It's a claim. I know there's a lot of companies that are claiming it today and, and it is partially true. But to get large enterprise organizations, which is the customer you serve, to adopt, I'm guessing there's more than just saying AI makes you efficient. Right? Like there's a bit. There's a stronger claim or value story you guys are telling. So how are you getting customers to adopt with that value story?

Speaker B: I'll go back to my uh, email days again because I remember when mobile devices started to come out, so they were uh, iPacks with GPRS jackets and we got email on there. There's obviously a huge cost associated with that. And I remember once presenting to the board on um, the values of it and kind of how it would make our uh, teams more efficient, they would be mobile, they could answer questions everywhere. So you even, I think before BlackBerry or around when BlackBerry was around, um, and when you're trying to just think

Speaker A: of emails and BlackBerry in my first BlackBerry to how exciting it was to have email on a black. Yeah, like it was so cool to

Speaker B: have your BlackBerry, but as an infrastructure guy that was always trying to sell to the board the benefits of those things. Like it's very hard to turn efficiency gains into hard metrics because probably like most of us, if, unless you're willing to um, reduce your workforce or um, remove another piece of software out of the balance sheet, like most finance leads or boards are not interested in the efficiency gains. I mean they are, but it's kind of a given to. I think as businesses have come under more cost pressure, they need to demonstrate where if you're putting AI in, what is it you're taking out?

Speaker A: Um, what line item are you replacing it with essentially? Or what are you, what cost associated with that or what profits?

Speaker B: So efficiency doesn't cut it anymore because you can sit there and say I'll make the workforce more efficient, um, and most CFO will turn around and go, does that mean you're going to, going to get rid of 20% of your team? Likes horrible to say.

Speaker A: It is horrible to say, but it is the reality of uh, software and technology right now. And I think people can relate because it has happened within probably most of the organizations we've been at.

Speaker B: And I think the challenge is that most people turn around and go well I'm going to make my, my team 20% more efficient. How do you actually demonstrate that they're doing 20% more work? Um, and I guess that's where because of the history of ServiceNow, um, we've got, we've been running enterprise workflows for years for customers. Like we've got a lot of uh, information that sits in their system that they can demonstrate how long a workflow's taken or how long it takes to get from A to B and the number of touch points through it. Um, it's actually quite easy as we put AI agents on top of that to demonstrate that kind of this is what you had and this is where you are uh, even to the point where we can bring in like technical debt and some of the systems that we're able to um, uh, interface with. Because one of our big selling points is that we interface with any cloud, any system like Workday or ah, Oracle, whatever your ERP is, your finance billing system. Like we can talk to everything, um, and any AI agent on top because remember, not one AI agent, it's always a mouthful, uh, is going to do everything. So you're going to have Copilot, you're going to have Claude, you're going to have uh, now assist, you're going to have move work. You know all of the, there's thousands of agents out there now who's going to get those agents talking to one another? So um, I think the, the interesting way to think about an AI agent is as a person in your organization. The difference is you can hire a thousand of those people in an hour, uh, and get them all stood up and trained and knowing what to do. The reality is they need access to systems, they need access to data. Um, they still need to go through an onboarding process of some description, talk to all the Other, uh, employees, AI agents in the organization, how do they do that? So if you think about it, it gets very complex, very Christmas, uh, very quickly. So the value of being able to stand up a thousand agents where you couldn't hire a thousand people that quickly is obviously a huge thing. How do you manage it, orchestrate it? And again, like the same with people. How do you make sure one of those employees has not gone rogue and is doing something that they shouldn't do, or accessing a system that they shouldn't, or not following a corporate process? That's the challenge with Agentic. And what we're, um, tackling really is that we'll bring that control framework around. It is the same as it would be if it was a person sat at a desk.

