Know Your People. Make It Fun. The Rest Follows.
How Method Built an AI platform that +85% of the Firm Actually Uses
I recently sat down with Jake Donaldson, Managing Principal of Method Architecture, and semi-finalist for the AEC Innovators Award, to talk about Modus, the custom AI platform his firm built in-house. Most AI rollouts in our industry die at adoption. Method’s is running at 85 percent, with 100 percent in sight by the fall. That alone made the conversation worth having.
Check out the other semi-finalists for the AEC Innovators Award and vote on your pick for the finalists at https://kpreddy.co/aec-innovators
The three things that stayed with me.
1. Adoption is earned, not mandated.
Method did not pick a tool and push it down the org chart. They surveyed staff at every level, leaned on the junior designers closest to the repetitive work, and built for what they heard. It is great customer discovery pointed inward. When a tool solves a problem someone actually has, you do not have to sell it. Adoption stops being a campaign and becomes a byproduct.
2. Kill what’s no longer serving you, and don’t mourn it.
Not everything Method built survived. As the models improved, some early tools became redundant, and rather than defend them out of attachment, they audit usage regularly and retire what no longer earns its keep. Learning to sunset your own creations is one of the more meaningful disciplines an innovator can build. A portfolio you never prune is not a portfolio. It is a graveyard you are still paying rent on.
3. The real payoff is human.
This is the part I loved most: Joy is a feature, not a distraction. Method designed Modus to have personality on purpose. It ships with a color changer and a “quack mode” that sets off rainbow dances and rains icons across the screen, completely impractical and completely the point. They built it to feel like a helper with character, and to remind people that this work is supposed to be fun. As the tools took the pressure off, that spirit spread beyond the software: people stopped just going through the motions and started actually enjoying the day again. Playful tools get used. And a firm where people are having fun is a firm doing better work.
And for anyone keeping score, the business case holds. Method measures value in time saved, roughly five-plus headcount of new capacity this year, and they are pointing that new capacity at growing the firm rather than cutting the team.
None of this was ever really about the software. It was about the people using it.
The full conversation with Jake is below (edited for length).
In this interview, Jake Donaldson — Managing Principal of Method Architecture — talks with KP Reddy Co about Modus, the custom AI assistant platform his firm built in-house. Jake explains why Method chose to build rather than buy, how the tool’s playful, branded personality helped drive an 85% adoption rate, and how the firm measures ROI through time savings rather than headcount reduction. He also discusses lessons learned, the evolving relationship with clients as AI becomes part of the design process, and advice for firms just starting their own AI journeys.
Walk us through how you all decided to build Modus. What got you guys to build something internally as opposed to buying an existing capability set? Why did you build it custom?
Modus wasn’t our first project — we started with an AI-powered accounts payable automation that connected a bunch of disparate systems. That pilot led us into building Modus. I think the driving force behind building our own custom platform was our culture. We call ourselves an “ego-free architecture firm” — we believe architecture at its best puts people and problem-solving at the heart of things, not the product itself. We didn’t want an off-the-shelf AI tool. We wanted something uniquely “Methoded” that we could brand, customize, and make fit each employee’s workflow — something that matched the cultural identity we’ve built over the last 10 years.
One of the first things I told our technical partner was that it needed a “color changer” — the ability to switch from light to dark to book mode. We also built in “quack mode,” which turns on rainbow dances and raining icons. Totally impractical from a usability standpoint, but a good reminder that yes, it’s AI, yes, it’s another tool, but we want people to bring some joy into their day and remember that architecture is a lot of fun.
What made you decide on a core personality instead of separate tools with their own interfaces?
We wanted to humanize it so it felt like a helper or assistant, like an icon or a little person that’s there — kind of like Clippy back in the old Word days, honestly. People respond better to someone who’s there to help. Hospitality is big in our industry and at Method, so we wanted an assistant persona for people’s day-to-day workflow. Under the hood, it’s really task-driven assistants and agents doing specific things — an array, a toolbox — but we’ve overlaid the fun, consistent Modus identity on top of that.
You have achieved an 85% adoption rate, which is pretty staggering compared to what we’re seeing with other AI rollouts in the industry. What would you tell other people trying to build their own AI tools today about how to achieve similar results?
I want to clarify — 85% today, our goal is 100% by the fall. We did the same thing a lot of companies do first: we just turned on Copilot in Office 365 for everyone. Besides being underwhelming in some specific ways, it just wasn’t getting the job done, and without training and buy-in, it was a tool no one used that we were still paying subscription fees for.
When we designed Modus, we didn’t start with one tool in mind — we talked to staff at all levels, not just top-down. Our junior designers doing the actual drawing work are the closest to the work product and know every annoying, repetitive process. Surveying and collecting data was really helpful, and then building tools from there so that when we roll them out, they’ve been purpose-built to solve a specific need. That makes adoption easy.
The other strategy has been not forcing it. As we build tools and come across staff struggling with a specific problem, we plug them in, they use it, hopefully love it, and word of mouth spreads. We’ve cherry-picked and rolled out organically rather than a corporate-wide mandate, and it’s been very successful.
