Over the past few months, I have noticed how quickly AI enters conversations about process improvement.<\/p>\n
A process takes longer than it should. A workflow has grown around spreadsheets and manual checks. Information exists, but is difficult to find. Students are not always sure which route to take, which form to complete, or which guidance applies.<\/p>\n
Often, colleagues are doing their best within processes that have become more complicated over time.<\/p>\n
And quite quickly, the question becomes:<\/p>\n
\nCould AI help?<\/p>\n<\/blockquote>\n
I think this is a good thing. It shows that people are engaging with AI, but also that they are looking again at processes, services, and sources of friction that may have been accepted as normal for too long.<\/p>\n
That feels important. If AI is encouraging us to ask better questions about how our work is organised, then that is already useful.<\/p>\n
Sometimes the answer is yes.<\/p>\n
But I think there is a more important question we need to ask alongside it:<\/p>\n
\nAre we solving the problem, or just adding AI?<\/p>\n<\/blockquote>\n
That distinction matters. AI can be genuinely useful, but it is not a magic layer we can place over unclear processes, disconnected systems, or poorly understood workflows and expect everything to improve.<\/p>\n
If we are not careful, we may simply end up speeding up a flawed process rather than improving it.<\/p>\n
In some cases, AI may be exactly the right tool. In others, the better answer might be process redesign, rules-based automation, AI-supported automation, better data, clearer ownership, or simply making information easier to access.<\/p>\n
The challenge is knowing the difference.<\/p>\n
Starting with the problem<\/h2>\n
The risk is subtle. Once AI is on the table, it can become the frame through which we view the issue. We start looking for a tool to deploy, rather than asking what is actually causing the friction.<\/p>\n
A better starting point might be:<\/p>\n
\nWhere are the biggest sources of friction for students and colleagues?<\/p>\n<\/blockquote>\n
That question feels more useful to me.<\/p>\n
It starts with people rather than technology. It asks where time is being lost, where confusion is being created, where work is being duplicated, or where students and colleagues are experiencing unnecessary barriers.<\/p>\n
Only once we understand that can we sensibly ask whether AI has a role to play.<\/p>\n
Not every problem is an AI problem<\/h2>\n
Many of the frustrations that exist in universities are not caused by a lack of AI.<\/p>\n
They are often caused by things that will be familiar to many people working across the sector: processes that have grown more complicated over time, systems that do not talk to each other, guidance that is difficult to find or interpret, manual workarounds, duplicated data entry, unclear ownership, or workflows that made sense once but no longer match how people actually work.<\/p>\n
These are real problems. They take time. They create frustration. They can make services feel harder to navigate than they need to be.<\/p>\n
But they are not always AI problems.<\/p>\n
Sometimes the process needs to be simplified. Sometimes the right answer is better integration between systems. Sometimes it is clearer guidance, a workflow review, or simple rules-based automation.<\/p>\n
AI can be part of automation, but not all automation needs AI. If a task is predictable, rules-based, and easy to define, a simpler automated workflow may be safer, cheaper, easier to test, and easier to explain.<\/p>\n
AI becomes more useful when the task involves ambiguity, language, varied inputs, classification, summarisation, or interpretation.<\/p>\n
If we skip that analysis and jump straight to AI, we may end up treating the symptoms rather than the cause.<\/p>\n
That does not mean AI has no value. It means we need to understand what kind of problem we are dealing with before deciding what kind of solution is appropriate.<\/p>\n
Where AI can genuinely help<\/h2>\n
There are many areas where AI can provide meaningful support, especially where work involves large amounts of text, repeated queries, complex information, or time-consuming administrative steps that are difficult to manage through simple rules alone.<\/p>\n
AI may be able to help by summarising long documents, drafting routine communications, identifying missing information, helping colleagues navigate policy guidance, analysing themes in student feedback, or supporting students to find the right information more quickly.<\/p>\n
Used well, these are not trivial gains. They can make information easier to find, reduce unnecessary steps in a process, and cut down the amount of avoidable administrative work that has built up over time.<\/p>\n
More importantly, they can create more space for colleagues to focus on the parts of the work that require judgement, care and expertise. Often, those are also the parts of the work that feel most meaningful, creative and enjoyable.<\/p>\n
But this should still be a supporting role.<\/p>\n
