How are your IT teams currently leveraging GenAI for specific tasks such as code generation, documentation, and troubleshooting? Could you share some examples of successful implementations?
Sort by:
We're not yet utilizing GenAI for code generation or troubleshooting, though our team expressed interest in exploring these areas. Currently, we are leveraging GenAI to overcome the challenge of a blank sheet of paper, helping us draft early-stage documents related to policies, processes, or internal communications. Additionally, GenAI is assisting us with meeting summaries, allowing us to avoid writing minutes from scratch. We're in the early stages, focusing on simple use cases, but we anticipate more exciting developments this year.
Similar to Yvonne, we've deployed around 3,000 licenses of Microsoft Copilot, prioritizing roles that are documentation-heavy, such as project managers and business analysts. The feedback on meeting minutes has been overwhelmingly positive, as it allows participants to focus on the meeting itself rather than note-taking. While we are using GitHub Copilot, I don't have specific details on its application yet. We haven't restricted access to ChatGPT, and I personally find it helpful for getting started on tasks. There's potential for basic automation and RPA, highlighting a gap in our understanding of existing tools like Power Automate. This realization presents an opportunity to upskill our teams and streamline processes.
In our environmental scan while developing our roadmap, we discovered that many tasks don't require GenAI. Basic automation tools like Power Automate can achieve similar results. People often overestimate the necessity of AI, not realizing the capabilities of existing technology. There's a lot of machine learning already integrated into our tools, and the natural language aspect is what's new. Regarding ChatGPT, we haven't blocked it either. We're conducting an experiment involving the use of GenAI models for generating exam questions for doctors. We want to compare the results from public GenAI models like ChatGPT with a private model we're training using medical journals and class notes. This will help us determine whether public tools are as effective as private ones. If residents use GenAI to predict exam questions, they're essentially studying, which is beneficial. We'll have metrics to evaluate the impact.
We're leveraging GenAI (GitHub Copilot, Claude 3.7, Gemin 2.5, etc.) primarily to accelerate proof of concepts and MVP development. It’s significantly reduced the time it takes to turn ideas into testable prototypes, allowing us to get solutions in front of users faster. That speed helps us focus on what matters most: nailing product-market fit through real feedback and iteration.