Demo (recommended to watch in full screen):

Salesforce x KPMG Hackathon Demo.mp4

Takeaways:

<aside> ✅

**This project helped me understand:

-** using multiple user flow diagrams to understand the different use cases the solution can be done (student vs staff)

• understanding NLP to identify which issues are prioritized compared to ensure faster support is given

</aside>

Project Overview

SPARK University is a diverse institution with students, faculty, administrators, and staff, each with unique needs and expectations. However, siloed and outdated systems created inefficiencies, slowed response times, and frustrated users. Our project focused on:

• Improving User Support: Developing an NLP-powered chatbot to provide instant answers to common questions and reduce reliance on manual processes.
• Streamlining Issue Resolution: Enabling faster troubleshooting and support across departments by centralizing information into one accessible platform.
• Enhancing User Experience: Leveraging natural language processing to deliver accurate, conversational responses, improving satisfaction for students, faculty, and staff.

Problem Statement

“SPARK University is a dynamic and diverse institution made up of students, faculty, administrators, and staff —each with unique needs, responsibilities, and service expectations. However, the systems used across departments are often siloed, outdated, or require manual work that slows down response times, creates inefficiencies, and frustrates users.

Your challenge: Design and build an AI-powered virtual agent using Salesforce Agentforce and Einstein AI that improves support experiences for any university stakeholder: whether it's a faculty member seeking help with grading tools, a finance administrator managing reimbursements, or an IT staff er troubleshooting system outages.

Your solution should streamline access to information, automate repetitive tasks, and deliver smarter, faster support experiences across SPARK University.”

Challenges:

1. How can NLP be used to grasp context of an user
2. How can various factors be viewed such as:
   • inventory
   • previous lending history
   • policies (reviewing loan policies)
3. Can existing training guides be integrated to handle simple requests (Ex: password resets)

User Profile:

• As this is an university environment with all ages, we made our tooled for everyone, regardless of age, background, or skillset.