I keep coming back to this point in conversations and in these posts because it is easy to agree with in principle, but harder to protect in practice. At Kent, one of our guiding principles around AI is that it should support human judgement, not replace it.<\/p>\n
That matters because, in higher education, many decisions require context, empathy, experience, and professional understanding. This is especially true where decisions affect students\u2019 progression, wellbeing, support, or access to opportunities.<\/p>\n
A useful way to diagnose the problem<\/h2>\n
<\/p>\n
Before deciding whether AI is the answer, it may help to ask what kind of problem we are really looking at.<\/p>\n
\n\n\n\n
\n \nIf the issue is\u2026<\/th>\n The answer may be\u2026<\/th>\n<\/tr>\n<\/thead>\n \n Repetitive manual steps<\/td>\n Workflow automation, which may or may not involve AI<\/td>\n<\/tr>\n \n Confusing handoffs between teams<\/td>\n Process redesign, clearer ownership, or workflow automation<\/td>\n<\/tr>\n \n Disconnected information<\/td>\n Better systems integration or knowledge management<\/td>\n<\/tr>\n \n Guidance that is hard to find<\/td>\n Improved search, clearer content, or AI-supported navigation<\/td>\n<\/tr>\n \n Large volumes of text<\/td>\n AI summarisation, classification, or drafting support<\/td>\n<\/tr>\n \n Complex scheduling or forecasting<\/td>\n Analytics, optimisation, or AI-assisted modelling<\/td>\n<\/tr>\n \n Sensitive decisions affecting students<\/td>\n Human judgement, possibly supported by AI, but not replaced by it<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<\/div>\n This is not a perfect framework, but it helps avoid a common trap: assuming that because AI could be used somewhere, it should be.<\/p>\n
The more useful question is whether AI is addressing the real source of the problem.<\/p>\n
If a process requires colleagues to spend hours manually copying information between systems, AI may not be the first answer. The real issue may be systems integration or a better workflow.<\/p>\n
If students are repeatedly asking the same question, an AI assistant might help. But it may also be worth asking why the existing guidance is unclear, hard to find, or written in a way that does not match students\u2019 needs.<\/p>\n
This matters because AI does not magically fix unclear information. If the source guidance is contradictory, incomplete, outdated, or written in a way that students already find difficult to understand, then an AI tool may simply reproduce that confusion in a more conversational format. In some cases, it could even make the issue harder to spot, because the answer may sound confident even when the underlying information is weak.<\/p>\n
Before using AI to answer questions, we may first need to improve the quality of the information it is drawing from.<\/p>\n
If a workflow is slow because every case is genuinely complex and requires careful judgement, AI should not be used to force speed at the expense of fairness or care.<\/p>\n
This is where universities need to be thoughtful. The goal should not be to automate complexity we have not properly understood.<\/p>\n
A well-designed process, supported by simple automation where appropriate and AI where it genuinely adds value, may do more for students and colleagues than an AI tool placed on top of a confusing workflow.<\/p>\n
Keeping students and colleagues at the centre<\/h2>\n
For me, the strongest argument for AI in universities is not productivity for its own sake.<\/p>\n
It is the possibility of reducing unnecessary friction.<\/p>\n
If AI can help students get clearer answers, access support more easily, or spend less time trying to understand university processes, that matters.<\/p>\n
If AI can help colleagues spend less time navigating avoidable administrative complexity, and more time on thoughtful, human, and meaningful work, that matters too.<\/p>\n
But we need to be careful about the story we tell. If AI is framed only as a way to make institutions more efficient, it may understandably make people anxious. Colleagues may worry about replacement. Students may worry about being pushed towards automated support when they need a person. People may question whether speed is being prioritised over quality, fairness, or care.<\/p>\n
That is why the purpose matters.<\/p>\n
AI should not be used simply to move people through flawed systems more quickly. It should be used, where appropriate, to help create better systems, clearer information, and more responsive support.<\/p>\n
That requires governance, transparency, good judgement, and humility. We need to be honest about what AI is good at, what it is not good at, and where human expertise remains essential.<\/p>\n
So, could AI help?<\/h2>\n
Yes, in many cases it probably could.<\/p>\n
But not everywhere, and not automatically.<\/p>\n
This is not an argument for doing less with AI. It is an argument for using AI more purposefully.<\/p>\n
The most useful AI projects are unlikely to be the ones that begin with the most impressive technology. They are more likely to be the ones that begin with a clear understanding of the problem, a strong sense of purpose, and a commitment to improving the experience of students and colleagues.<\/p>\n","protected":false},"excerpt":{"rendered":"