It’s no secret that ChatGPT and other Generative AI tools have taken the world by storm, providing time savings and increased efficiency. Generative AI is a subset of artificial intelligence that generates text, images, videos, and audio in response to user prompts and questions. AI programs like ChatGPT use deep learning techniques to find patterns in data and replicate those to generate content. This can provide benefits across a wide range of business functions, and one area where we’re seeing major impacts is during large-scale system implementations. Today, we outline four recommendations for effectively using ChatGPT during system implementations. 

Leveraging ChatGPT for System Implementations

1. Writing Assistance + Idea Generation 

ChatGPT can provide writing assistance in many ways – from making text more concise, improving verbiage, or drafting an email. It can help summarize meeting minutes and notes into bullet points, which is particularly useful during requirements-gathering meetings. You may also use it to simplify training materials to ensure they are user-friendly.  

If you are struggling to begin your task, getting ideas from ChatGPT may give you the spark and starting point you need to overcome writer’s block. We have seen our clients use this strategy during the project planning phase to determine what is in and out of scope or for brainstorming Key Performance Indicators (KPIs) that they should keep in mind. When using ChatGPT in this way, it’s a good idea to get justification for the recommendations by asking it to also choose which KPIs are best for the given scenario or provide any cons or gaps in the KPIs generated. This will help to mitigate any potential hallucinations.

2. Research & Data Analysis  

System implementations are a great opportunity to identify and implement efficiencies and increase business performance. However, finding and being able to access a wide range of data is often a challenge when trying to identify trends.  

Let’s say that you are doing research on company performance over a given amount of time or comparing businesses in a certain field. ChatGPT can parse through its training data and provide immediate results. It can also assist with data analysis by providing recommendations on how to normalize data, listing visualization techniques that can best portray given data or processing data using specified tools like SWOT or Risk Matrixes.  

While ChatGPT can perform a certain level of statistical analysis, it’s important to confirm that the output is correct. Make sure you ask it to cite sources and provide any math or code used so it can be reviewed by an expert on your team.   

3. Understanding or Generating Excel Formulas or Programming Code  

ChatGPT is very efficient at creating Excel formulas and generating programming code. Submitting prompts like “Give me an Excel formula that does …” or “Generate PowerShell programming code for …” will reduce the time needed for online research and programming trial and error. Excel is often used during data migration, so ChatGPT can be helpful with field mapping, data conversion, and report creation.   

ChatGPT can also assist with troubleshooting formulas and code that you write. Simply load your formula or program into the tool, and ChatGPT will automatically resolve the issues for you. If you are using Excel, Copilot is another generative AI solution that integrates directly into Excel to provide similar functionality. 

4. Summarizing Digital and Image Inputs 

Tools like ChatGPT can summarize document and visual images into high level and/or specific text descriptions. For instance, reading a large PDF file of a system administration manual can be an arduous task that makes finding answers to specific questions difficult. Instead, you can upload a PDF of the manual into ChatGPT, then submit your specific questions to speed up system design and troubleshooting. Make sure to provide all necessary context about the files to get the best results.  

Going Forward 

While generative AI tools provide value for both personal and business use, there are some best practices to use to ensure you get the best results in a compliant way. Make sure you’re providing clear input and use context. If you don’t get the results you want, ask the question in a different way with more specific parameters like you would in a conversation.  It’s also imperative to avoid sharing confidential organizational information. As always, generated content is meant to be a starting point. Be sure to add your personal touch and assess the results to make it your own. 

ChatGPT and other Generative AI tools can be very beneficial across numerous business functions when used with clear intention and care. If you need support with your Generative AI strategy and plan for your business, the Data Analytics team with Clarkston Consulting can help.  

Contributions by Patrice Freeland