Speaker A: Yeah. And I think that's what a lot of companies have to realize. I always say this as well. Like every agent I've ever built or trained or anything, it's like a junior employee. You can definitely enable them, you can give them the right frameworks, uh, like you said, the compartment in which they are supposed to do the work. But just like any new employee, there is chances of human error. There are chances of AI error as well. And I think putting up the guardrails, making sure you're checking in exactly what they are deploying and how that looks and what that data is and how they're executing against that, is just like hiring tens of thousands of new employees. You have to make sure they're trained up and they have the ability to execute on what you want them to execute.

Speaker B: On.

Speaker A: Which I think sometimes people forget when it comes to deploying a lot of zero human touch processes or frameworks or whatever they're building. So, um, I love that you also highlighted that it's financial gains, because I think a lot of people keep looking at AI as efficiency gains. And yes, it is. It's efficiency gains. Step one. But to really get the spend that you need to get and the adoption that you need to get on your customer level, you need to translate it to, not only is it going to make your team efficient, but it's going to help you grow your profit margins by X or it's going to help you increase revenue by Y or whatever the claim is. But I do think that we keep telling our customers you'll be more efficient. But what does that actually translate to? Which is, which is, like you just said, it needs to be one layer deeper that makes sense to, to the CFO or the board or whoever it is that that's Reporting up to. And um, I think they have a

Speaker B: low tolerance for that now. Like, you know, they don't have money slosh brown to spend.

Speaker A: Yeah, there's. But there's no more time or money. It just. We're in a tough economy, a tough world right now where it's like, we don't have the leeway to be like, let's just spend tens of thousands, hundreds of thousands on this experiment. It's just, it's not going to work.

Speaker B: Everything is under scrutiny at the So I don't think I finished answering your question. I guess the. So that, that journey that we've been on from traditional workflow into agentec means that we can actually benchmark with our customers from how long and how much it costs them to do it before. And you can enter some of that information in and then we can benchmark how long it takes now so you can demonstrate, uh, the real cost savings.

Speaker A: Um, I love that, I love that like we have to benchmark, but I love that you're just asking because I think sometimes we get nervous to be like, oh, is this right? Is this wrong? Is this resulting in what you expecting it to get? And so sometimes we just need to ask our customers like, hey, this is new for both of us. Where do you stand today, where do you stand tomorrow? What's it look like a week from now? And how can we make sure we continue to grow efficiently for your business? Um, but I love that I've sat

Speaker B: in a lot of customer presentations where we've been like, you know, here's the hundreds of millions you're going to save with, with this technology and not service now all over the years, um, and you almost get laughed out of the room because then they're not, uh, they're not validated by anyone on the customer side or, you know, it's not.

Speaker A: Yeah, there's no bad, there's no real like check in of like, hey, this is where the reality is of where the hundreds of millions you're going to save. We're actually at this marker. Yes, we might save hundreds of millions in 10 years, but right now we're at 50 million or whatever the number is

Speaker B: a customer to think, you know, if someone shows up and shows you a saving 100 and even if you could say a 10 for that, then that's still a substantial number. Um, so I think sometimes all of us do it. We can get a bit too bogged down in the data and arguing that it's, you know, it doesn't take 10 minutes it takes five. I still have 50% saving. So I think as long as it's, as long as it's directionally telling you the right thing, I think people need to take more of a leap of faith uh, and then start. Otherwise you can spend months or years talking about the cost savings and not doing it when in reality you're getting left behind and not making progress on it 100%.

Speaker A: You can sit there and theorize on it, you can sit and create as many hypotheses as you want. We can be like yes, Maybe there's like 100 million saving or this is what we can do. But if you don't actually start and get going and start pulling the trigger on some of these things you're going to just be, especially right now, we're going to be left so far behind. Um, but I want to come back to boundaries on your AI agents because I think that's really, really important because I think some people are mostly doing automations with AI. I feel like that's probably where most of companies are. But truly deploying agents that are zero Human Touch is getting started but maybe not adopted as quickly as probably most companies would Like. How did you guys get started on Zero Human Touch? Full AI agentic workflows where, where you had them just again go out and do those things. And where did you kind of draw the boundary here between human and agentic agents doing the work?

