DAVID HAYES - CDO, The Open University

Data plays a huge part in what we're doing at The Open University

David is leading The Open University’s data function at an important juncture when society is requiring more flexibility and AI could support them to respond to student needs with greater ease. He talks to us about his role in this organisation where data is how they first get to know their distance learners.

Sponsored by innova-tsn

A varied career

Like many who have risen to the rank of CDO, David didn’t start in a data role but rather grew into it. He explains, “I started work in the water industry as a finance graduate and that probably sent me into the world of analytics.”

After a brief stint in telecoms, David moved into banking. It was while he was at Abbey National that their acquisition by Santander saw David’s career fortuitously transition into data.

He’d worked in data-heavy accounting roles before moving across to be an analyst in an operational area. Santander then needed people with a deep understanding of the organisation to support data migration.

David picks up the story, “They created a central team called Data Quality and Structural Systems, which was very much a Santander model. And we found ourselves being the team that were responsible for making sure that, when data migrated from Abbey National systems to the Santander platforms, it was all OK.

“That needed really strong finance skills. It needed somebody who understood data and could manipulate it, and it needed somebody who understood the organisation.”

David had the skills and expertise that fitted those requirements, “That got me into data. We migrated millions and millions of records from many systems into one in lots of cases, and then we needed to control all of that data once it was in.

“The Santander model meant that you had to put much more emphasis in controlling your systems and your data. So, we ended up leading teams doing that. Then the banking crisis hit and banks evolved to need a CDO role. It started very much as a governance role – so, can we establish governance over our data, over the provenance, over the quality, over the way we report, etcetera?. I was leading the Data Quality and Structural Systems  team at that time and was in the fortunate position to be invited to take the role of CDO.

Sponsored by kainos

“I grew the CDO role at Santander, and whilst dealing with the regulatory reporting and governance challenges all banks faced, the world of data lakes and data science exploded. We began to build a more end-to-end data function in Santander and I left quite a large team there in 2019 to move to The Open University.”

Education is important to David. He’d considered teaching for a career and coaches children’s basketball and football in his spare time. So, this move fitted with his personal area of interest – and this particular role was appealing too.

“What the OU does and why the OU exists is amazing, the chance to be a part of that and to be able to help people change their lives through learning was too good an opportunity to miss.”

While there was already a data presence in the organisation, David tells us, “We've been able to grow quite a significant data function here at the OU, initially through consolidating some existing teams, making great use of our partner Infosys to support us to build capability and capacity, and then by gradually bringing in new skills and experience from outside the organisation as the demand for data and AI has increased and we have delivered more and more.

At the end of the day, when it comes to delivering with data and AI it’s all about the people and the teams, that was true in Santander and it is true today at the OU.”

Data to deliver on a mission

The Open University has a mission to make learning accessible to all. With almost 200,000 students aged from 18 to 80 and over from all 4 nations of the UK and beyond, they’re the largest university in the UK by student body.  As a distance learning provider one of the unique features of the OU is that all of their teaching materials are available online and years of pedagogical experience is captured in those materials David points out, “This means that data plays a huge part in what we're doing. The CDO would say this, but I tend to say that we know our students through the data they leave with us.

“So, we've been using that in the prospect to registration, to completion phase of the student’s journey for some time. We track data all the way through that lifecycle as best we can. We can see where they're engaging.

“There's loads of room for improvement in that journey, but we use data throughout typically to support the student, but more often to support the people who are helping the student – whether that's the teacher, or the lecturer, or the support staff that are in our contact centres.”

Sponsored by infosys

And there are a lot of these. Around 10,000 people work at the university, with about half of that number in teaching roles and half organising and helping the university.

Another important factor here is that a lot of students choose to study at The Open University (The OU) because it is, as its name suggests, more open to them – because they’re able to study part time around work or other commitments or because of disability, for example.

David continues, “In the region of 30% of our students are registered with disability, so they’re a really, really significant population of our university. We provide all kinds of learning adjustments and there's a very human touch to that. We make available a lot of humans who can speak to people, but we try to support those people as much as we can with information, systems, tools and processes to know where the student is in their learning journey.”

