How to prioritize, build, and measure AI: Insights from the PendomoniumX NYC customer panel

Over 350 software leaders gathered at PendomoniumX NYC to build community and learn how to navigate the AI era. A standout session featured Pendo customers sharing real-world insights from their AI implementation journeys.

Panelists included Alo Mukerji (SVP of Product Management, US Bank), Kelly Delaney (VP, Learning Experience Design, Savvas Learning), and Mike Bartels (SVP of Workflow & Insights, Nasdaq).

Here are their key insights and practical takeaways for your own AI strategy.

How much is your company investing in AI? How are you balancing that with other software investments?

Mike Cartels (Nasdaq): We consider ourselves a technology company, so AI is central to our strategy. For internal operations—engineering efficiency and customer success—we’re at a 10 out of 10. Every software investment considers AI strategy and value creation. We’re currently around a 5 or 6 for product features since we’re still delivering existing roadmap items. We’re at an 8 or 9 for AI product integration.

On the product side, we’re closer to a five or six in terms of actual investment, only because we have the rest of the product roadmap to deliver. Not everything our customers are asking for is AI first. But we’re at an eight or nine when we think about the roadmap after the next couple of quarters and integrating AI into our product.

Kelly Delaney (Savvas Learning): We’re very focused on AI—30% or more of our roadmap may be based on AI features—but the actual investment in software isn’t going to change. We’re considering how to make our product experience more efficient for customers. Their needs haven’t changed; we’re just finding better ways to serve them.

Alo Mukerji (US Bank): We’re in an industry with heavy regulations and compliance, so getting AI tools approved is very hard, never mind using them in practice. We’ve focused more on internal use cases and efficiencies, which are easier to approve. On the customer side, we have a special group focusing on AI, more on the data side, trying to figure out different ways to ultimately deliver customer value.

How do your customers feel about AI?

Alo Mukerji (US Bank): Our customers deal with financial data and transactions, so they don’t want their privacy or data shared. But they do want speed—they expect us to do things faster, whether developing features faster, testing faster, or delivering value faster. They want the benefits as long as we protect the data.

Mike Cartels (Nasdaq): We drove AI conversations with customers twelve months ago. Now, they expect AI as the first or second agenda item in roadmap discussions. Curiosity is high, and hopefully in 12 months, we’ll be talking about who’s not using it.

Kelly Delaney (Savvas Learning): About a year ago, we pulled together a panel of kindergarten through fifth-grade reading teachers. 15 out of 16 used AI regularly, but not the AI startups you’d expect. 

They were using Canva AI and other tools already embedded in their workflow. The insight is that they’re finding AI tools within the workflow they’re already in, rather than going to separate tabs or tools.

Are you using AI to solve existing or new problems?

Mike Cartels (Nasdaq): Both. We’re finding new ways to help customers solve problems they initially purchased the software for, but we’re also seeing other things they were doing outside our product and bringing them in. 

For example, we discovered customers spent dozens of hours summarizing board materials before meetings. We built AI capabilities to solve this, bringing that workflow into our product to increase stickiness.

Alo Mukerji (US Bank): The problems are the same for small businesses and restaurants, but the expectations around speed to solution are different. What’s changed is the possibility of solutions we can offer. 

We’ve always talked about providing proactive insights to help customers save time and money, but that seemed like a pipe dream before. Now, it’s a real possibility—we could potentially give customers actionable insights that would be transformational for their business.

Kelly Delaney (Savaas Learning): The problems are absolutely the same, but really intractable problems that we’ve wanted to be part of the solution for are now within reach. 

Gathering disparate forms of student data and providing meaningful feedback to teachers, or creating the depth of curriculum to serve neurodivergent students—things that would have required an impossibly tall stack of textbooks—are now possible.

How are you using Pendo for discovery in your AI journey?

Kelly Delaney (Savvas Learning): First, I have to say the product discovery course is transforming how we think about ourselves as a product organization. We’ve been looking at funnels to understand natural workflows, then determining where we can add AI solutions or the right product solutions within those natural paths through our curriculum offerings.

Alo Mukerji (US Bank): Pendo is critical to our product discovery. It allows us to do things much faster because it’s a partner tool, helping us work around regulations. We use it on the support and customer success side for data about customer questions and all our tracking data, and we’ll be using Listen soon. As Pendo has more AI features, it helps us be more AI-forward in our practices.

Are you measuring AI features differently than traditional features?

Mike Cartels (Nasdaq): We use traditional metrics like utilization, frequency, and how often features are used, but we’re really trying to track time saved specifically. Time is a valuable currency for any profession. 

This is important because we have a strong discipline to measure return on investment. We’re trying to figure out the monetization path. Is it good, better, or the best pricing? Price increases over time? Standalone modules? We want to attribute distinct revenue to our AI investments.

What advice would you offer for organizations on their AI journey?

Mike Cartels (Nasdaq): First, set up governance early, especially for regulated industries. Second, the power of technology is on an exponential path—things you tried three or four months ago that didn’t work might work now, so go back and try again. Third, you won’t come up with ideas for how to use AI if you’re not using it yourself. Experiment with ChatGPT, Perplexity, or any LLM in your personal life, then translate that experience to customer value.

Kelly Delaney (Savvas Learning): Learn with your team and model what learning looks like. Share what you’re learning—learning is collaborative. Set up chat rooms, task forces, and share-outs. Don’t be afraid to try new things and model that lack of fear. We’re all at the starting line, so surround yourself with people with intellectual curiosity and learn together.

Alo Mukerji (US Bank): Focus on where the customer value is, even in internal use cases. When AI feels overwhelming, focus on what’s going to help the customer, and the rest will come. 

This panel was held on June 17th, 2025. Read more from PendomoniumX NYC, or explore everything in the Summer ‘25 Release.