The reason why I am covering use cases as my next topic is first to understand the areas in which these could be looked at and how could I get some (Proof Of Concept) POCs going.

To pick the first one or two POCs to work with, we need to identify use cases that serve our needs. Let’s look at the WHAT first, and then we could go to the HOW.

Here are the most commonly looked at functional areas as the first ones to leverage Gen AI

Customer Service

Enhancing Question Answer / Knowledge Applications by replacing automated Chat Robots that are keywords based to Gen AI chat that will be more humane and, at the same time, be able to understand the context and be very specific to what the customer is looking for. It could also be multi-lingual.

Email Draft Responses is another one that is being considered by many. Pre and Post-sales customer service representatives are the faces of the company at all times.

• How the email is responded to has to be humane and specific to the point in a very objective manner. One might ask, don’t we have templates? Templates probably have canned responses but are not context-based. Mostly, it would be subject-line based.
• The way a human thinks and responds is context-based on the body of the email. Gen AI is the closest that can come to the thought process of a human being and being able to do with massive data sets

Project Management

Automatic Status Report Generation — Our day-to-day work is all running on systems.

• The meetings are on the calendar.
• Timesheets are being filled.
• Technical work output is normally checked into applications
• Non-technical work output is captured in many different ways.
• All our work is captured as data at some place
• That data is fed to the LLM and a draft status report is generated and stored for the user to review and send it to anyone to whom they have to send it to

Code Generation

This is the first thing the technical teams start looking at

• Generating baseline code saving 40% of the time
• Automating unit test cases

Document Generation

• Automating release documents for software based on the change in code base between releases
• Automating End-user guide updates on any changes that are made

Each one of you can come up with so many use cases. My approach is to pick something small and start creating usable functions, deploying and continuing to enhance them. My suggestion to others is the same.

The subsequent write-up in this series will cover the approaches I am using to start building one of the use cases.

🙂