Speaker B: I think you've got to start where you're uh going to say confident but where it's, it's low risk and you've got the confidence is not going to do too much damage. Yeah, want of a better term. Um, so finding a low, I'll call it a low value workload but I mean kind of risk and um, uh, uh governance kind of position. Like if you can find something and then start there and build the confidence then I think that's where we're seeing the most traction. Where you know, kind of say triage and incident ticket like ServiceNow's heritage is very much an ITSM. Um and kind of incident triage is a great one where typically it would come into a help desk agent and they would look at it, read the context and then figure out which team it needs to go to. Um, obviously add in extra context. If they started getting different uh, like a multi stream of the same incidents then it would be able to pick it up and go hang on a minute, there's something else going on here or raise a problem record it's probably and they can Go in and you know, look across the infrastructure and find out that the common part is this, I don't know, network switch causing the issue as an example. Like it would be able to do that analysis, talk to other agents as well if it needed. So you might have an agent which is doing traditional monitoring of your IT environment. It could talk to that agent and go and um, I'm getting all these tickets. That agent could go, well I'm seeing this issue. They would almost have that conversation and then figure out what they're going to do. So it's again like think of it in human terms. Like if I saw that coming in and I was a help desk agent, I'd pick up the, the telephone to the network team. Is something going on your side? Yeah, we've got a problem and you connect the dot. Like all of that can happen in milliseconds, you know, kind of agents doing it and at scale. So I think like incident triage is a great place to start because it's again it's like um, it's a sign,

Speaker A: a low hanging fruit I guess. Or yeah, I don't know, I'm thinking of a better word too. But it is, it's low uh, like I guess triage incident, like if you, it's not going to cause major outage issues for your customers, but it's going probably solve something that has a friction point where an agent can do it quite quickly rather than waiting for a human.

Speaker B: Yeah.

Speaker A: Do you keep human channels open through that or how does like I guess in that loop of like again let's say triage, the, the agent takes care of it. Where's the human loop, if any. And do customers ever get frustrated with agentic responses rather than human response?

Speaker B: Again, I think it depends on the risk around it. So if you're comfortable that an agent can make the decision and you've got those guardrails in place, then let them get HM on with it. And I think you can start by asking say uh, I've done X, Y and Z and this is what I think we should do. Or there's three options that we could do and you present that back to a human to make the decision, the ultimate decision. So it's still done all the generative work but it's asking you for confirmation before it would, you know.

Speaker A: Got it. And when does it decide like confirmation. Like when does it like decide to just execute, let's say as an agent versus get the.

Speaker B: Yeah, I think that's why you need to build those guardrails in, like you can execute up to this point, um, when you go beyond that and then you need to check for confirmation. So it's a bit like, uh, feel bad saying it. Like if you had a whole load of apprentices, uh, starting that just come from school, you said earlier, like, they, they almost need training up and telling them, like, how far they could go. So it's no different to a human. I think the difference, it's really not.

Speaker A: I, I joke every single time that something that I've built comes back and it's just like 80% there. I'm like, no, I didn't say exactly this to you. I realize I should have probably been even clearer because you don't know what I'm thinking and I have to give even more strict guardrails to exactly what I want you to run and execute on. And, um, like you just said, it's like they don't know any better. They only know exactly what you've told it and you've got to kind of revise that as well. So.

Speaker B: And I think the very thing anyone who's using, um, you know, kind of the, the natural language, um, systems in, in angular, I use Claude every day. I use Chat GPT at home. Um, and I've weirdly, purposely made the split between personal life and work life.

Speaker A: I think everyone has, by the way. I think, I think when Claude rolled out, it's like m. I don't even know anymore. Like the million things they can do from cowork to code to design to everything. It's rolled out. I think in the last quarter, not even probably the last 6 weeks or 12 weeks or something like that, everyone suddenly made this shift between ChatGPT is like my personal agent and Claude is like my everyone I talk to. This seems to be the trend right now where we're separating our AI, our LLMs, I should say, into personal versus work.

