Falling in Love with the Problem: How Leaders Navigate AI

Victoria Mensch: Hello and welcome to Decisions at the Frontier, by Silicon Valley Executive Academy. Today we have a fantastic guest: Rani Pangam, currently a VP at ServiceNow. We'll get into the points our listeners are really interested in — innovation culture, measurable enterprise value, bridging business and technology. So I'm very excited about this conversation and learning from you, Rani. Welcome.

Rani Pangam: Thank you so much, Victoria. Very humbled and honored to be here.

Victoria Mensch: Transformation is definitely something you live and breathe every day, and you've been doing this throughout your career as well. What are you seeing right now from the transformational mandate that companies are going through? What is at the top of the list of priorities for you in your work?

Rani Pangam: Great question — and like you said, a lot of leaders listening would have been part of transformations for most of their careers. This one is unprecedented. I know it's going to sound cliché because everybody says it, but the pace of change is truly unprecedented. Right before we even get to actually test out and understand the latest changes in AI, something new is already on the horizon and it's launched. So one is the unprecedented pace at which it is changing, which makes adoption that much more challenging to drive. A big thing we're seeing right now: in 2025 there was a lot of piloting going on, and the initial focus was on taking productivity out, making people more efficient. We're seeing those conversations pivot, where boards are asking the C-suite, 'What's the new business model you're going to create with this, so that we can actually improve our top line?' How are we driving revenue, not just driving productivity — which is now an assumption? From productivity to yield is what I would say is a big change. And we've got to scale. You can't just pilot, you can't just experiment. There's a lot of scaling that is being expected — and all of this while being safe, while driving governance, because you cannot just put in untested models and put them in front of your customers. You've got to make sure there are enough guardrails around it. That's a big part of the shift we're seeing. Even when I talk to customers — and I see this in my company as well — another big change: data has never been that important in the history of companies. Everybody knew you need a good master data strategy, a unified data strategy, but the ability to bring together that fractured data, cut down silos, is even more prominent now with the agentic work, because the agents need access to that data seamlessly, to bring the work together and avoid any friction. And I think time to value — because AI is so much exponentially faster than any prior technologies, customers expect that time to value to be that much faster. Although the ChatGPTs of the world started in the consumer space, somehow companies have started to have those expectations as well: 'We could do it that quickly in ChatGPT — why are we not able to do it within the company?' So time-to-value expectations have gone higher as well.

Victoria Mensch: A lot of what you're talking about is definitely on the minds of the executives we talk to and the ones going through the academy. One thing I wanted to focus a little further on is really that switch from piloting to deployment. AI is a big topic, it's top of mind, and there's a lot of pressure on both sides on how to implement it and how to use it. But generally, when we're talking about emerging technologies and implementing them within the organization as part of the transformation strategy, that jump from a pilot to deployment and adoption is usually a big challenge within the organization.

Rani Pangam: 100%. And I think scaling is very key — scaling without losing the hunger for innovation. Because it's changing so fast, you also need to be constantly on your toes, innovating and understanding what's coming very quickly at you, and then making it robust enough that it's ready to scale. A key part of that is really understanding how the innovation teams you would establish in your company operate: what is that small yet nimble innovation team that does the bottom-up, grassroots-level, scrappy, quick pilots to even understand if the technology is relevant — because what you don't want to do is waste time on useless stuff. But that team is interdisciplinary enough that they have the context of different business needs: how would this work in sales? How would this work in finance? How would this work in product? How do you build together that SWAT team that can quickly disposition and quickly build innovation into their sprints, to bring that technology into usable, valid use cases? And then based on that, every two weeks, at the end of the sprint, say: 'We believe this is the right technology to bring in and scale in XYZ — let's partner up with the team that actually scales it, to bring it across the enterprise.' These models existed before; what I think now is that these models of SWAT teams — these innovation teams — are becoming a necessity, and having these test labs be a necessity, rather than something on the side or a fun thing. Now it's absolutely a necessity to have those innovation teams, have them be super nimble and agile, and be able to disposition what that new technology would even mean for the enterprise.

