A year ago, AI work at Gapps was mostly two people. Elmo Visuri, our Tech Lead, said at the time he hoped the team would soon have more people in it than just him and Kasper Pöyry.
That's what happened. The AI team now has a Lead Solutions Architect who designs the architecture, a Tech Lead who builds the agents, a Consultant who spent two decades in localization and language technology and now focuses on business AI use cases, a Professional Trainer working on change management, and a Business Unit Lead who has just started.
Ask Elmo what changed in his everyday work, and his answer is pleasantly unglamorous: he no longer has to be part of everything.
Ask Leena Peltomaa, who joined Gapps in August as Senior AI Solutions Consultant, who builds the actual solutions; she doesn't need long.
The actual implementations, and when it's agreed together with the customer that yes, you really need this agent now, then Elmo is the wizard who can build them.
Elmo's most recent build for a customer was a multi-agent architecture. He's also frank about the pace: technically, things are moving so fast it's hard to keep up.
Lauri Raitasuo, Lead Solutions Architect, places himself in the middle.
I draw the slides, which Elmo completely hates, and then I'll do, say, a quick bare-bones version of a note taker in the Google environment.
Lauri has been a coder and an architect, among other things. About ten years ago, he and his team turned recordings into text for a knowledge bank and ended up with their own language model.
We accidentally made our own language model and didn't realize it was some kind of hype thing.
Leena's part is the framing around the work. When a customer asks about use cases for, say, Gemini Enterprise, or how the EU AI Act will impact content creation, she gathers the team's insights for that conversation. The critical task is to understand customers’ core problems and needs and deliver real value.
Atte Front, the team's Professional Trainer and Change Management Expert, designs training programs, leads train-the-trainer initiatives and runs hands-on sessions on all things AI, Gemini and Google Workspace. And Jani Muuronen, the new AI Business Unit Lead, describes his own job in one line:
My job is to help the team shine and to clear obstacles out of the way.
Lauri says that after he started at Gapps, he missed having someone to test and develop ideas with. Leena became that person, and the first proof came quickly.
I was on a fairly tight schedule, for my own reasons, with a presentation to a customer about Gemini Enterprise. Leena took over the preparation, and it went really well. The customer was very satisfied. It was considerably better than what I could have done myself in the same time.
The reason is practical. At the customer, Lauri jumps from one meeting to the next and listens to everyone's worries.
It's good that the other person has a bit of time to focus on what we actually write on these slides.
Leena's route into an AI team is unusual. She has 20+ years in learning, localization, and globalization, 15 of them leading people, and a parallel career as a literary translator with more than 20 published book translations, many of them Swedish crime fiction. Before Gapps, she worked in pre-sales and solutions, designing AI solutions for multilingual communication.
In the interviews for her role, Lauri noticed something early.
Many said that AI saves time, that AI makes things more efficient. But she was the only interviewee who brought up the new business opportunities, that that's what companies should be looking at.
Elmo puts it plainly: a team with more varied backgrounds just gets stronger, and not everyone has to "speak code". He also says Leena has picked up Gemini Enterprise really well.
Leena's first weeks went straight into customer work. She got a foot in the door with one big customer's projects.
Thankfully, I've been given the opportunity to do real customer work already, and talk about the AI stuff. That's been nice.
Much of the team's work right now is turning what they've already done for customers into services the next customer can pick up. Leena and Lauri have been sparring on exactly that: how to sharpen ideas that have been sold or delivered but aren't off-the-shelf products yet.
The clearest example is the AI Adoption Framework. The frame is in place, and we're productizing it step by step.
There's clearly a need for it in the market. Someone to help with: okay, how do we actually adopt AI, what should we do in practical terms, " Leena says
Lauri describes the arc the team works across. It starts with thought leadership, moves into adoption, and ends up in concrete tools and services. "Almost a kind of full-stack thinking, little by little," as he puts it.
And there's real demand for the slow part. Leena points out that there is currently a fair amount of buzz about so-called citizen developers, i.e., the idea that anyone can now be a developer thanks to no-code and low-code technology.
In everyday real life it isn't necessarily that simple. For people who only have experience in chatting with Gemini about something, there's still a big barrier to building even an agent, let alone actual applications.
Lauri's version of the same point comes from a building site. He has built two houses, so he knows what happens when amateurs and professionals use the same tools.
Now everyone has a cordless drill. Drills everywhere, screws everywhere, but that doesn't necessarily mean the result is good. And the tool doesn't give you the ability to judge whether it's good or not.
Leena has data on this. Last year, for her thesis research, she asked 400 marketers how they use AI to produce text. 95% use it for the first draft and for sparring. Very few let anything through without using their own judgment to fix it.
Being able to see what needs fixing, or what makes it good enough for my purpose, that's the professional skill.
Leena expects the team to mature as it grows: roles separating more clearly into specialisms, and enough delivery capacity to run more customer projects at once.
Jani's job is to make that growth repeatable. He started as a software developer in the 90s and moved back and forth between expert and leadership roles. For the last couple of years, he's worked as an entrepreneur and interim leader in growth companies, building his own AI agent tools along the way.
What excites me most is the chance to make AI services repeatable and able to scale, so that growth doesn't require a new person every time.
Interim work taught him one rhythm. He wanted a different one.
In interim work you come in, fix the structure and leave. Now I wanted to build for the longer term and see how far this can be taken.
Hiring still matters, though. When Lauri describes the team that AI adoption work needs, he describes it in people.
It should be a similar kind of package: one Elmo, one Atte, one Leena, maybe one Lauri.
The reason is the work itself. AI adoption gets built into people-based organizations, and someone has to stand between the two.
We know how well AI understands people, so you need an interpreter or a transformer in between. Not a terminator, a transformer.
In practice, the team looks for people who truly understand how AI transforms businesses and who can evaluate and measure the business outcomes of AI adoption.