“If you understand the problem, AI can help expedite the learning phase. Whether that’s research, designing, or building.”
Lewis Gavin, Head of Growth at easyfundraising, keeps coming back to this idea when he talks about the shift happening inside product and growth teams—especially as AI becomes more embedded in day-to-day work.
Formerly a lead data architect, Lewis now spends his time growing the business on all fronts, from topline growth and retention to engineering. His path was intuitive, yet untraditional.
“I wanted to transition into growth because there were a lot of paths to delivery that we already identified. I already had a good understanding of our product, and how the customers use it. We just needed someone to start embedding a growth mindset into our teams.”
That shift coincided with a broader change inside easyfundraising. Shipping velocity—not ideation—had become the core constraint.
It wasn’t a lack of ideas or ambition holding the team back. It was the gap between understanding what users needed, and turning that understanding into something teams could align on and actually ship with confidence.
“The general issues were bottlenecks, iteration speed, and how quickly we could actually ship,” says Lewis. “I now run a team of 10 looking after our core cause experience, with the explicit goal of making those experiences as good as possible.”
Proving the Problem
easyfundraising’s product organization was split by platform—website, app, and browser extension—but customers lived across all of these.
Moving to customer teams meant insight around each type of customer no longer had to travel across teams and disciplines. But the teams still needed to improve how they listened to customers.
Before adopting Reforge Insights, there was no shortage of customer feedback—it just took a fair bit of manual effort and copying and pasting context between apps.
“We were basically swimming in survey responses. Customer feedback has always been very important, but we only had one person manually tagging every survey for ‘insights,’ aggregating surveys and app store reviews, and trying to distribute them manually. It was hard to get into what the actual takeaways and pain points were,” notes Lewis.
“We actually tried to build something like Reforge Insights before we knew it existed.”
Having taken Reforge Learning’s course on growth, Lewis was no stranger to the brand—but wasn’t aware of the SaaS products.
“We looked at AI and feedback systems, trying to create a system that could find a way to discover what users were experiencing. A lot of our features are directly inspired by user behavior, so we wanted to get this right,” says Lewis.
Trying From Scratch
“We were looking at AI and customer feedback tools and trying to find a way to make it easy to ask a question and understand what users were experiencing,” says Lewis. “AWS even has special budgets for teams working in this space.”
The team went as far as running multi-day coding sessions to build an internal solution.
“We did get something working,” he says. “But it was clunky. It wasn’t great. It didn’t quite do what we wanted.”
Then Reforge Insights landed in someone’s inbox.
“We ended up on a call with Brian and Chun, talking through what they were building. It immediately clicked.”
From Bottleneck to Leverage
Once Reforge Insights was in place, the bottleneck disappeared.
“Now anyone can figure out what our users are talking about,” Lewis says.
He built a custom report focused on the users and behaviors he cares about most. Every week, it lands in his inbox with emerging trends—effectively a running pulse check on customer pain.
“It’s basically: here’s what your customers care about right now. It helps realign my focus on what is important.”
Product teams use it to prioritize work. Marketers use it to pull real customer quotes. Instead of insights living with one person, they became shared context across the company.
But even with clearer signals, there was still a hard problem to solve: turning that shared understanding into something real enough to validate—without slowing everything back down.
What Fast Actually Looks Like
Before AI-powered prototyping, speed broke down in familiar ways—despite everyone having clearer signals about what mattered. Validation often came late, after ideas had already gathered momentum.
There was manual effort and specialised skills required to build prototypes, which could be time-consuming. As a small company, we’re always resource-bound, so finding ways to democratize prototyping across the product org was super helpful.
User testing followed a similar pattern. The team relied on third-party panels and proxy users—often people who had never heard of easyfundraising—interacting with static screens or lightly clickable flows. To be clear, the feedback here was still really useful, but we never did it frequently enough, often due to time constraints of building functioning prototypes.
Lewis describes a moment where a recurring customer pain point surfaced around referrals. Instead of kicking off a long discovery cycle, he moved straight from insight to execution.