He points out that all this work is ongoing, “There's loads we can do to improve what's available to people, either the students or staff members. And we're on a constant mission to try and improve that.”

The possibilities to personalise learning experiences

As the conversation turns to AI, David tells us that at the moment, The OU is using AI behind the scenes to identify students who are at risk and then prompting teaching staff to intervene.

David explains, “We provide information into our teaching staff using a tool developed in the data teams that we call the Student Progress Tool, so they would know, for example, that someone is maybe at risk and needs a bit of support, and that might be because there's certain flags that we've seen in their engagement with the online learning, or it doesn't look like someone is going to submit their assessment. We drive that information into our teaching staff and we've got some great staff who make sure that they take the opportunity to contact students.

“So that's where we're currently focused. We believe that there's a great opportunity to go much further with that level of personalised support to make that information more visible to students themselves as well, so they can start to be prompted more directly or very carefully – because we want to be really careful not to prompt students in the wrong direction. There is quite a bit of testing that's needed for that.”

Looking a little further ahead, “We believe that it's going to be possible at some point in the future to personalise more the way students engage with material directly. So, two people could be engaging with the same course, and maybe one person could engage with it in a different way to another, but both learn the same thing.”

“We provide different materials today. So, you might get a video or some text, as a simple example, but you can work through the flow of the course today. We see a future whereby someone's learning preferences could influence the kinds of things that they look at and use, and then we could be using more personalised tools to give you the ability to test yourself as you go more than we do today.

“So, using AI in that sense to both flex the materials you're seeing and the tools you're using to test yourself, but not to replace the important role of the teacher in that space. And not to undermine the pedagogy and the content.”

David explains that the art of learning is contained within the university’s material because of all the years of “learning about learning, and learning about online learning” that they have, “Because we have this history of pedagogy and we've got units across the organisation who really know what it means to teach online at distance using materials as well as the teaching methods that we have.  The opportunity we see in front of us now is to try and answer the questions ‘How do we translate that using AI? How do we use AI to augment that and improve it whilst retaining the richness of the model we've got?’”

These approaches could play a part in keeping busy students actively involved in their learning journey. Drop-out rates for part-time distance learning are higher than in most other areas of learning, David tells us. Many of The OU’s students are working, and some are in periods of their lives where there are big life changes such as moving jobs or home, getting married or having children, or other factors which could disrupt their learning. This is why they want to make continuing to study as easy as possible, “There's lots we can do. We believe that encouraging engagement is probably the most important thing.

“Anything that we can do to encourage people to be more engaged in material, whether that's changing the material, changing how we deliver the material, or prompting people to do things at certain times, we're interested in trying those things out carefully.”

Trialling AI in course creation and delivery

While they’re yet to roll out AI in assisting teaching or content creation, they are currently trying both of these. The university’s OpenLearn offering supports this, “OpenLearn is a free learning resources website and we have a huge number of learners that use OpenLearn for free. So, it allows us also to test some things out.

One of the great parts of working at the OU is that we have access to world leading experts in educational technology and pedagogy.  The Knowledge Media Institute (KMI) and the Institute of Educational Technology (IET), both OU organisations, have together developed some prototypes for AI assisted learning and using their expertise we have  been trying out AI-assisted learning in Open Learn.   There's some early positive engagement results coming from that – students seem to value it on the whole and it seems to drive more engagement, which is what we would expect.”

David points out that there are some competitors outside the UK that are already offering augmented teaching using AI, but this is something The OU is approaching with care, “We're on a path to try to get that into our learning environments and there's every opinion about that, so we're going very, very carefully. Naturally there are some very bright people here, so there's some really well thought through arguments. You've got every opinion, from those who think that AI should not be considered at all and banned, to those who feel we should be embracing AI and everybody should be doing it now.

In 2025 the OU appointed a new Vice Chancellor, David Phoenix, and we are currently developing a revised strategy for the future.  The foreground to that strategy remains the mission and purpose of the OU, and it is really clear that AI will play a really important role in that future, both as part of what we do and a major external influence in why we need to do certain things.

