Atlas the Chatbot

Atlas the Chatbot
Project Type: 
Academic Technology
Data Science
Web Development
Blended Learning
Teaching Tools
Resource Centralization
Faculty, TAs, Students, Staff
Mae Bethel

Conversational AI to guide faculty and TAs in improving learning experiences


Faculty and TAs increasingly seek out practices and tools for tech-enhanced learning. While these resources exist on sites like the Teaching Resources webpage, finding the right resource – and navigating discovery questions along the way – can pose an unexpected challenge.


The initial vision for the chatbot was spurred by conversations with instructional teams who wanted on-demand support to identify teaching tools and pedagogical approaches. As a first step, the Digital Learning Solutions team developed curated lists of resources through blended learning research, which was formulated into an Airtable database. The team then mapped out lines of conversation through which an instructional designer would guide instructors towards the right resource. This conversation was externalized into dozens of pathways within the back-end of Dialog Flow, a conversational AI interface. Various technologies were added to the stack, most significantly Bot Copy, a customizable UI that guided users down a decision tree based upon typical discovery patterns.

The team performed user testing to determine best pathways over time. Once a classic decision tree was mapped out, team members also trained the natural language processing AI to recognize and classify additional terms unique to questions about learning outcomes and classroom strategies.

Finally, the team deployed the app on the Teaching Resources website, integrating it with the underlying WordPress code base. Additional user testing was performed to optimize patterns for identifying resources for abstract ideas such as improving attendance or promoting engagement.

Technologies used 

Dialog Flow
Bot Copy
Google Assistant
Google Analytics


The Digital Learning Solutions team created a chatbot that guides faculty and other teaching staff in identifying teaching tools. Through the conversations, the team tracks the most common questions and results. This data leads to novel pathways for discovering new tools and predicting nascent requests for innovative tools. As a result, the team can track and target needs based on the data generated from the chatbot.

The chatbot has also been featured in several presentations such as the ATXpo, Stanford Academic Technology Communities of Practices Showcase, and Pandemic Pedagogy Research Symposium. It has generated interest in other schools at Stanford for a variety of use cases, as the team has consulted with other schools to guide their thinking in novel uses of conversational AI.

The welcome message will put the user on the path, saving time and keeping the user engaged.
An example of a resource page with hyperlinks for adopting tools or pedagogical approaches.