Lutra’s plunge into AI shows that it isn’t just for developers but for anyone with a curious mind and a willingness to learn
Volaris Group introduced the AI Accelerator in 2026, a program designed to upskill developers, product teams, and business leaders and help companies advance their AI maturity and fluency. This is part of a series about the AI Accelerator and its impacts on our businesses.
Compared to most companies that Volaris Group acquires, Lutra sits in an unusual spot inside the portfolio.
While software is a product that they offer, it is not purely a SaaS company. It is not purely a professional services firm either. The New Zealand-based company employs process engineers who are responsible for what happens inside the fence at water and wastewater treatment plants, from the moment raw, untreated water arrives to the moment clean water leaves.
Their registered engineers solve chemical problems on real infrastructure for clients, including councils, districts, and cities, who in turn serve communities that depend on safe drinking water or treated wastewater every day.
“That’s our profession, our passion, and our mission in life,” says David Romilly, Lutra’s Chief Engineer (pictured here). Their work is mission-critical in the most literal sense.
Putting AI in the hands of self-proclaimed “geeks” who love a problem
When Quentin Griffiths, Lutra’s CEO, brought AI to the team as a company initiative, nobody needed much convincing. Process engineers, David says, are “geeks and nerds who love tech and innovation.”
Lutra held an internal Innovate Day, where they gave the team creative liberty to generate ideas, and used it as a launch pad for the idea that they would bring to the AI Accelerator in Vancouver.
The initial proof-of-concept idea they brought forward was called Project Mussel, a nod to bivalves that filter dirty water to make it clean. The concept is to help local councils better manage the fragmented technical information scattered across documents, reports, memos, and emails tied to a single asset or piece of infrastructure. In this line of work, a pump with ten years of history or a blower with a maintenance record spread across three systems all produce a trail of information. The problem they’re trying to solve is pulling it together quickly enough to matter to the client.
Lutra’s experience at the AI Accelerator
Lutra brought six people to the AI Accelerator in Vancouver: two on the product track, Romilly on the leader track as Chief Engineer, and three on the developer track. The catch was that the team they’d signed up for the developer track weren’t developers. They were process engineers with deep client knowledge, plus one lead data analyst with Python skills. When they arrived, the Volaris Business Transformation team running the AI bootcamp were concerned they would struggle with learning during the week.
As the leader on the team, Romilly was undeterred. He was confident that the AI Accelerator would work for them. His read on why? Process engineers are trained to take a problem and find the most efficient solution. Add deep customer knowledge to that instinct, hand them the right tools, and the team had all the ingredients for success.
We walked in with an idea, but with open minds. We ended up coming up with a completely different product, coded it, and now that’s running as a potential solution in our market.
On top of that, they arrived with no preconceived development approach to unlearn and nothing to stifle their creativity. They simply adopted the framework and started pushing.
And although they went into the program with Project Mussel in mind, it wasn’t the product they ended up with at the end of the week. “We walked in with an idea, but with open minds,” David says. “We ended up coming up with a completely different product, coded it, and now that’s running as a potential solution in our market.”
The product they ended up pursuing was Ripple, a document management tool designed so that information imported from any organization gets filtered back in to update a single trusted source. With their solution, there is no hallucination of outside data and no guessing. They produce a clean record of what information is known and where it came from.
By day two, the Business Transformation team checked in on the team, who had defied their initial expectations. This time, their feedback was more positive: “Amazing work.”
That product was Ripple, a document management tool designed so that information imported from any organization gets filtered back in to update a single trusted source. No hallucination on outside data. No guessing. Just a clean record of what is known and where it came from.
A family of tools taking shape at Lutra
Since their work at the AI Accelerator, the team has been building out what Romilly describes as “siblings” to Mussel. Each one addresses a specific slice of the problem.
One is Coral, a CMMS (computerized maintenance management system) tool being developed in collaboration with a contact in the Pacific Northwest of the US. The scenario Romilly describes for using the tool is as such: an operator walks past a blower and hears an unusual noise. In the current world, that observation might not be captured and could be easily forgotten. With Lutra’s solution, the operator speaks into their phone, the system routes the report to a supervisor, and the AI suggests a likely cause, logs it for the next maintenance cycle, and captures the paperwork that would otherwise fall through the cracks.
Another “sibling product” is Pearl, focused on training and customized document recall for smart operations. It’s a tool David’s team now plans to merge with Coral. And Mussel itself continues to evolve, with Romilly vibe coding a local version alongside his team.
Lutra’s CEO is just as hands-on. Quentin Griffiths is already writing code himself on a product that’s generating significant interest across the water sector. The company’s grassroots adoption of AI has been driven by non-developers, with developers brought in to support rather than lead.
What AI changed for Lutra
Romilly describes a moment that captures the shift for him. In his work, he’s demonstrating to multiple clients that months of work can be done in weeks. He’s unbashful about this claim, and believes the efficiency will promote more engaging conversations, more insightful findings, and that the growth of the industry will benefit more than a single entity or organization.
Working so closely with the client organization gave him something he wouldn’t have had otherwise: direct visibility into what their problems were.
Uncovering untapped capabilities
One detail from Vancouver stayed with Romilly long after the event. The lead data analyst who joined the developer track was someone who sits on a different side of the business. He doesn’t interact with the engineering team day-to-day and had never worked closely with them before. Yet, he fit in immediately.
“I walked away telling the executive team, ‘That person is underutilized,’” Romilly realized. “We need to do more collaboration together across the office.”
Lutra sees AI adoption at an “utopian transformative level” as being gradual and strategic, rather than something that will happen overnight. But Romilly and his team are fast movers, so he’s suggesting to his clients to take risks to reap the benefits that are measurable, tangible, and exciting.
For Lutra, the AI Accelerator wasn’t just an opportunity for them to prove they could create product prototypes. It revealed to them the capabilities that were already there within their people, waiting for the right conditions to show themselves.
This article was adapted from an internal piece by Hanna Ha, Corporate Knowledge Lead at Volaris Group.
