Searching for Better AI Returns
Students in mpd² are collaborating with consulting firm CGI to help turn AI adoption in businesses into an economic growth engine.
Economic data from the AI arms race paints a bleak picture.
US companies spent more than $200 billion on AI in 2025 and are expected to add more than half a trillion dollars to that in 2026. But a growing number of prominent economists are calculating that the returns are much less than forecasted.
Some estimates projected AI investment being responsible for anywhere from 39 to 92 percent of gains to the US economy, according to The Washington Post. That same article explains some economists think those investments may not have impacted economic growth at all.
That’s where students from Northwestern Engineering's Master of Product Design and Development Management (mpd²) program come in. A team of students is part of a Thesis project sponsored by consulting firm CGI and Mike Hyzy, its vice president of AI strategy and product development. The goal is to help move AI from cool gadget to cooperative colleague at the core of an economic growth engine.
“Getting people to use AI is only half the problem,” Mike said. “Heavy AI use can actually dull critical thinking. People start outsourcing their judgment to the tool. They stop doing the hard cognitive work that made them valuable in the first place.”
The students are diving deep into a system Mike created for CGI called A3F—the AI Adoption Acceleration Framework. A3F’s goal is to address why most organizations struggle to get people to effectively use the AI capabilities they invest in.
The word “effectively” is the key, Mike said, because while adoption is a problem, it isn’t the one currently leading to such small returns on investments.
When organizations started rolling out AI, they followed the traditional change management playbook and quickly learned it didn’t work, he said. Usage would spike after a launch, then drop off as people returned to familiar workflows.
Mike said it's important to understand the root causes of why people won't try AI, then tap into their motivations to help them become AI native.
A3F helped address that. But gaining adoption created a new problem—employees who lean too heavily on AI and lose critical-thinking skills.
“We're evolving the approach,” he said. “The goal isn't just adoption anymore. It's teaching people to use AI in a way that elevates their thinking rather than replacing it. We want AI as a thought partner, not a crutch.”
The mpd² students are helping pick out blindspots Mike and his CGI colleagues may have developed in A3F over time.
“Bringing in a team that had no history with the framework, no institutional knowledge, no reason to be polite about what wasn't working, that was the point,” Mike said. “I wanted them to find what we couldn't see.”
And that is exactly what they are doing.
Students spent the first quarter of the three-quarter project delivering journey mapping, competitive analysis, persona development, market sizing, and design requirements for the next iteration of A3F.
The students spent their second quarter in the design phase, translating their first quarter insights into solutions.
“A3F isn't a classroom exercise. It's a framework we're deploying across thousands of users, with real adoption data, real friction points, real stakes,” Mike said. “The students aren't working on a hypothetical; they are working on something that will ship.”
That makes the CGI/mpd² partnership a highly desired win-win, Mike said. CGI benefits from the fresh outside perspective, while mpd² students earn practical experience in using critical-thinking skills to help mold the AI-driven future.
That experience is desperately needed right now, Mike said.
“What separates people is critical thinking—the ability to identify a problem worth solving, break it down, figure out where AI adds value and where it doesn't, and ship something that works,” he said. “If you can show that progression, from messy problem to working solution, you'll stand out.”