Case Study: AI-Powered Customer Discovery

Speed to Judgement

How Xylem’s product managers make better decisions, faster
Heidi Smith
“Moiz’s coaching empowered us to make roadmap decisions through a discovery sprint. We applied it to our existing roadmap immediately. My team walked away with skills and AI tools they’ll actually use, not just a framework they’ll forget.”
Heidi Smith · Director of Product, Xylem

Speed to revenue requires speed to judgement

Xylem is a leading global water and energy solutions partner. Its Sensus brand delivers remotely-managed metering products that give utilities the right data at the right time. A key growth strategy is increasing sales of new water, gas, and electric meters.

That growth depended on rewriting a legacy software platform owned by Heidi Smith’s team. The rewrite had to deliver a great customer experience, and stakeholders needed it shipped last quarter. When Heidi saw her team would have to justify every feature they kept or cut, she wanted those calls backed by evidence, not opinion.

The Problem
01

Where’s the Evidence?

Stakeholders needed customer-backed proof for each feature.

02

Which Features to Cut?

The team needed to know what customers truly valued.

03

How to Use AI?

AI wasn’t part of real product decisions yet.

The Approach
AI-Powered Customer Discovery sprint that pairs discovery fundamentals with AI to do it faster.
The Result
Roadmap decisions, in a fraction of time
1 week
not 3–4 weeks, to make a decision
13 PMs
trained to use AI in discovery
100%
would recommend this to other teams

How Andrew made a high potential, low risk product decision

Before
3–4 weeks
Now
1 week

Andrew juggled internal and external meetings, with no structure for getting to a decision.

Andrew reaches a decision he presents to engineers as high potential, low risk.

Andrew Fleming
“After learning discovery fundamentals, I know exactly what I need to de-risk before launching a feature. And with AI, what used to take me several meetings over 3–4 weeks now takes me a week.”
Andrew Fleming · Technical Product Manager, Xylem

How did Andrew do that?

He found the riskiest assumption about his product idea, tested it with customers, and made a decision based on what he learned.

When he was redesigning mapping features, his assumption was that customers needed a full GIS-level rebuild. When he interviewed customers, he found the real need was far narrower: users only needed to make custom layers, not an entirely new system.

That single insight saves months of engineering time. Now, he can make a recommendation of what customers truly need, accelerate the roadmap, and allow Xylem to start selling meters much sooner.

Discovery fundamentals
have not changed…
Define
Assumptions
Customer
Empathy
Prepare
Interviews
Talk to
Customers
Decide
What changed is the speed to do it.

AI removes the manual grind: analyzing interviews, surfacing patterns, pressure-testing assumptions.

The decision still belongs to the PM.

The hours don’t.

The Xylem product team applied timeless discovery principles with AI tools. The result was faster, better judgement, and the organizational credibility that comes with it.

AI is a force multiplier for product judgement.

“Highly recommend this. I’ve read the books on customer discovery, but putting it into practice is harder than it looks…it takes real judgment, communication skills, and a shift from building to learning. Moiz helps you put that into practice, using AI tools at every step.”

Abid
Sr. Director, Product Management · Ecomedes

“This is a great way to leverage AI in your customer discovery process. Moiz is a great instructor and he provided us with structured tools that can enable a PM to get stakeholder buy-in in a build versus pivot decision making process.”

Aditi
Sr Product Manager · Real-Time Innovations Inc

“Moiz reshaped how I think about using AI in product work. Beyond the obvious use cases, he taught me how to apply AI to the higher-leverage parts of the job — customer discovery, insight synthesis, and decision-making. I walked away with reusable workflows and a stronger mental model for how AI can augment good product judgment.”

Dalton
Product Manager
Want to be the PM who makes decisions like this?
I run training and coaching to show you how.
DM me on LinkedIn →
Moiz Kapadia has trained 100+ PMs on customer discovery. He has 10+ years of PM experience ranging from Oracle to pre-seed startups. Moiz is a Fractional Head of Product and Advisor.
moiz@moizkapadia.com
linkedin.com/in/moizkapadia