Developing Artificial Intelligence (AI) for business processes

As a member of staff at the University of Essex, you are no more than one internal email away from some serious AI expertise. You could; of course, write that email yourself or get a Large Language Model (LLM) to do it for you. Even if you don’t use AI to write your email, you probably already use it for at least one of the following tasks:

• to summarise reading
• to create dancing cat videos
• or poetry about cats dancing
• to communicate with your energy supplier or your bank
• to plan a presentation or even the first draft of a blog…

It follows that as many of us begin to embed AI into our daily lives, we will ask the question ‘how can I use it at work?’

Mindful of our Information Security Policy (.pdf), this shouldn’t mean pasting data into your nearest LLM, but neither should it follow that we disregard the use of AI to support us with some tasks. We already have guidelines for students in place for the use of AI as it relates to teaching and learning; and we have created new guidance to help you navigate how AI might be used to enable efficiency in your day-to-day work. Key to this guidance are the principles we have set out below.

Integrity: the way we use and incorporate AI must align with our values.

Co-piloting: it’s fine to use AI as a ‘jumping off point’ for example to create a first draft of some meeting notes. But these will then require human intervention and approval. Work initially generated by AI should be reviewed and checked before being circulated.

Accountability: if we make a decision which has been reached with the support of AI then we are accountable for it.

Transparency: we are clear and open about our use of AI.

Information security: we comply with our policy and all associated legal requirements.

Our Steering Group ‘Utilising our Expertise in AI’ brings together Academic and Professional Services colleagues to ensure oversight of this work as the possibilities that AI presents become clearer. Chaired by the Chief Information Officer, this group will consider how we further embed AI to support staff at work in the future.

As a first step, we’re delighted to launch Zoom AI Companion to all staff in January 2025. This can help you to take meeting notes, document actions and summarise discussions. We have been trialling this feature for meetings of Professional Services Group (PSG) since May 2024 and have noted that it reduces the time required to create minutes, which has meant more time on value-added tasks. Zoom AI Companion is user friendly and covered by our existing licences. Zoom AI Companion is included as part of our current standard University package, so we would like to take advantage of this as much as possible.

We have also approved a trial of Copilot for Microsoft 365, an AI-powered productivity tool which has an LLM to understand, summarise, predict, and generate content. Copilot for Microsoft 365 pairs with the applications that many of us use every day, such as Word, Excel, PowerPoint, Outlook and Teams. Our small trial group includes the Governance Team, who will use Copilot for Microsoft 365 to help support our committee structure whilst monitoring its effectiveness. We hope that other teams will soon be able to benefit from the use of Copilot for Microsoft 365. In the meantime, we encourage colleagues to consider using Standard Copilot; which has more limited features but may still be helpful in your day-to-day work. You can access Standard Copilot via your web browser.

The Utilising Expertise in AI Steering Group is considering further projects, and we welcome any comments and suggestions that you may have on the guidance, or our approach to utilising AI for business processes. For further information see the Moodle course Quick Guide: Generative Artificial Intelligence. We will form a network of AI champions in the New Year and are looking for colleagues who can provide user case studies, so we can share best practice and learn from each other. Please do get in touch if you would like to be involved, email Fiona Harvey at fjharvey@essex.ac.uk.