<span class='p-name'>CollaboraThink: Using AI to help people brainstorm and make decisions</span>

CollaboraThink: Using AI to help people brainstorm and make decisions

This summer, I’ve been involved in a hackathon organized by Carnegie Mellon University (CMU) and the University of Tartu with three areas of focus (Medicine and Health, Education and the Future of Work, Finance and Economics).

For those that do not frequent a hackathon, it is often a realm where developers (and sometimes educators) connect in a dynamic fusion of innovation and knowledge exchange. These events serve as thrilling crucibles where participants from diverse backgrounds converge to ideate, collaborate, and create transformative solutions to real-world challenges.

From my experience, in these events, developers and educators team up to innovate educational solutions in a time-bound sprint. Their creativity, coding, domain expertise, and insight drive them late-night sessions, brainstorming, and prototyping. This synergy enables the development of cutting-edge tools, platforms, and applications to reshape the learning landscape. Hackathons foster teamwork, problem-solving, and dedication, leaving a lasting impact on both participants and education.

In this post I’ll share some of my insights and ideas as I worked through the Education and Future of Work track.

Initial Pitch

To begin the hackathon, participants needed to introduce themselves to the group with an initial pitch to indicate an area they’d like to explore with others.

For my initial pitch, I was inspired by the Generative Agents: Interactive Simulacra of Human Behavior work by Stanford and Google.

The researchers created generative agents that can simulate human behavior. These agents can be applied to gaming, virtual world environments, and more, with the ability to form opinions, initiate conversations, build relationships, and coordinate group activities autonomously. The researchers developed a new framework using a large language model, allowing the agents to store experiences in natural language, synthesize memories into reflections, and utilize these to plan behavior.

My idea was to create a virtual environment where educators could practice interacting with students. This could be used in teacher development to mimic learners in a learning environment.

Develop an AI community to serve as a situated, simulated learning environment or assessment tool. 


After the initial pitch, I connected with teamed up with Ikram Shah, a graduate student at CMU and we agreed to modify his Ideastorm chat platform to suit these needs. Ideastorm is an AI-powered chat platform designed to spark creativity and generate fresh ideas.

CollaborThink is designed to empower individuals in brainstorming and decision-making processes. 

By leveraging the capabilities of AI technology and data source integrations through Langchain, it offers natural language understanding, creative suggestions and decision analysis to enhance work-related decision-making. Langchain is a framework built around large language models (LLMs) that can be used for chatbots, Generative Question-Answering (GQA), summarization, and much more.

CollaborThink’s flexible and customizable nature ensures it adapts to diverse use cases and user preferences, making it an invaluable asset for individuals seeking efficient and effective decision support.

Situated simulation-based learning environments like this allow humans to learn, interact with virtual agents, and role play the knowledge, skills, and discourse practices they’ll need to utilize as they interact with other humans in pursuit of their goals.

CollaboraThink Use Cases

CollaboraThink allows individuals to share ideas with a virtual partner, and receive (sometimes critical) feedback on work process. We identified the following possible use cases.

  • Use case 1: A K-12 student is working on a wicked problem in a classroom (climate change) and they have a project to complete. They can work with CollaboraThink to understand and complete the project. The tool can engage in dialogue with the learner about what they know, need to know, and give feedback on the ultimate product.
  • Use case 2: A classroom teacher is building a lesson to teach to a group of students. The educator can engage in discussion to better understand the ways in which various individuals with varying backgrounds will understand, question, and pushback on materials presented and how learners will process this information. 
  • Use case 3: A business professional is preparing to pitch an idea to a group of investors. CollaboraThink can help prepare the professional as they pitch the initial idea, perhaps work across different languages and cultures. The professional can also include an agent that will critique their work and allow for “red teaming” the idea before it goes to the actual pitch. 

Wrapping Up

Ultimately, Ikram and I were not selected to move forward as one of the five teams that moved to the final round of the hackathon. I valued the opportunity to interact with the other members of the hackathon, and hopefully we can bring this tool to the market.

If you’d like to stay on top of areas like this, you should be reading my weekly newsletter. You can follow here or on Substack.

Leave A Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.