Speaker B: Well, I think, I think you've seen how powerful it gets on the things that is learning about you as well. Like, I was out on Saturday with a friend and we were, we were kind of saying, like he was saying that he feels like he wants to erase, um, his history and start again with ChatGPT. We were asking it like, how, how long have you been building up, uh, knowledge about me? And it was going back to the beginning of 2023. So.

Speaker A: Yeah, but it's crazy if you think about, like the, the history it knows about you, whether you use it personally or professionally. I'm like, wow, there's so much I Remember when I moved ChatGPT memory to Claude? When I started using Claude, I was like, like, it's crazy how much these LLMs know about every little thing. And again, what they execute based on what you've told it to execute versus not and what you should or should not do. So it's just interesting to see how far we've come in just a few years.

Speaker B: 100%. And ChatGPT is doing my gym program, a golf swing where I'm going on holiday. To your point, like what, how I make a cocktail. I don't want work knowing all of that stuff.

Speaker A: It's funny, it feels like the differentiator of like, it reminds me of like when you had other social media versus LinkedIn. Like LinkedIn is what you can safely post, not work, and then what you put on Facebook or other social media is what you don't want LinkedIn to see. I love that. I love that. Okay, so you have agents that have, have guardrails and your, your clients are opting in to use said agents versus humans. And you have zero touch as well for some of these, which is great. But let's talk about rollout. Like how do you get customers actually into using these agentic agents, AI first rather than again, humans, because you work in an industry and your customers heavily regulated large, large enterprises that have global footprint. I'm just thinking the change management of that is probably very, very tricky. So I'm curious, how are you incentivizing your team? What's worked, what hasn't worked? How are you actually getting customers to adopt? Because I think that's what a lot of software companies are struggling, especially if they've been around and now they're moving to AI first. Getting customers to change is truly the hardest part of all of this right now.

Speaker B: And I think with AI is you end up in these interesting conversations around AI can help you transform your business processes. Um, but also you've got to start using AI to help it improve the process. Like what do you do? Do you fix the process first or do you just give it to AI and let it improve it as it goes? So, um, I haven't got the answer to it.

Speaker A: That's such a chicken and egg question. I just feel like that is such a do I just get going or do I have AI fix it or do I fix it? Who knows?

Speaker B: There's an argument for like giving it the process and um, you know, AI is going to naturally figure out an efficient, more efficient way of doing it. Like if it's been Running it for a month. You can probably just go and say like, what can I do to speed this up by 20%? And it'll, it'll recommend, you know, a million ways.

Speaker A: It will. You could genuinely ask it. I ask it all the time. I'm like, where are the gaps here in this analysis? Where should I fix for next time? You know? And it does, it gives you a list of recommendations.

Speaker B: Yeah. The other argument is that you give it a broken process and that's a broken process on steroids. Like it's going to repeat it and run it, um, you know, more times than you could in any human. So there's an argument for broken data and broken processes are just going to accelerate the mess. So I don't know, like, again, it'll be interesting to see, um, how um, companies work out which is the best route for them. But again, like not using it at all is, is just getting you left behind. I think that's what I'm seeing. And we call it an AI divide. I've seen more and more research that the AI divide is growing in companies as well. The adopters and the fast adopters are accelerating, uh, a way which you would expect with a, ah, platform or a capability like AI. Because to the, what we were just saying, it's going to make it better and going to improve. Um, they've obviously accepted some risks to get on that journey quicker. Um, but it's like any innovation, you know, the innovators are going to innovate and they will innovate quicker. But AI is supercharging that the people and the laggards that are not adopting it and not transforming their businesses, m. Are getting left behind faster than ever before. So.

Speaker A: And do you feel like they're, they're, they're. Even though they're being left behind, they're now quick to come back because they're watching themselves get left behind. Or do you think that these customers are just. We'll figure it out later.

