C#
Better C# Switch Statements for a Range of Values
By hackajob Staff
If you’ve worked in product or engineering for more than five minutes, you’ve probably heard the same promise on repeat:
AI will supercharge discovery.
AI will automate research.
AI will understand your users.
But here’s the reality we unpacked in this DevLab episode with Elizabeth Ferrao, Product Lead at C Spire: AI can speed up a lot of things, but it still can’t replace actually knowing your users.
And if you’re building products today, that gap matters more than ever.
In this conversation, we dug into why customer understanding is still the real moat for product teams, how fast prototyping can trick us into thinking we’re solving the right problem, and why hearing someone’s voice can tell you more than any AI summary ever will.
Let’s get into it.
The stuff AI can do… and the stuff it really can’t
Elizabeth is honest about it: she loves the productivity boost. AI is great at the boring parts: managing outreach, reminders, scheduling, transcribing notes, and helping draft an interview guide.
But the moment you ask AI to replace judgment, nuance or human interpretation?
It falls apart.
As she puts it:
“I give AI every transcript, every note, all the context… and it still misses the big insight. It doesn’t know our customers like we do.”
That’s the core thread running through the entire episode:
AI can automate the admin. It cannot automate understanding. And in product, understanding human behaviour is the work.
Continuous discovery isn’t optional anymore
One of the strongest points Elizabeth makes is about scale. Not the scale of systems, but the scale of insight.
Most PMs talk to five or ten customers and call it done. But humans aren’t static. Their needs change. Their expectations change. The world around them changes.
For Elizabeth, discovery is never finished:
“We are constantly evolving. Our customers are constantly evolving. You can’t stop talking to people.”
And she means literally talking: hearing tone, energy, frustration, joy. Transcripts won’t catch that. Summaries won’t catch that. And AI definitely won’t catch it.
Fast prototypes don’t fix the wrong problem
Prototyping has become ridiculously fast. You can spin up an idea in minutes. But if you’re not deeply connected to your users, you’re probably validating the wrong thing.
Phil sums it up well:
“It doesn’t matter how fast you ship if you don’t understand the problem.”
Elizabeth agrees. She’s built for home users, wireless customers, and B2B buyers. Three massively different audiences using the same product for completely different reasons.
Without continuous conversations, every decision becomes guesswork disguised as strategy.
Why hearing someone’s voice still matters
One of the most powerful parts of the episode is when Elizabeth talks about the non-verbal side of research.
AI can summarise your notes, but it can’t feel what you felt in the moment: the pause before someone answers, the laugh, the annoyance, the "oh wow, yes" moment you didn’t expect.
She puts it simply:
“You need to hear someone’s voice. That’s where the insight is.”
Bring engineering into the room (literally)
Elizabeth invites her engineers to sit in on early interviews. Not because she wants them in discovery forever, but because nothing beats hearing a user describe a pain point in their own words.
Inviting engineers into these sessions helps everyone see the same challenges through the customer’s eyes. It encourages empathy and reduces unnecessary back‑and‑forth because teams are reacting to real stories rather than assumptions. When product and engineering hear the same input, decisions become clearer and debates tend to resolve much faster.
“If product and engineering are debating something, sometimes the quickest fix is: let’s go ask a customer.”
Why customer understanding is the only real moat left
With AI getting better at generating specs, writing code, and speeding up execution, the one thing teams can’t outsource is the thing that matters most:
Knowing your users deeply enough to build something that actually works for them.
Elizabeth calls it the moat, the durable advantage product teams still own.
Competitors can copy features and even your design patterns, but they cannot replicate the level of customer understanding your team builds over time. That knowledge shapes better decisions, reduces wasted effort and ultimately becomes the thing that keeps your product ahead. It only comes from doing the work consistently.
Final takeaway: talk to more humans
If people remember one thing from this episode, Elizabeth hopes it’s this: hear your customers, genuinely and directly. Not summaries or assumptions, and not AI interpretations of your notes, but real conversations with real people.
The pauses, the tone, the enthusiasm and the frustration all tell you something no tool can. These interactions might be unscalable and occasionally inconvenient, but they remain some of the most valuable inputs in product work.
If you build products, write code, lead teams or just care about building things people actually use, this conversation is worth a listen. It’s a reminder that in a world full of AI-powered shortcuts, the work that matters most is still deeply, stubbornly human.
