Following on from the AI and Me panel session at the Oxford Digital Festival 2024, we’ll be providing further use cases and inspiration for the generative AI tools available here at Oxford through the AI and Machine Learning Competency Centre. Members of the Generative AI Enterprise pilot have been telling us about how they’ve been using ChatGPT Enterprise and Copilot for Microsoft 365 to improve and enhance their work. In this article we’ll be looking at how generative AI has been improving colleagues’ productivity and efficiency in the workplace.
Colleagues in the pilot have been using ChatGPT and Copilot to great effect for streamlining various tasks, saving time and effort across the board. Margo Socha, Senior Programme Teaching Materials Manager from the Saïd Business School, has found ChatGPT particularly useful for generating course images to add to course cards and banners in their Virtual Learning Environment (VLE), Canvas. On how she went about creating these images, Margo has said:
Writing prompts for this task has been a bit of a learning curve and a 'trial and error' process, but with some patience, the results have been impressive.
Margo has also been using the tool to help with transcribing infographics and images, a task which used to be done manually and take considerable time. She neatly summarised the use of AI with:
Whether it's for creative work, technical troubleshooting or writing assistance, AI powered tools have quickly become an integral part of how we operate, making a noticeable impact on productivity and efficiency. I can see this becoming an even more valuable tool as we continue to explore its full range of capabilities.
Becky Gainsford has also been saving valuable time using ChatGPT in her role as Project Archivist, Metadata Support in the Special Collections team at the Bodleian Libraries. Some of Becky’s work involves digitising old, paper-based catalogues so these can be made discoverable on the Bodleian Archives and Manuscripts website. Becky has used ChatGPT to automate these long-winded and repetitive tasks by asking it to create python scripts. These would help with adding barcodes to catalogues as well as ensuring that dates are correctly formatted.
While Becky could have written any of these python scripts manually, ChatGPT was able to do in seconds what would have taken her a couple of hours. She does note however:
The scripts always need manual modifications and bug fixing but the time saving is still large… Prompts require careful thought, and I would always treat results with caution and fully test anything produced. To assume that it is always correct would be dangerous but it’s a great tool with potential for producing real efficiencies in my role.
Copilot for Microsoft 365 has also been tested in the Generative AI Enterprise pilot. Alex Green, from the Department of Physiology, Anatomy and Genetics, has been using Copilot to navigate a clunky request tracker currently being used as the divisional ticketing system. Alex uses Copilot daily to format search queries for tickets through an Outlook folder where he filters all his tickets. This would then link to emails which Alex can summarise using Copilot to provide him with a clear understanding of the ticket without having to navigate through the messy formatting and sometimes lengthy text. Copilot was also able to help create ticket response templates by referencing internal documentation and previous emails to make the process much more efficient, allowing consistency in responses.
These certainly aren’t exhaustive uses for either tool but will hopefully provide some inspiration to help you get started in your generative AI journey. Licences for Copilot for Microsoft 365 and ChatGPT Edu are now available for purchase through the AI and ML Competency Centre.