Speaker B: I think they don't even see it. I think they're not aware how much they're getting left behind as well.

Speaker A: But I think it's a great time to incentivize your customers because of the bubble that technology is in right now. Everyone feels, I mean, I know personally if you work in tech, you're feeling left behind by not being AI curious by not building something new with AI without vibe coding out the weekend, whatever it is. There's so much use of AI outside of our jobs that we feel the need to really integrate it into our lives, but it's in our work lives, I should say. But it's interesting that you have customers that are definitely just so unaware of

Speaker B: this because I think the cloud journey did a lot of damage to customers as well. Like in the um, cloud people were giving it to their developers and they were starting to develop on it. And we're seeing it a bit with AI tokens. It's following a similar path.

Speaker A: So do you think tokens are working well for adoption? How are you, I guess, making the transformation?

Speaker B: I think what I was going to say is when we released cloud, people had their developments, they could spin up development environments in seconds, in minutes, they were able to do stuff they'd never done. And then companies saw their commitments burned down quicker than they expected or cost rise before they could measure the return on investment. The same thing is happening with AI tokens. Anything you see in the press, especially over the last few weeks, is tokens are out of control. People are burning through tokens and then what's the return on investment on those tokens again? Come back to the value conversation. Say CFOs are starting to see these big AI bills land on the desk and they're starting to kind of go, what is the ROI of this million pound I'm spending with um, uh, a natural language model. I think that's the dichotomy. And what we tend to see in our customers is they don't, they don't. They commit to um, a lot of AI and not out because they're trying to hang on to the cost control.

Speaker A: Yeah, I was just about to ask like, what's. It's one of the hottest topics right now in customer success is people are ripping out their tech stacks, throwing away their customer success tooling, trying to build their own thing with Copilot or Claude or a, uh, data lake or all of the above, basically. And this is again coming back to what you said 20 years ago, we're just kind of building on our own, which AI is just enabling us to do much faster. But it's a problem because there's, there's maintenance, there's support of whatever you're building. And for those of you who have been around for perpetual software, like you can, yeah, sure, build a data lake, do all these things, but who's going to maintain all of this for you over time? And I think people are underestimating the like, it, like regulatory risk, the overhead risk, just risk in general. So is that something that you're talking to your customers about with ripping out certain tools and just putting AI in or using again, ServiceNow with AI components. How are you navigating customers doing that?

Speaker B: I mean a super. So there's obviously this SaaS apocalypse. Um, yes. Uh, because who's going to need a SaaS provider, uh, when you can just ask AI to build it? To your point, maybe this is a, uh, naive opinion from me, but, um, all of the big software providers are using AI to develop their software as well. So like, we haven't missed that boat and just kind of all our customers are using it and we're not using it in our own development teams and stuff. So in theory our product's gonna um, develop quicker and we're gonna bring more into it. But more than that, it's built on M, you know, tens of years of experience of running this type of platform to your point of regulatory knowledge, how you maintain it. Um, so I kind of don't buy the narrative that SaaS is dead because they'll always. It's a group of people that have got experience and they're using AI to build that. Now, M, again, naive because who knows what it will look like in three.

Speaker A: Yeah, I think it's a tool. I think much like everything else, AI is a tool. Will it kill all of SaaS? Who knows? No one can really tell the future, but at the same time it's evolution. Like, I think we're, we're evolving. I don't think it's going to rip out SaaS completely, but I think there's a risk for sure.

Speaker B: And I think the uh, I can't remember the second part of your question, but I think it was like kind of how we're going the teams. Like it is about going out and you know, finding the use cases that are uh, low risk, high activity. How do we got this narrative with my team and I say you can use AI to supercharge your job. Don't let I replace you. And like the job roles that we do will accelerate because we're using AI to help us do more and do it quicker. As a leader though, I'm starting to see where AI is, uh, damaging people as well in a way. Like for me, I have loads of people send me, um, more information than ever. I was talking to someone this afternoon about, I think over the last decade you've had to consume more and more information not only as a leader, but anyone in a corporate environment is coming along alongside email. Now you have two platforms to manage or um, Slack or any of the other platforms. And then now that AI is there. All of that information can be produced quicker. What I'm tending to see is uh, AI laziness or a um, lack of ah, marking your own homework.