Victoria Mensch: What I'm hearing is that this SWAT team is kind of the dedicated team running those pilots and experiments. Is that a fair statement?

Rani Pangam: I think that is fair. If you're going to invest in that at that level for the company, the investment in a SWAT team — in an innovation team — is a very small percentage, and you can always cap it: 'I would seed-fund this innovation team and it'll always be capped at this amount,' because it's based on ROI. If you know the value is exponentially higher and you're actually going to create a whole new business model and open up a whole new market, that's a completely worthwhile investment — because you want to know that this will scale. You want to know that this is safe. You want to know we're going to bring it in with the right guidelines or guardrails.

Victoria Mensch: From your experience, what would be a good structure within this team? Who do you want to have on the innovation SWAT team?

Rani Pangam: When we look at innovation SWAT teams, traditionally people just think of the technology people, because we're talking about technology — not a new business process coming in. There are definitely architectural needs, security needs, engineering needs, as you assess the technology and build it into smaller, experimental-type use cases. What I also think is essential here: because AI without context is just another tool — it's almost useless — for AI to have the interpretation and the business context, it's important to have product leads with strong business context. How does our go-to-market motion work? How does our order-to-cash motion work? How do we close the books? How do we actually run our P&L? Having that business context — whether it's the product lead or actually having a business lead in there — is hugely important, especially with AI.

Victoria Mensch: And what will it take for it to leap from that pilot to a scalable project?

Rani Pangam: That honestly is so dependent on industry. I wish I had a better answer for that — I know I sound like a consultant when I say that. But having worked in different industries, I have just seen such a difference. If I'm in life sciences and I'm heavily regulated, and this technology will go into a medical device that actually impacts someone's health and life, I'm having a whole different barometer of how I evaluate the scaling versus being at a product software company where it will be a feature in a software product. So it's very dependent on the industry, and also the company's appetite to handle the risk associated with introducing new technology — because it does: there's tech debt with what you already had, and now there's change, adoption, implications to people's jobs. It has to be weighed in that context, and only each company knows best how to assess that. But there has to be that assessment.

Victoria Mensch: I'm hearing risk management and risk assessment, and that lends itself to a question of how do I experiment without compromising the core business. In the Silicon Valley innovation playbook we usually talk about the willingness to risk — a bias toward action. But there's a huge difference between a startup that doesn't have anything to lose and a large company that has an established business, a core business it really needs to protect. What have you seen that works? How do companies approach this, and in your roles, how did you manage that tension between those two efforts?

Rani Pangam: You want to keep that healthy tension, because you want to hold the innovation team accountable — not just slapping technology on for the sake of it. One of the areas where we've tried to do this: take your number one players, your high performers, to actually try that use case in real life. With high performers, you don't really have the other implications of why the technology is not working — you just know this person is really good at what they do. And when they adopt the technology, when they showcase the results, it becomes an inspiration for others to say, 'Wow, look how easy that made my life.' An example: you get your best quota-attaining sales rep to actually use an AI sales coach — one that prepares you before going into a customer meeting, with all the right briefing, all the right collateral you should know, your talk track for that executive by persona, the industry challenges that customer is having, their immediate pain points, their strategic priorities. Instead of hunting for all that information, you have it at your fingertips — and now you actually have a narrative going into the meeting with the C-suite. The success of that meeting is the evidence: it's not that hard. It's a natural language conversation I have with this AI sales coach, it's in my pocket, I have it whenever I need it — and look how successful this person was using it. So show success first, demonstrate that, and then: 'If that sales rep is your friend, you're excited about what your friend did — now I'm going to excite two more friends.' Use that chain reaction to get the excitement going. It's almost like finding champions — relying on those champions that see the value and are willing to spread that value.

Victoria Mensch: Exactly — finding champions in the organization who see the value and are willing to spread it.

