EdPlus AI Project Hub

Business problem we are solving for

The success centers for ASUO are often inundated with student inquiries. The high volume of calls and the breadth of the questions, ranging from course schedules, admission requirements, student benefits, fee structures to technical issues, require an extensive amount of time and resources to address. We aim to improve the efficiency, response time, and resource allocation in our student support call centers by integrating an AI copilot that can handle inquiries, aggregate resources, and provide scripted responses to keep the conversation flowing.

Summary of project

This project involves building a proof-of-concept (POC) AI Copilot for university student support call centers. The AI copilot will be designed to understand, retrieve, and present relevant information from multiple tools and databases used by the call centers. The copilot will also have the capacity to generate a scripted response that the support coach can use or modify to answer the student's queries efficiently. We aim to collect data on the AI model's performance, identify areas of risk, and answer high-level questions before releasing the product to the public.


Testing Feasibility: A POC allows us to test the feasibility of our idea on a small scale before committing substantial resources to a larger project.
Identifying Risks Early: By starting with a POC, we can identify and mitigate potential risks or challenges early in the development process.
Ensuring Effective Use of AI: Not all problems need AI, and a POC allows us to ascertain that our use of AI will add value and is the right solution for this issue.
Understanding User Interaction: A POC will help us understand how our users (both students and support staff) interact with the AI CoPilot, and what improvements may be necessary to ensure optimal user experience. ​Refining the Model and Optimizing our Data: The data we gather during the POC stage will help us to train and refine our AI model to ensure it delivers accurate, relevant responses.
Evaluating Integration Challenges: The POC allows us to evaluate and address potential integration challenges with existing call center tools and systems.
Gaining Stakeholder Buy-in: A successful POC can be a powerful tool to demonstrate the potential of the AI CoPilot to stakeholders, helping to gain buy-in for a full-fledged implementation.
Iterative Learning and Improvement: The POC will allow us to adopt an agile, iterative approach to learning and improvement, tweaking the model and its implementation based on real-world feedback and performance.

Primary Features:

Aggregated search across all coaching tools and resources ​The Copilot will provide a centralized search for a number of different resources and databases available to the coaches Response suggestions for coaches ​Using the content provided by the coach, the co pilot is capable of providing a short script for the coach to follow or modify that answers the students question and keeps the conversation flowing. ​Automated Follow Up Questions ​Depending on the questions and resources being searched for within the copilot, we can recommend additional questions that may be helpful to a given conversation.

Scope of project

The project will include the following stages:
Requirements Gathering
Understand and document the needs of the call center and the types of queries the AI needs to handle.
Data Aggregation and Preprocessing
Aggregate relevant data from various tools and databases, and clean and preprocess the data for training the AI.
Model Development
Design and train the AI model to understand queries and generate appropriate responses.
Integrate the AI model with the call center system for a seamless handoff of tasks.
Testing and Evaluation:
Conduct tests and evaluate the AI's performance.
Deployment and Monitoring
Deploy the model to the call center and monitor its performance, iterating and improving the model as necessary.
There are no rows in this table

Resources/teams involved

Project Manager: Responsible for overall project coordination, communication, and timeline adherence.
Product Owner: A project stakeholder who will act as the voice of the user and make decisions about the product direction.
AI/ML Engineers: To build, train, and test the AI model.
Software Developers: For system integration, API management, and front-end interaction.
QA & UX Research: To conduct rigorous testing and ensure the AI model meets all specifications.
Support Staff and Coaches: To provide expert advice on call center operations and end-user experience, and validate the relevance and effectiveness of AI responses.
Coaching Copilot Tasks

Conceptual Mockup:

Data Sources 2
Connection Type
ASU Websites
CSV Upload
CSV Upload
CSV Upload
CSV Upload
CSV Upload
Ross Early
CSV Upload
There are no rows in this table

Quark Data set:


Additional Considerations

Data Privacy and Security: As the system will have access to sensitive information, robust data privacy and security measures are critical.
User Training: Training for support staff and coaches on how to use the AI Copilot and understand its limitations.
Continuous Improvement: Collecting and analyzing data on the AI's performance post-deployment for continuous improvement.
Scalability: The system should be built with scalability in mind to handle

Next Steps

From POC to Phase 1

Build data pipeline into crucial copilot data sources
Migrate from Zip of Guru data to live API Access
Migrate from scraped website data to automated content export
Migrate from GPT3.5 to Google Vertex
Build responsive UI
Implementing auto tagging AI into data sources
Standardizing format for data sources

Want to print your doc?
This is not the way.
Try clicking the ⋯ next to your doc name or using a keyboard shortcut (
) instead.