Speaker A: Yeah your brain as I like to

Speaker B: say it in, in anger. I was chatting to again my friend on Saturday. Like I'm starting to get actually quite frustrated with um, some of my AI tools that are building stuff for me because it's like they, I asked them to do something and it's like I'm being ignored now. I feel like it's got worse in the last few weeks. I don't know why, nothing. But it's, it's not outputting it as I want and it's like, I think because I'm conscious as well. There was this thing around, was it 20% of compute capacity and AI is wasted on pleases and thank yous and

Speaker A: um, all the rest, all the niceties that we treat it like it's uh,

Speaker B: like human doing my environmental bit by not being too polite to it, just

Speaker A: being direct, just ripping up the bad day. Just do what I'm asking you to do basically.

Speaker B: Um, so there's this worry for me that we're, we're going to drown each other in information because my team is sending me stuff with AI. I'm then using AI to simplify it and summarize it it for me because I've got too much. Yeah. I'm then talking to it like I wouldn't talk to a person because I don't need to and yeah be my bit. Um, but then I think that the boundaries will blur at some point and if we're not careful I think one, there's still a human at the end of it. To your question earlier, like where does the human sit? Someone still needs to consume all that information and I'm seeing the level of context switching and um, uh, amount of information consumption that someone needs to cope with kind of exponentially go up in the last few months. Um, and it's like it's exhausting for the human brain to be able to keep up with that. Uh, so I think that's risk number one. If we don't build agentic workflows correctly to do the work and we still expect humans on the end of it to do it all, that's not going to be sustainable because it's like again it's like me hiring a thousand people in my team and they all come to me on day one kind of going what do you want me to do then? And I say go and do this. And then they come back with a question like, it is still going to be difficult.

Speaker A: It's too much. No human can hire a thousand people and expect to ramp them up all at once. Like, how would you do the same thing with again, an entry level AI bot or agent or whatever.

Speaker B: Socially, if we start talking to each other like we're talking to AI, which is a risk, like, I don't know, I think there's bigger cultural and um, uh, well, being pieces that are not being considered.

Speaker A: There is, but there's a huge shift to everyone just sending more like whatever happened to just like two or three sentence, get the point across, communicate efficiently. It's AIs made this again, like you said, content or I don't know what the word is, but it's making conversation cheap or information cheap. And it's then causing us to send paragraphs instead of three or four sentences that get the same thing across. Like no one needs to sit through 500 slides in a deck. We could probably get it in five.

Speaker B: Um, and we'll all be on the same holiday locations eating at the same restaurants.

Speaker A: Yeah, because AI is going to keep dictating. That's exactly where we're going to go after these 500 slides that it's put together for us. That's going to. But there's a line that needs to be drawn and there needs to be almost again the lack of outsourcing our brains, which I get it, it is accelerating how we're working. It's making us really push the needle forward in technology, in customer success. But there is, yeah, there is risk, uh, at every corner if you don't, if you just let it run rampant and then just outsource your brain, as I like to say, or AI laziness, as you've pointed out too. Um, but I'm curious, with all this AI agents adoption from your customers, I'm curious, how is this AI product layer that you're building, how does this affect kind of your renewal and upsell conversations? Because I think now more than ever customers are expecting higher roi, higher outcomes output. That's the conversation that's happening. Is this kind of shifting your value conversation because of AI, or how is it kind of either helping or hurting the renewal and upsell conversation now that

Speaker B: you have AI, we've built AI capabilities into all of our um, uh, product skus now. So, uh, you get AI by default when you're purchasing anything. Um, I think that's a desire for us to embed that AI piece into our traditional um, platform.