Rani Pangam: The other thing also is embedding it in the flow of work directly. If I'm getting ready for a meeting and I have no other way but to use that AI sales coach to get ready, then that's just how you operate. If you make it seamless — if you put it in the flow of work of how I get ready for a meeting anyway: there's a portal I go to anyway to get all this collateral, and suddenly that portal is a completely natural-language, AI-driven portal — then you've just made me adopt AI without me even showing any resistance.

Victoria Mensch: So there's a top-down element as well. It's not just bottom-up — 'see if you want to adopt it' — but a top-down commitment: this is the way we're going, and we're going to make it stick.

Rani Pangam: Oh yeah. It's not just bottom-up: there's a top-down commitment. This is the way we're going, and we're going to make it stick.

Victoria Mensch: I know we talk about AI a lot, but it's still a newish initiative, and it's really at the intersection of so many different functions. What do you see works best — where does the ownership lie for the AI agenda and implementation?

Rani Pangam: I don't know if the decision around who owns it has changed. It's always been the P&L owner: who's getting the ROI at the end of it. There might be multiple beneficiaries, but there's always a primary sponsor who's held accountable for the ROI. If it is a commercial-led, revenue-driving AI technology, it's got to be your revenue officer, or that key regional revenue lead, whatever your model is. If it is a capability that's going to go into a product that gets launched to a customer, it has to be your chief product officer. So whoever owns the P&L and the ROI for it has to make the decisions. Because it's changing fast, you don't want to make this decision by committee and take so much time to make the decision that you've actually lost the competitive advantage you were going to get with that technology. I don't think it's been any different in the past: even when you look at these big steering committees, you can have multiple beneficiaries, but there's always still one executive sponsor and decision maker that has to be able to veto if there is a stalemate.

Victoria Mensch: You've been involved in transformation efforts for many years, in different companies. Why are you drawn to this?

Rani Pangam: I think the change with AI is that the outcome is of course very attractive compared to what technology has done in the past, but I think the process to get there has become very challenging too — and that's what makes it very interesting, at least to me; I can't say that for everyone. The ability to just keep up with what's going on... I know there's a lot of talk about the risks and harmful effects of AI, but the excitement of certain areas where we never thought it was possible: even finding what's possible with curing cancer. There are conversations now about longevity — in the next 20 years we are right around the corner of a breakthrough where humans could live forever. So I think the process of learning has now become important to all of us. The best leaders in the world are now also figuring out what this technology means, what breakthrough it's going to have, and what's the next frontier for it. The process of learning is very exciting to me — more than what some of the other technologies have done in the past.

Victoria Mensch: So embracing that uncertainty, and almost learning to thrive within it — understanding what you can control and what you cannot control — and having that element of curiosity and continuous learning: that's what actually defines success nowadays, would you agree?

Rani Pangam: 100%. And I would say that is what makes us human: the ability to question. AI has its downsides, and if you don't know what it does or how it does it, how will you define what the harmful side of it is? You don't want that to happen to you. How will you make those choices? And especially for those of us who have kids — we've got to raise them in a world where they know nothing different, where everything is done with AI and the entire physical world around them will change. So it's almost essential that we all take it upon ourselves to really learn it and grow from it: how will we ensure that it's used in the right way, and not be subject to the downsides of AI?

Victoria Mensch: It's almost a point of decision — which values to take, and how to take the effort according to those values. One of the things you mentioned that's interesting is time to value, and the pressure to show value. What I've heard from executives is that now there are expectations for us as humans to perform at the same levels — to produce results as if we were machines. That's actually a big element that contributes to burnout in a lot of organizations, something I've personally experienced, and why I became an advocate of a burnout-proof leadership strategy. So I'm curious: is this a challenge for someone in your position? Do you have tips and tricks for dealing with that pressure and stress and really getting into sustainable performance, so you can be satisfied and fulfilled in your job and not burn out?

