How can generative AI help diverse small businesses learn about the procurement process in higher education? A team of CPS graduate students is trying to find out by designing what we’re calling a Diverse Supplier Chatbot.

Why design a chatbot for this purpose? Currently, many diverse small businesses have no idea what’s involved in selling goods and services to colleges and universities. This is an approximately 170-billion-dollar market, but the procurement process is complex. For a typical small business, with limited knowledge and resources, it can seem like a black box.

As Saloni Jain, who is studying analytics and statistical modeling, observed: “This experience has been a deep dive into the realm of supplier diversity initiatives, something I wasn’t very familiar with before working on this project. It was eye-opening to see how the Lab’s initiatives can create symbiotic relationships between small businesses and educational institutions. The fact that our lab is actively facilitating these connections, particularly within the New England area, has added an extra layer of meaning to the work we’re doing”.

So, why not tap the power of generative AI to create a practical, easy-to-use chatbot to answer questions like these?

  • How can I get certified as a diverse supplier?
  • What are the basics of the procurement process in higher education?
  • Whom should I contact to learn more about procurement opportunities at a specific college or university?

Building on the work done by a team of Informatics capstone students in the spring of 2023, an interdisciplinary group of College of Professional Studies students tackled this challenge over the summer as a co-curricular project.

Designing a chatbot has multiple components. The team learned that building the components of the system is one thing, but connecting them into an operational prototype is quite a different challenge. For example, how would a small business owner engage with the chatbot? As a team member Nidhi Bhalodia, an interactive design student, noted: “The user research and exploration of various diverse supplier programs helped us map the pain points of the suppliers and address them through a structured conversation question flow. We had to define a user journey aiming to connect suppliers with the educational institutions they might want to do business with and guide them through the process.”

Other challenges included creating a local knowledge base of relevant, accurate, and up-to-date data and understanding the capabilities of a new technology. “As for the technological aspect, working with the ChatGPT API was a fascinating experience,” noted Saloni Jain, “Figuring out the ‘right questions’ to ask was critical. This balance between technological sophistication and user accessibility was a rewarding puzzle to solve.” Arjun Malarmannan, a project management student, added: “With Bing integrations in ChatGPT getting rolled out, this project has got a huge potential in the future.”

In reflecting on her experience as the team’s project manager, Jessica Gautama commented that “I worked in aligning the wants and needs of different stakeholders to the team’s capabilities and driving the team to a successful completion. Overseeing this project enriched my understanding of the complexities involved in designing an AI chatbot and my personal development in collaborating with team members from diverse cultures.”

Another team of students will aim to test a prototype of the chatbot during the fall quarter. Follow updates on this project and other news and initiatives on the Lab’s LinkedIn page – no prompts necessary!

Student team members: Nidhi Bhalodia, Akshat Bhatt, Jessica Gautama, Yesha Gosaliya, Saloni Jain, Arjun Malarmannan, Xuan Tang, Jin Xu, Yueying Zhou

Faculty mentors: Professors Xiaomu Zhou and Youngbok Ryu

Student/alumni consultants: Katharine Lee, Renee Rosenblum, Ankita Shinde