The Product Team Diet Plan: Is AI Eating Your Roadmap?
The future belongs to the builders — and the builders use no-code.
Jacob Klug, the co-founder of Lovable’s largest agency partner, Creme Digital, is determined not to go to university.
“I was always the kid doing something in the back of class. In high school, I [tried] so many different businesses… People always knew me as that kid.”
During the COVID-19 lockdowns, while still in high school, he noticed local businesses struggling to transition to digital service delivery. Responding to this need, he built his first app for a client using the popular low-code tool, Bubble, and the rest was history.
“I'm not super technical. I definitely couldn't build [a web app with code], but I was familiar with Bubble, and so we kind of built the first tool for our client. And that was a very… awakening moment.”
Klug and his cofounder launched Creme Digital when he was in Grade 12 with the goal of undercutting larger product development firms on price and time-to-market.
When the duo discovered Lovable in 2023, they were so taken with the chatbot-style development tool that they pivoted their agency away from Bubble and embraced what’s become popularized in 2025 as “vibe coding.”
Except, it’s a whole lot more than just vibes. Klug’s team primarily uses Lovable, Bolt, Superbase, and Cursor to deliver client projects, and he now has four developers working for him.
"Our positioning in the market was, how do we be the most high-end version of no-code tools?"
“We're now Lovable's biggest agency. And, yeah, we basically, in the last few months, have just been building only on Lovable and Bolt.”
As he tells his story, the 21-year-old looks like a seasoned startup founder—sitting at what looks like a minimalist desk, wearing a plain t-shirt, with a podcast microphone hanging by his head.
His eyes dart between our Google Meet window and the incoming messages on his screen as he explains to me that Creme Digital scaled to over “a couple million,” pretty quickly.
The company pivoted from Bubble to Lovable approximately five months ago, said Klug. “It's been going, obviously, pretty well.”
So well, in fact, that Klug’s parents have finally agreed to take university off the table.
He doesn’t see things slowing down—in fact, he sees the trend of AI tools in product delivery accelerating to favour cross-functional, delivery-oriented professionals as opposed to individuals with deep skill sets.
He adds that he sees the roles of the product manager, designer, and developer combining into one and new roles, including the prompt engineer and the “design engineer,” emerging.
"You need somebody who has the visual abilities and taste, but also is good at prompting, and also has enough technical requirements to get things over the line."
Klug describes a vision for the future that’s exciting to some and terrifying to others. However, not everyone agrees.
Leaders think no-code is like a junior PM: young and brilliant, but needs supervision.
As my former manager, James Rubec, populates the Google Meet window on my MacBook Air, I note that he’s taking this call from a different location than his usual, sunny office situated in his family home.
It’s Easter Monday today, which is an official holiday in much of Canada and an unofficial holiday throughout the rest, and Rubec is kind enough to take this call with me on the final day of his long weekend.
Having known Rubec for many years now, I’m not surprised to see his sweet dog squirming about behind his chair or that he opens our conversation by telling me about the egg hunt he and his kids enjoyed earlier that day.
Rubec is my mentor, my friend, my former manager, and was actually the first person I thought of when I began researching the impact of AI and Automation on the product management role.
Here’s why—Rubec is a strong advocate for the role being one of a bridge builder, team unblocker and data cruncher, just as much if not more than the traditional responsibilities of feedback collection, feature ideation and backlog prioritization.
Rubec’s process-driven style has rubbed off on me, and to this day, I still approach the role with the same “whatever it takes” attitude, which is how I often find myself automating sales processes just to mine them for data, working on press releases for feature launches and creating marketing collateral for upcoming events. I imagine he still does the same.
If he’s right, his favorite parts of the job are about to become the core of the job.
Today, Rubec, who has been the head of product at the media intelligence company Fullintel since 2022, is using AI to do many of the things he used to ask me to work on way back when—and he’s doing them much faster than I ever could have delivered them.
One of the primary ways he leverages ChatGPT is by synthesizing structured and unstructured customer feedback into datasets used to make business decisions.
“I will take my verbatim customer feedback transcripts from calls, and I will dump them into a singular prompt. I can then just ask my chatbot what [customers have] said about this feature set, and it will report back verbatim and give me all of that data,” he described.
Rubec and I share a passion for process automation—especially around customer feedback loops—and as he speaks, I can sense his excitement about his newfound superpowers.
According to Rubec, the product development team of the future will consist of, “one senior product manager, one senior developer, devops, and devops support, one developer, one PM—and they kick ass and take names all day.”
What about the image of the future that Klug so eloquently painted? Well, it makes Rubec nervous because he’s been around the block a few times and knows what happens when you can’t answer the question “why did that break?”
I will not put myself in a position where I have to explain to someone why something does or does not work and say it's because the AI did it.
James Rubec, Head of Product, Fullintel
Is the agentic army for us? Maybe, but it’s going to be ok.
Is there a middle ground between scaling “to infinity and beyond” and simply letting AI run quietly in the background to enhance each contributor’s abilities?
Erun Fernando, Director of Product Management at the marketing technology company, Cision, has been experimenting with AI tools both at work and on his own time.
He predicts that the future will consist of a handful of strategic humans guiding a team of AI agents towards delivering an output. He uses the classic PRD as an example.
“I can [now] spin out a PRD in the car while driving somewhere, right? I literally just throw on [ChatGPT] voice transcription, and I'll blurt out. I'll talk [for] maybe 15 minutes…and then get something really awesome out of it. Or, at least a draft."
He continues, saying that he sees product managers moving towards strategic functions as opposed to execution, which means that large companies probably won’t need as many of them.
All these positions are moving a little bit more towards directing a workforce of AI bots, more than ‘I'm just going to go in and just hack this Excel sheet myself,” he explained.
“I am putting [my bet] on the fact that ultra lean teams are going to be a significant force in the next five years of SaaS, where what you often were doing with teams of, say, 20 to 70, maybe you can do with a team of three to 10.”
At this point in our conversation, the question on my mind is the one on everyone’s mind.
What happens to those of us who are no longer needed, or who are never offered a job to begin with because the labour market needs have decreased so dramatically?
Don’t fret, fellow PMs. The answer might lie in the two-person business model.
“I think it will be the kind of world where five people making $2 million [per year] is a very stable and very exciting space to exist,” Fernando reassures.
Brock Rowlands is the Director of Product Development at TheAppLab, as well as an educator and thought leader with The Data Integrators.
He agrees with the assessment that AI tools have made running a lifestyle business—think an agentic recipe builder or a closed marketplace for sports fans to discuss their favourite team and trade high-value cards and memorabilia.
“It lowers the bar to somebody who could be a creator,” he said. “Somebody could build a utility, and that could be their career.”
Rowlands has been in the no-code world before no-code was cool, having worked as the Principal PM on a no-code tool for the manufacturing space in 2020. Regardless of where AI takes us in the future, he feels that understanding core technological concepts will always be valuable.
“I think understanding technology from a systems perspective, so understanding inputs and transformations and outputs, and understanding how computers talk to each other… will always be a currency.”
If that sounds familiar, it’s because we writers said the same thing a few years ago. I'll let you be the judge.