Rani Pangam: I wish I had a magic wand. I'm going to tell you some very stereotypical answers. The first thing: this has to be a very fun learning process for us too — that's very important to me. If there's some cognitive overload and I start seeing fatigue instead of the excitement to solve the puzzle, that's a sign to me that this is getting way too much and I'm losing the interest in it. The tactical aspects of how I stay sane — I don't think I'm doing anything rocket science. It's the same as what people did about two years ago when you saw a lot of executives talk about mental burnout. One: ruthlessly prioritize the time on your calendar to have the headspace. Whatever helps you have a clear mind, where you can have the capacity to learn and grow — make that part of your routine, the way you work out, the way you brush your teeth, the way you spend time with family. There's no shame in blocking time on your calendar to just learn. Make sure you do that — and you don't have to do it on your personal time; if you have to do it on your work time, whatever that takes, make sure there's time for it. And it's not just doing it for the sake of it, but doing it because you're curious, you want to learn, you want to understand it well. So I make sure I prioritize that time pretty ruthlessly on my calendar. And I'm very, very clear about: let's not meet if we are not trying to further an agenda. If we're just meeting because I want to know status — guess what, AI can take tons of summary and give me what I need right at my fingertips. I don't need to meet for that. So that asynchronous way of working has become even more prominent. It became prominent in COVID already, but now even more, because AI is there to summarize all of that reams and reams of data and give me what I need — so I'm able to still make decisions offline. But if there is something that requires brainstorming, that requires design, that requires putting diverse thoughts together, we have to meet. And sometimes those meetings are even more efficient, because you've done a lot of the summarization ahead of time and everybody's able to input ahead of time on their own, which they don't need to meet to put those inputs in. Then we bring all those inputs in and really focus the time with crystal-clear understanding that the meat of this meeting is to take all these inputs, argue, debate — that's really why we need physical time together — and then make a decision, and then let's go back. So those are two tactics that came about in COVID anyway: just be very religious about it — the calendar and carving out the time, and prioritizing that space for curiosity.

Victoria Mensch: Those are fantastic. If you could go back for a second to the AI adoption: in your experience, what are the metrics that you would track to measure AI adoption — whether it's going well, whether it's not going well?

Rani Pangam: Some of the metrics are common to most: we're looking at usage — how much daily usage, how much monthly usage we're seeing, the number of tokens that are being used, things like that. But we're also getting to the point where we want to know: how quickly did you get to the first time you had the aha moment with that technology — and where in the workflow is that aha moment, and how quickly were you able to get to it? Being a product company, that's something we're starting to look at: when users use this product, usage is the easy KPI — we know you used it — but when did you get to the defining moment in that workflow where, yes, we know that this is when the product is going to stick around longer? Whether it is an onboarding process where you've done these three specific steps during onboarding as a hiring manager that are essential for us to learn that you're really adopting the product. So that is definitely a pivot we're seeing: going beyond just usage and API calls and classic adoption metrics, to looking at how quickly you're getting to that first aha moment while you use the product.

Victoria Mensch: So it really goes beyond just activity — whether you used it or you didn't — to how you use it, and almost your emotional reaction to it.

Rani Pangam: Exactly. And I think it makes you think more about the outcome you're trying to attain, because using technology for the sake of it is useless — we've known this through years of transformation. So you have to be a lot clearer: if I am deploying this technology, what is that aha moment that tells me the customer is actually embracing the product? And using that specific feature in the product also means that you're actually using it — you're not just going in there and exiting; you're going in there and you're in the workflow of things, actually processing what I need you to do in the tool. So that's why that whole time to — not just time to value, but time to that first aha moment — is crucial.

Victoria Mensch: Well, thank you so much, Rani. Those are really interesting and very useful points from your experience leading transformation — especially around AI, this fast pace of AI disrupting the status quo, and the potential for it to really transform not just making things cheaper and faster, but really transform the workflows and the way we work and the way our companies work and the business models. So if there is one thing you want to leave leaders with at the end of our conversation — one principle for leading through transformation, regardless of the technology — what would that be?

Rani Pangam: Like I said, I think AI is definitely unprecedented. So this is the first time I'm feeling that you have to fall in love more with the problem than the solution, because there's so much learning involved. That's the one I would say: don't focus too much on the technology and the solution — fall in love with the problem, and how you're going to learn through that journey, and make yourself better at it every single day. That would be my one takeaway.