“I took the insight and dropped it into Build,” he says. “I asked it to visualize what this could look like.”
In no time, he had branded, realistic prototypes. Using Build’s browser extension, those prototypes captured real pages and flows from the site—automatically reflecting easyfundraising’s actual design system. They looked authentic enough that users wouldn’t question whether they were real.
He screenshotted the prototypes, dropped them into the engineering workflow, and the feature moved from idea to build to approval—inside a single day.
“For me, Build solved the frontend gap,” Lewis explains. “I’m not a frontend guy. I needed something that looked like us, that we could actually show to users and that I could use to ship quick experiments without burdening our UX team”
That was the difference from other tools.
“There’s a lot of AI prototyping out there that’s useful for ideation,” Lewis notes. “But it usually doesn’t look like your product. We needed something that felt real enough to align teams and validate decisions before writing code.”
He elaborates, “There’s a lot of vibe coding out there, but none of it looks like us. We needed prototypes we could put in front of real users without caveats.”
Read more about this experiment on Lewis’s Substack.
Where it All Comes Together
The shift wasn’t without friction. Like many companies, Lewis noted that easyfundraising has felt the impact of today’s unhelpful “AI is replacing jobs” narrative. However, they have found that AI simply accelerates the work they were already doing, freeing up their scarcest resources to focus on building bigger and better things.
Over time, the benefits became harder to ignore. Ideas were more fully formed before development began. Internal feedback loops tightened. And instead of discovering problems late—when fixes were expensive—teams aligned earlier, with more confidence in what they were building.
Lewis is the first to admit that product wasn’t always his lane. But the combination of deep product understanding, AI-powered insights, and fast prototyping changed the equation—and the shipping paradigm entirely.
“If I need something to just work, I can just get in there and fix it,” he says. “All of this—the new AI tools, the experimentation mindset, the willingness to try things—is not only increasing clock speed, but also improving research and outcomes.”
He comes back to the same idea.
“If you understand the problem, AI can help expedite the learning phase. Whether that’s research, designing, or building.”
Related reading and technology:
“If you understand the problem, AI can help expedite the learning phase. Whether that’s research, designing, or building.”
Lewis Gavin, Head of Growth at easyfundraising, keeps coming back to this idea when he talks about the shift happening inside product and growth teams—especially as AI becomes more embedded in day-to-day work.
Formerly a lead data architect, Lewis now spends his time growing the business on all fronts, from topline growth and retention to engineering. His path was intuitive, yet untraditional.
“I wanted to transition into growth because there were a lot of paths to delivery that we already identified. I already had a good understanding of our product, and how the customers use it. We just needed someone to start embedding a growth mindset into our teams.”
That shift coincided with a broader change inside easyfundraising. Shipping velocity—not ideation—had become the core constraint.
It wasn’t a lack of ideas or ambition holding the team back. It was the gap between understanding what users needed, and turning that understanding into something teams could align on and actually ship with confidence.
“The general issues were bottlenecks, iteration speed, and how quickly we could actually ship,” says Lewis. “I now run a team of 10 looking after our core cause experience, with the explicit goal of making those experiences as good as possible.”
Proving the Problem
easyfundraising’s product organization was split by platform—website, app, and browser extension—but customers lived across all of these.
Moving to customer teams meant insight around each type of customer no longer had to travel across teams and disciplines. But the teams still needed to improve how they listened to customers.
Before adopting Reforge Insights, there was no shortage of customer feedback—it just took a fair bit of manual effort and copying and pasting context between apps.
“We were basically swimming in survey responses. Customer feedback has always been very important, but we only had one person manually tagging every survey for ‘insights,’ aggregating surveys and app store reviews, and trying to distribute them manually. It was hard to get into what the actual takeaways and pain points were,” notes Lewis.
“We actually tried to build something like Reforge Insights before we knew it existed.”
Having taken Reforge Learning’s course on growth, Lewis was no stranger to the brand—but wasn’t aware of the SaaS products.
“We looked at AI and feedback systems, trying to create a system that could find a way to discover what users were experiencing. A lot of our features are directly inspired by user behavior, so we wanted to get this right,” says Lewis.
