Generating Brand-Agnostic Images for CDW
A team of MSAI and MBAi students collaborated on a unique capstone project for the IT solutions company.
Leaders at information technology solutions company CDW routinely are confronted with a dilemma. If they only have one picture to showcase a certain type of product — like a laptop — how do they choose which brand to portray? Apple? Lenovo? Microsoft?
What if generative artificial intelligence (AI) could be used to create a brand-agnostic image?
Senior digital executive Rick Lockwood posed that question to students in Northwestern's MBAi Program — a joint-degree program offered between Northwestern's Kellogg School of Management and the McCormick School of Engineering — and Northwestern Engineering's Master of Science in Artificial Intelligence (MSAI) program.
D.K. Marsh (MBAi '24) was one of the students asked to answer CDW's question. When he thought about image generation, he immediately thought of DALL-E — an AI-system from OpenAI that uses text prompts to create realistic images. Marsh played with the model soon after its launch, but he stopped when he couldn't think of many use-case scenarios for the new tool.
Then he heard about CDW's project.
"CDW had a real problem they could solve with AI image generation," Marsh said. "It opened up a lot more possibilities that I hadn’t thought of, which I found fascinating. Additionally, the project was going to be very 'visual,' which I thought would be more engaging to work on."
Marsh collaborated with four other students — two from MBAi and two from MSAI — to answer CDW's question. The foursome was one of 17 teams of MBAi and MSAI students to work on a client-facing problem for their capstone project. The team participated in a poster session at the Capstone Showcase and was named co-winner along with a group that worked with investment management firm Vanguard.
Image creation was — and continues to be — a new use-case for generative AI, and when the team began the project, updates or new models launched on what seemed like a daily basis.
"The biggest challenge for us in the beginning was that most models just weren't good enough to generate realistic enough images for the use-case, and the only one that was was quite hard to set up and work with," Kushal Agrawal (MSAI '23) said. "Fortunately, DALL-E 3 was released just in time for us to pivot to it."
That pivot came halfway through the quarter and was a defining moment for Lockwood, group product manager and head of digital merchandising. To him, it epitomized why he enjoys working with the MBAi and MSAI students.
"The way this space is moving so quickly, those are changes that are hard to stay on top of," Lockwood said. "It was so cool to see that change and adjust and hear the students say, 'Hey, here's how we can improve, even though we're halfway through.' I think that opened a lot of eyes (at CDW)."
The late shift taught the students a valuable lesson as well — one Marsh thinks many people forget when it comes to innovation.
"AI is changing fast and it’s impossible to pick the best horse in the race at this point," he said. "There needs to be a balance between picking the best AI, and picking the best solution that works for the business. We ended up picking DALL-E over other options, not necessarily because it was the best image quality, but because its quality was satisfactory, it was the fastest option to build a solution around, and it was easiest to be maintained by CDW once we were done."
A major discussion with the students was about AI not replacing employees, but instead creating systems to help those employees work more effectively.
That theme is a constant one stressed through the MSAI and MBAi programs.
"While AI is poised to transform many traditional roles, it is imperative that we prioritize ethical and moral considerations at every step of AI development and use," MSAI deputy director Mohammed Alam said. "We strive to educate our students on the significance of responsible AI, ensuring they understand that AI technology should enhance, not replace, human creativity and labor. It's about augmenting the human experience, not diminishing the value of professionals. This balanced approach respects existing professions while making possible solutions and advancements never possible before."
Lockwood was impressed by the students' maturity during their project. This was the second year he worked with MBAi and MSAI students for their capstone.
He hopes it won't be the last.
"These programs have been great to dedicate time to innovation, and I think these students in particular are very much on the leading edge of some of the latest technologies," he said. "We offer the option to come in and present … and the folks they're presenting to are very high up in the organization. I've never had any reservations about putting one of these students in front of our executives."
Karie Kuiper, senior product manager at CDW, agreed.
"The students’ findings really opened a lot of doors for us here at CDW when it comes to working with AI and helped illuminate which areas of our digital experience may benefit most from generative AI specifically," she said. "They helped us lay a great foundation that we’re continuing to explore."