Speaker A: Mhm.

Speaker B: And I use the word force. Encourage our customers to, um, obviously start leveraging that and thinking about that in an enterprise context. Um, I think the way we incentivize our teams is massively changing as well. So adoption's always been a thing. If I think about my customer success teams, they're heavily driven on adoption and things like nps where we can see if customers are satisfied with what we're doing. But, um, more importantly than ever, AI adoption is the race. We need to make sure that the teams are focused on, um, the value conversations, but also finding the use cases. I think that's where it's shifting. Like how do you find the use case, which is low risk, high value. Can remove the. And I, and I do mean remove, remove the workload from people. Um, and not just be more efficient or generate more content. Um, I don't. Even internally, I'm saying it with like, we're producing a lot of agents that make my teams more efficient. Um, I need them doing, like, we need them doing the work. And I think that's where we're starting to get to. Um, but it's a journey again. It comes back to any software rollout I would have done through my career. You start where you feel safe and then as you build confidence that your team skills get there and your, uh, customers are adopting it and ready for it and you're not causing a problem, then you roll it out. Um, but ServiceNow have really focused on that control framework around it. Again, it's our heritage on control compliance, security. How do you know these agents are doing what they said they should do or what you told them they should do? Um, and we've talked about it a few times in your chat box. Again, when you're using these language models, you sometimes have to encourage them or tell them they're wrong. I think we've all done one gone. That's not right. And they kind of go, oh, yeah, you got me. Uh, sorry about that. Didn't mean to get it wrong. Like in a corporate environment, not acceptable. Like you can't have, um, someone who's making an ambiguous decision. It's got to be right. So I think that's the interesting shift that we're on is how do you make sure the agents are doing what you told them to do and not something that you shouldn't and you do. I still think it's like a human. You don't want to stop the creativity if you want. You want them to be able to do what they need to do, to assimilate information and make a decision based on that. Um, but there's got to be boundaries.

Speaker A: And that's what I think that's the name of the game here. I think that's like with anything that anyone's rolling out, whether it's customer facing or internal, there has to be boundaries and you have to, like, there has to be checks and balances. You have to check its work just like you would check any other human intern, new person working within your organization. So, and I love that you brought up Value Conversations as well, because the last time you and I worked together, I came in to run a value conversation workshop with your team. So from your side as a leader, what made you make that call? Like, what was the gap you were trying to close within your team? What made you decide to bring in external training on value conversations rather than handling it internally? What, uh, made that happen for you?

Speaker B: There's always a power of someone, a new face or someone external coming in and saying something. We could say it a million times internally.

Speaker A: But yes, I've heard that from many leaders I've worked with. It's nice to hear an external voice.

Speaker B: An external viewpoint always brings in a reality check. And it's good for us as well to just make sure that we are on the right path of what we're doing. Because, um, I've got a load of experienced people in my team. I've been lucky to be hiring, um, for the last year. But, um, we don't know everything, so it's always good to get an external viewpoint. I think, um, the other thing again, the value conversation is changing. We're lucky that we we've got a strong value framework, um, in particular in ServiceNow, like pretty stronger than most of the other organizations that I've worked on. We're building it into the product. Um, so we're using AI to demonstrate value as well. Um, and again, I always think about this in my personal life. We mentioned Spotify earlier. Like Spotify Wrapped for me is great example, one of the best, uh, digital value frameworks where if you ever thought you weren't using Spotify at the end of the year, and it shows you how many hundreds of thousands of hours you've wasted listening to artists and who your favorite was always embarrassing to see who your Spotify racked.

Speaker A: Yeah. But it's also just reaffirming that whatever I'm paying is worth it because I

Speaker B: am using it for another year.

Speaker A: Yeah, exactly.