Trying From Scratch
“We were looking at AI and customer feedback tools and trying to find a way to make it easy to ask a question and understand what users were experiencing,” says Lewis. “AWS even has special budgets for teams working in this space.”
The team went as far as running multi-day coding sessions to build an internal solution.
“We did get something working,” he says. “But it was clunky. It wasn’t great. It didn’t quite do what we wanted.”
Then Reforge Insights landed in someone’s inbox.
“We ended up on a call with Brian and Chun, talking through what they were building. It immediately clicked.”
From Bottleneck to Leverage
Once Reforge Insights was in place, the bottleneck disappeared.
“Now anyone can figure out what our users are talking about,” Lewis says.
He built a custom report focused on the users and behaviors he cares about most. Every week, it lands in his inbox with emerging trends—effectively a running pulse check on customer pain.
“It’s basically: here’s what your customers care about right now. It helps realign my focus on what is important.”
Product teams use it to prioritize work. Marketers use it to pull real customer quotes. Instead of insights living with one person, they became shared context across the company.
But even with clearer signals, there was still a hard problem to solve: turning that shared understanding into something real enough to validate—without slowing everything back down.
What Fast Actually Looks Like
Before AI-powered prototyping, speed broke down in familiar ways—despite everyone having clearer signals about what mattered. Validation often came late, after ideas had already gathered momentum.
There was manual effort and specialised skills required to build prototypes, which could be time-consuming. As a small company, we’re always resource-bound, so finding ways to democratize prototyping across the product org was super helpful.
User testing followed a similar pattern. The team relied on third-party panels and proxy users—often people who had never heard of easyfundraising—interacting with static screens or lightly clickable flows. To be clear, the feedback here was still really useful, but we never did it frequently enough, often due to time constraints of building functioning prototypes.
Lewis describes a moment where a recurring customer pain point surfaced around referrals. Instead of kicking off a long discovery cycle, he moved straight from insight to execution.
“I took the insight and dropped it into Build,” he says. “I asked it to visualize what this could look like.”
In no time, he had branded, realistic prototypes. Using Build’s browser extension, those prototypes captured real pages and flows from the site—automatically reflecting easyfundraising’s actual design system. They looked authentic enough that users wouldn’t question whether they were real.
He screenshotted the prototypes, dropped them into the engineering workflow, and the feature moved from idea to build to approval—inside a single day.
“For me, Build solved the frontend gap,” Lewis explains. “I’m not a frontend guy. I needed something that looked like us, that we could actually show to users and that I could use to ship quick experiments without burdening our UX team”
That was the difference from other tools.
“There’s a lot of AI prototyping out there that’s useful for ideation,” Lewis notes. “But it usually doesn’t look like your product. We needed something that felt real enough to align teams and validate decisions before writing code.”
He elaborates, “There’s a lot of vibe coding out there, but none of it looks like us. We needed prototypes we could put in front of real users without caveats.”
Read more about this experiment on Lewis’s Substack.
Where it All Comes Together
The shift wasn’t without friction. Like many companies, Lewis noted that easyfundraising has felt the impact of today’s unhelpful “AI is replacing jobs” narrative. However, they have found that AI simply accelerates the work they were already doing, freeing up their scarcest resources to focus on building bigger and better things.
Over time, the benefits became harder to ignore. Ideas were more fully formed before development began. Internal feedback loops tightened. And instead of discovering problems late—when fixes were expensive—teams aligned earlier, with more confidence in what they were building.
Lewis is the first to admit that product wasn’t always his lane. But the combination of deep product understanding, AI-powered insights, and fast prototyping changed the equation—and the shipping paradigm entirely.
“If I need something to just work, I can just get in there and fix it,” he says. “All of this—the new AI tools, the experimentation mindset, the willingness to try things—is not only increasing clock speed, but also improving research and outcomes.”
He comes back to the same idea.
“If you understand the problem, AI can help expedite the learning phase. Whether that’s research, designing, or building.”
Related reading and technology: