Table of Contents
What separates standout AI products from the rest? It's not just smarter models—it's a more intentional, value-driven mindset. In this edition of the Founder Loop, our CEO Ryan Millner and Pete Giordano from Scale Venture Partners break down what it really takes to build AI products that drive outcomes, win customers, and earn investor conviction.
Here’s a bit of background on our speakers, in their own words:
Meet the speakers
Peter Giordano: “I work at Scale Venture Partners as a CMO EIR (Executive in Residence), and I help our portfolio companies unlock growth. My career spans roles from the early days of the Internet—at companies like Net Gravity and Double Click.
I spent a decade at VMware, and most recently, Google Cloud, where I led all of the go-to-market (GTM) strategy for Google's Workspace product. I'm a 2-time founder during market downturns, giving me some firsthand experience on navigating resilience and critical transitions.
I've learned that my superpower is helping startups translate very complex problems into a clear strategy. Lately I've been focused on helping companies translate their AI innovation into business value.”
Ryan Millner: “I’m one of the co-founders, and CEO of Unwrap. Previously, I was at Amazon on the Alexa team—trying to make Alexa smarter. I got to see the evolution of natural language processing (NLP) with ChatGPT, and everything that's come out of that over the last 5-6 years.
Really excited to chat with Pete today about Unwrap. We are a customer intelligence platform, helping companies better understand their customers by analyzing feedback.”
Engineering impact is more than a model—it’s a mindset
Giordano: “Let's get a little bit deeper into Unwrap’s origin story. I always find that moment fascinating—the moment you realized there was this fundamental gap in how companies understand their customers.
So take us back to Amazon, while you were working on Alexa. What was it like? What was the specific issue? What was the frustration you felt that then planted the seed for Unwrap?”
Millner: “During my time at Amazon, I was a product leader and one of my core responsibilities was to understand our customers. I needed to know things like: did people think Alexa’s answers were too long or too short? Was Alexa missing an event they wanted to ask about? Were our visualizations on their multimodal devices valuable?
But I felt like I really couldn’t do a good job of actually understanding them—and that wasn't because of a lack of effort or desire. It was just really hard. Alexa has tens of millions of users. You can’t sit down with all of them and have a conversation about what they like and don't like about the experience. You can’t read every Reddit post, support ticket, or review.
So when it came time to build the roadmap—plugging in what we thought our customers wanted—it was very anecdotal, based on intuition.
The most frustrating part was that our customers were telling us exactly what they thought about the product, all over the web. They were very active in the subreddit for Alexa. They were leaving support tickets and reviews every day. They were tweeting. I just didn't have the tools to understand it all, fully.
So the lightbulb moment was: what if we took this technology we knew really well at Alexa—called natural language processing (NLP)—and pointed it at this corpus of customer feedback data?
I thought, ‘That would give me so much more insight into what our customers wanted and loved about the product—and allow us to build a better product’, which was really our goal.
We then talked to a bunch of product leaders inside and outside Amazon to try to figure out, is this a Ryan problem, an Amazon problem, or a broader issue? It very clearly resonated with others, which gave us the confidence to leave Amazon and start this company.”
Giordano: “We've gotten a chance to know each other a little bit over the last several weeks, and one of the things that has come up, as I’ve heard you tell this story, is that that origin of Unwrap was not just born out of that insight in the moment, but also from your desire to create a result—a benefit for the customer.
It’s this idea of a results-first philosophy. I’ve seen firsthand the opposite—so many AI solutions that have technology in search of a problem. How do you and Ashwin define results-first thinking? And how does that help you shape the day-to-day decisions that create this product that provides all these benefits to your customers?”
Millner: “As a product manager or product leader, you are not measured on what technology you use or how you implement something. At the end of the day, it's really just, are you solving the problem? For Ashwin and I, this has been core to our entire career and what we’ve built at Unwrap.
Our focus is on solving specific problems for our customers, and we don't really care what the best way to solve it is. I think that's important for a couple of reasons:
- Your customers don't care. They don't care if there's a thousand humans on the other end of a dashboard. Or if it's all powered by AI. They just want it to work, work well, and solve the problem they came to you for.
- The AI space is evolving so quickly. If you get married to a particular approach, that way may be obsolete in 6 months or a year.
For us, it’s always about solving the problem in the best way possible today and staying close to different ways of solving the problem as technology evolves. We use different NLP techniques throughout the platform. We’re really well versed in LLMs, fine tuning larger models, building models ourselves, using smaller off-the-shelf models when applicable.
Sometimes, the right solution doesn’t include AI at all. All that matters is that the end result works.”
Giordano: “There's an old ad from the 1940s, and it read something to the effect of, ‘Nobody wants a drill, they want the hole.’ Customers just want the result. I love the journey you’ve described there, but I want to go a little deeper.
What does that actually look like in practice? What does it look like to start to apply that results-first mindset? What are companies today missing if they're relying on their traditional analytics? Or existing, basic customer success tools?
The alternatives that you run into today—what are they missing and why? Why is finding a partner that’s solely focused on outcomes so valuable?”
Millner: “When I was at Amazon, there were two things I wanted to do as a product leader. One was to understand, with the existing 10 issues or priorities I knew about, which one was most important to our customers.
The much harder problem was finding things that are problematic for our customers, that we didn't yet know about. When we started Unwrap, that was the problem we thought was more valuable to solve.
Existing customer intelligence platforms may say, ‘Here are the 10 things that customers are complaining about.’ But you’ll know about them already—those buckets of complaints. It was really valuable to show our customers new issues or friction points they've never seen before—in a really digestible format.
Instead of someone having to come in and build their own dashboard or write their own query, we wanted to proactively send alerts and emails to people saying, ‘Here's a new issue that you aren't aware of. It's growing to this percentage, and we think it's important to fix it.’
When we're onboarding a new customer, if we're able to show them something new in the first minute, their eyes light up and they're hooked. That's when we know we've done a good job.
We’ve relied on a combination of humans and technology to make this possible. We did not take the approach of automating Unwrap 100% from day one. We weren’t going to prioritize automation over quality and we weren’t going to have no humans in the loop.
The quality wouldn't have been high enough, and the customers would have suffered because they weren’t going to get the right insights. Instead, we took the approach of delivering the best insights to every single customer—and then, figuring out how to automate that process more and more over time.
It was easy when we had five customers to have more humans in the loop—to check whether the insights were correct. Now, that's obviously not possible. But we've never lost sight of our quality bar. Luckily we've gotten better at scaling the automation of Unwrap to now where it's pretty much fully automated. But we didn’t take that approach from the beginning. We weren’t willing to compromise on quality, and that holds true today.”
Giordano: "Wow! I love that. There's so many gems there. One of them—the idea of, at the very beginning, resisting the urge to automate and instead doing the more laborious, manual tasks so you get the right result. Can you give me an example of an insight that you were able to reveal to a customer that was mind blowing to them?"
Millner: “We were working with a wearable company covering one of their launches. They were leaning on us to make sure the launch went smoothly and to capture any friction points—and hopefully, fix them quickly.
For a hardware company, these launches can be 2-3 years in the making—they are a really big deal. By analyzing all of their support tickets and comments on social media, we were able to identify a couple issues that were preventing upgrades to the new system. That's a lot of revenue on the table potentially—if people aren't able to experience the new product as smoothly or quickly as they'd like to.
During the launch we were able to identify issues that kept popping up and they were able to fix them quickly. That type of coverage wouldn't have been possible without a platform like Unwrap—because you're not able to, at scale, understand the entire corpus of customer feedback.”
Giordano: “Yes, that you're getting the entire corpus, that’s part of the difference. The customer intelligence space is pretty crowded. It's a mature market. And every one is talking about the ability to provide insights. When you're a customer, you’re looking for clear differentiation. Customers want something that is going to help them stand out, better compete in the market—but they also want something that’s highly valuable to the company.
What is it about the approach that Unwrap is taking that's fundamentally different from some of these other alternatives?”
Millner: “One advantage is that my co-founder and I, we're both product managers. And we're primarily selling to product managers and customer experience leaders. We are building the product that we would have bought personally—going back to the fact that the hardest problem for us was the “unknown unknowns,” the uncovering of things that we didn't even know existed—that problem is pretty hard to solve.
Imagine you have tens of thousands of phone calls and you have hundreds of thousands of support tickets and tweets and reddit posts. Technology is much better at saying, ‘This is a piece of feedback about login problems,’ and then going and finding all the feedback about login problems. That’s an easier problem to solve than saying, ‘Show me 10 things I don't know about.’
We have developed a proprietary algorithm to find new issues for customers. The only reason we're focused on this is because it’s what we wanted when we were at Amazon. It's all about being proactive and servicing insights—whether they're new insights, whether they’re growing or shrinking—automatically to our customers.
Our customers don't want to have just another dashboard to log into. They don't want to have to come, build a dashboard or write a query themselves. They want to simply get an email or a slack that says, ‘Hey Pete, this is what's broken over the last 2 days. Here's how you can fix it. Here's the impact of that.’
That's the most delightful experience—and that's what we're constantly building towards. It goes back to being as proactive as possible, and requiring little from our customers.”
Giordano: “That's wonderful—and brings us full circle. Throughout my career, I’ve noticed for founders I work with—whether at Scale or other accelerators in the ecosystem—there is an always-present tension. A tension between the urgent things customers want and the long-term things that are actually needed to deliver those results.
How does a team of your size and experience navigate that tension? How can you use that results-first philosophy, while also navigating those decisions to find the right balance to address what customers are asking you for?”
Millner: “It's applicable for our customers and for us as a company, because we face the same problems as our customers. You're really trying to figure out: I have a set number of resources to build the product—how do I most effectively use those resources? Do I spend it on fixing bugs and all these little points of friction in the product? Or do I launch brand new pillars of the product that are going to be entirely new experiences that maybe our customers aren't asking us for today?
The answer is both—you have to do both all the time. The trade-off for us is centered around what’s going to keep customers happiest—and making sure that they continue to renew—while also finding new opportunities to serve more of their needs. Finding those new revenue streams that maybe we didn't have before.
For us, it's always about understanding our customers’ needs at a really deep level and understanding the areas where we’re meeting their needs and where we could improve. For example, we previously did not have a way for teams to understand how effective their support teams were at handling tickets.
We’re analyzing a lot of our customer’s support tickets—it was natural for us to say, ‘We're going to help them understand how to make their agents more effective.’ That was a new line of business for us.
But that doesn't mean you can abandon your existing customers' needs and all the friction points they're currently experiencing. It's a constant balance—and again, it comes down to understanding what they care about and what needs are not met today.”
Watch the full session to hear the Q&A portion of the webinar with Ryan and Pete!