Speaker B: I think that's the biggest challenge for any software provider. And again, I've seen it in every organization I've worked in is how do you bring value into the product? Um, and you see some of these consumer products do it brilliantly. Um, but it's obviously a lot more simple for them than it is in,

Speaker A: ah, an enterprise or a major enterprise organization. Yeah.

Speaker B: Um, we've got to get there. Like, it's got to be in your face at every turn.

Speaker A: Yeah. And it's always nice to learn new ways of doing that. Exactly. And just making sure that you stay sharp on it, whether with AI or without. So I love that. I loved working with your team. Um, but I want to wrap us up with our quick fire questions, where I challenge each one of my guests to answer each one of these final questions in a sentence or less. Are you ready for a challenge? Yeah. Okay. You can. You can ask the language model to answer these for you, but it has to be a sentence or less. Are you ready? Okay.

Speaker B: Go for it.

Speaker A: Okay, let's do it. If you could predict the future, what do you think customer success will focus on next year?

Speaker B: Uh, net new revenue.

Speaker A: Net new revenue. Interesting. I've not heard that. Okay, we can't dive into that. Could be a whole different podcast. Next question is, which app or software do you use every single day or every single week? It could be on your phone, laptop, doesn't matter.

Speaker B: Claude and chatgpt Low we talked about.

Speaker A: Yeah, I feel like that's the number one answer these days for that question. Okay, and next question is, if you could change one thing about customer success, what would you change?

Speaker B: Uh, if I could wave a magic wand, I'd get our customers to understand what we're there to do.

Speaker A: I feel like that could be the mandate over the last 20 years in customer success. Okay, next question. AI very topical. But what is one cool or unique way that you've used AI today?

Speaker B: Today? It's been down for half of today, actually. So, uh, it's made me realize how, uh, I'm never going to cope if it disappears again. I'm using it to analyze, uh, AI adoption across our customer base and what we need to do to drive that up.

Speaker A: Ooh. M. Like analysis. Like a gap analysis. I love that. Okay, and then my final question is, who should be my next podcast guest?

Speaker B: Oh, podcast guest. I've seen some of your, um, podcasts.

Speaker A: Ah,

Speaker B: I don't know. I'd have to come back to you on that one. I know you speak to Marco all the time. Marco's okay.

Speaker A: I can get Marco on here. Amazing.

Speaker B: Well, thank you, Marco. What's going on at LinkedIn?

Speaker A: Yeah, exactly. With Marco. You got to hit me up now because Sparky said that you've got to be my next guest. Um, but thank you so much, Neil, for your time, your energy, sharing what's working and what's not at ServiceNow. I love this conversation. If my listeners have more questions or want to get in touch, what's the best way to get in touch?

Speaker B: I've already got me a LinkedIn because, uh, as much as AI is, uh, assimilating my inbox and my teams, uh, actually LinkedIn, probably the lowest. William.

Speaker A: Uh, let's all flood marking's LinkedIn, but I will make sure to put Neil's LinkedIn down below. Thank you again, Neil, for this conversation. I really appreciate it and a big

Speaker B: thank you for having me. But also the work that you, you referenced that you did with the teams, like, they're still talking about how they're having these value conversations. So thank you for everything you're doing for us as well.

Speaker A: Amazing. Thanks. Thanks for taking. Thanks for tuning in to the Customer Success Pro podcast. I hope you picked up something valuable to take back to your team. If you enjoyed this episode, it would mean the world to me if you took just 10 seconds to leave a review on Apple or Spotify. It helps more CS pros like yourself discover the show. And creating new episodes takes a lot of work. So leaving a nice review keeps me motivated to keep creating. And don't forget to hit subscribe on Apple, Spotify, YouTube, or wherever you listen to podcast episodes. I drop a new episode every Wednesday packed with practical tips. And if you've got a topic you'd love for me to cover or want to be a guest on my show, send me a message. All the details are in the show notes. I'd love to hear from you. And hey, if this episode helped you share it with a fellow CSM or CS leader, remember, sharing is caring. Cheers to your CS journey and I'll catch you next week for our next episode.