How have you uncovered the team members who are struggling with unique things you think you can help them with?
Lots of surveys and asking questions. We also have a rigorous quality review process — when we see repetitive errors or big problems in drawings, we’ll talk to the team and do a postmortem: how did we get here, what decisions were made, where are the gaps these tools could close. Staff get excited about it. One of our most successful tools, called Puckman, was built for a very specific client and product type — it might only last a year or two before we sunset it, but because the redlining process was so repetitive and needed to happen in real time, staff loved it immediately. The feedback was: “I can error-check my drawings before my boss sees them.” People inherently want to do a good job, and when you give them tools that help them do it better, it’s a win-win — and that’s where adoption takes care of itself.
Are there any things you’ve built that you’ve retired, or capabilities you planned for that you decided against? What have you learned from that process?
At the beginning, we had a lot more silos between different tools — an HR one, an IT one, an office admin one. Those are helpful and need custom knowledge bases, but they don’t all need to be separate tools, mainly because the frequency people ask questions in each area is so sporadic that it doesn’t make sense to maintain dedicated tools for all of them. We do informal audits every quarter or so. One of the last projects I did with Steps was actually building an analytics and feedback dashboard so staff can flag good or bad responses, and we can see usage trends by tool. If a tool isn’t being used, I’ll talk to a few people, understand why, and we’ll intentionally sunset it without losing sleep over it. We’ve tried to make purging a routine part of our process.
You referred to this as a job satisfaction tool first and a financial one second. How are you guys measuring ROI for the investment put into developing a new platform?
Time is our currency. For everything we’ve built, we try to map out beforehand how long the existing process takes and who’s involved — that’s clarifying for what we build, but it also gives us a clear before-and-after picture. Over the last year or two, we’ve done regular surveys of staff on time savings per task/week using different tools, so we can map hours saved across staff levels and billable rates. We’re on track to have built in roughly five-plus headcount worth of additional capacity by the end of the year through time savings. That’s created an interesting dynamic — our goal was never to replace people, just to eliminate headaches so people can do more of what they love. But then we need somewhere to put that extra capacity, so we’ve been aggressively focused on growth to fill that bandwidth rather than negatively impacting headcount. We’re growing our top line and keeping headcount flat.
Is there anything unexpected that’s come out of the free time people have now — new products, new service lines, anything you didn’t expect?
From a business perspective, our managers get to spend more time with staff, and staff get to spend more time with clients. But maybe more important than the business side: I’m seeing more people feel less stressed about participating in company cultural events. We have a “culture crew” that hosts monthly stress-relief activities, and in the past people were going through the motions because they were so stressed about deadlines. Now that pressure is easing off, and people feel like it’s okay to goof off in the break room or play ping pong — it’s given everyone more margin. We’ve also had more people than expected reach out about this specific product and consulting offerings — that wasn’t really on our radar since we built it for our own purposes.
Is there anything that doesn’t show up in the videos or other assets you shared that you’d like our audience to better understand about what you built?
What’s hard to see from pictures or video is that Modus is continuously evolving. We’re constantly scrubbing which tools are working versus not, and tools that were once separate now sometimes make sense to combine as workflows change. What you don’t see is the back-and-forth with staff — someone realizing “because I can do this in Modus now, it’s changed my workflow in other ways, so I have to rethink other parts of it.” That’s not something we solve within Modus itself, but it has a ripple effect spreading outward that’s been exciting to watch. It’s greatly and positively impacted all of our staff in unpredictable ways.
If you could start all over again, is there anything you would do differently looking back on your experience?
Short answer, no — I’ve learned so much through every failure along the way. But pragmatically, yes, there’s a lot I wish I’d known up front. I wish I’d had more clarity early on about working with an outside partner, and about being more open-minded — with AI and innovative tools, your new workflow doesn’t have to replicate how you did things the old-fashioned way. Sometimes there’s a more elegant technology solution than the manual process, and I wish I’d had that mindset from the start. I was also very obsessed early on with token usage and subscription costs, but now that I’m seeing the returns — including from the investment in technology partners and consulting — I lose zero sleep about doubling and tripling down on that investment each year. Me today would tell me yesterday: don’t sweat the money, focus on the output and impact, and the investment will be worthwhile.
Have you noticed a difference in your relationships with your clients? Are they aware that you all are AI-enabled at this point?
It depends on the situation. A lot of what we started with was back-of-house operational stuff, so clients didn’t necessarily notice a difference. But some of what we’re doing with AI-powered renderings has changed things — in the past, renderings were almost an afterthought, time-consuming and often outsourced at real cost. Now we can produce them in real time, same day, as we’re designing. That’s huge because we can bring those images to the client meeting up front. We see our role as architects as being master translators — trying to get inside the client’s head and extract a vision they can’t fully articulate themselves. Being able to frontload visualization with AI helps us communicate more effectively and bring their project to life. It’s also been interesting to see clients start experimenting with AI themselves — with certain clients we’ve been able to collaborate and communicate about that, which has evolved the relationship in positive ways since we share the same goal of producing great quality work.