“We are generally the better persuaded by the reasons we discover ourselves than by those given to us by others.” - Blaise Pascal
image by ChatGPT
The question came before Joe, our PM had finished with his second slide. "Where is AI in this roadmap?"
The CTO wasn't hostile. She was genuinely puzzled — and behind her puzzlement was something harder to address: a conviction that the roadmap should look different.
Joe explained what we were building, why the current priorities addressed more urgent customer needs, why AI features without the right data foundation would underperform. But no one seemed convinced.
A week later, Joe tried something different. Before our next stakeholder conversation, he played three minutes of a recorded customer interview. A senior ops manager at one of our mid-market accounts, describing in her own words exactly the problem we were trying to solve. She was specific, frustrated, and completely unscripted. At no point did she mention AI.
This time nobody asked where AI was.
Why the customer's voice travels further than yours
As product leaders we are trained to build a case backed by research, metrics, trade-off analysis, under the assumption that a better argument wins.
Sometimes it does. But in a meeting where senior stakeholders have their own hypotheses, and strong convictions about what the product should do, the product leader's argument is seen as advocacy.
A customer's voice is treated differently. It has no agenda the room can question and rarely needs defending.
A three-minute recording of a frustrated user describing their week does something that a well-crafted slide cannot: it makes the problem feel real before the solution is proposed.
But why does customer evidence land differently? It's not that it's more persuasive. It's that it's safer. A stakeholder who backs your recommendation is taking a personal risk. A stakeholder who backs a customer's expressed need is simply being responsive. One is advocacy. The other is service.
For product leaders, this is a specific and learnable shift. Before the next contested roadmap conversation, the question worth asking is who could make this case better than I can?"
Watch It In Action
Teresa Torres author of Continuous Discovery Habits, gave a keynote at Mind the Product San Francisco. In it she makes a precise and uncomfortable point: most product managers present their final answer to stakeholders and then wonder why stakeholders push back.
The problem is that by skipping straight to the conclusion, they've left out the customer evidence that makes the conclusion feel inevitable. Watch from the 9-minute mark.
Try This Week
Before your next roadmap or prioritization conversation where you expect pushback:
Build a folder I keep three recordings ready: one for the AI question, one for the "why are we spending time on this?" question, and one for the "can we just ship it?" question. Find your version of this.
Find one piece of unscripted customer voice that speaks directly to the problem you're trying to solve. Not a quote from a slide, but a recorded interview clip, a support ticket, a verbatim survey response. Two to three minutes is enough.
Open with it before you present your recommendation. Not as supporting evidence at the end — as the opening. The sequence matters: problem felt, then solution proposed.
After playing it, stay quiet for a moment. Resist the urge to immediately interpret or explain. Let the room respond to the customer first.
Joe now starts every contested roadmap conversation the same way: with a recording. He's learned that his job isn't to be the most convincing person in the room. It's to make sure the most convincing person gets to speak.
From Elsewhere
In Welsh, hiraeth describes a longing that has no direct English translation — a grief for Wales, its land, its history, and a way of life that may be gone or may never have fully existed. It sits at the intersection of homesickness, nostalgia, and a deeper sense of loss tied to Welsh identity. The Welsh have a word for it because, for them, it has always been worth naming. More on hiraeth here »