We're active now in the AI space and because the pace of change is going so fast, who knows where we'll be in five years’ time, let alone wait another 10 years? So, we need to embrace it, we need to run with it and we need to do it smartly. And I think that's where the university is setting itself up now. So that, for me, is really good.

“We believe AI is something we need to do. But we are necessarily going carefully with that. We neither want to undermine students’ futures by not having it or undermine the quality of what we do, so we've got to get that mix right.

“In parallel to trailing AI assisted learning, we're trialling the use of AI so that we can fast track the creation of content, but we don't want to take the academic rigour out of the content we create. AI still has a reputation for hallucinations, making stuff up and oversimplifying things. So, we’re naturally being very careful with that.  We have a rich history of content we can use as trusted input to the AI we create and because of that we are confident that it will be possible to put the guardrails we need around any AI created content to ensure it works well for our students.

David thinks it’s really important to find the right balance between AI-driven outcomes and human connection in their learning environment which values empathy and academic guidance.

“We're learning through trialling the different AI ideas that we've got. Everyone's learning about AI in parallel. What everybody's saying, and we couldn't agree more with this, is the importance of keeping a human in all the way through the process is so, so vital. AI can shortcut a whole host of amazing things that I think can make our working lives a lot easier, but it doesn't replace humans. It doesn't do some things that humans do, and I think that's really important to hold on to.”

Putting AI on the curriculum

As well as the potential to support content creation and student learning, The OU wants to embed AI in their curriculum, “Teaching about AI in the context of the subject you're learning, whether it's History or Classics or Psychology or French or whatever it might be.

“Our students are going to be using AI in their future and we have an obligation to make sure they understand that and an obligation not to ignore it. But we also have an obligation to teach about it responsibly and do things ethically.”

We wondered how AI could or should be used in student assessments? “We take 3 approaches. One, you can use AI in this assessment. Two, you could use AI in assessment. Three, you mustn't use AI in this assessment.

“There's been a move for a very long time now to make assessment much more authentic and that might mean, for example, producing a video or doing some kind of group work – doing something that maybe is a bit more real-world like. The advent of generative AI is fast tracking that a little bit, partly driven by the fear of cheating, but more importantly driven by the reality that it is part of everyday life, so there's even more reason to make assessment even more authentic.”

While David transparently admits that he’d like to do more with upskilling people across the organisation in terms of AI capability, work is underway in this area. He gives kudos to the  university’s central academic teams, who are playing a big part in this, “They're doing a huge amount of work to improve AI literacy amongst the academic areas of the university. They've run events to try and provide first of all understanding, then learning, then guidance on how you can embed AI in learning, teaching and assessment.

As an example, “They ran a two-week ‘Faiesta’ session in September this year, where many of us – were talking about ‘What are we doing about AI in the organisation? How do we put our arms around it? What can you do in your role? How can you take it forward? Why do we think this matters? Ask us questions’ and so on. These were really good sessions.”

He points out that working with the wider teams across the organisation is key for the digital teams at The OU, “We don't deliver things in isolation. Typically, if you're going to deliver an application with AI in it, it takes all kinds of skills to do that, not just data people.

“There are such inspirational people around the university, so the ability for us to collaborate and get stuff done, I think is the crucial part. Without that, nothing happens.”

Thinking beyond studying

Looking beyond the learner experience, The OU have been using more traditional AI in their marketing and prospect funnel for some time. This enables them to identify cohorts of prospects or students they could reach out to and prompt during the signing up or re-registering process.

They also have internal use cases associated with the use of generative AI to make policies, procedures and processes quicker and easier to engage with.

David continues, “We're experimenting with Microsoft Copilot like many organisations are to see what we can do in those spaces. And we have relationships with other providers who are now embedding AI more natively in the things they do. So, for example, the HR team are looking at embedding AI in our HR systems to help just make those processes a bit slicker than they are today.

“And we're looking at a platform approach to Agentic AI as well. We're just at the start of the pilot approach of testing something out in that space that we think will help us to create, manage and control Agentic AI. We think there's a host of areas where we can improve our existing process using agents, but we've got a huge learning journey to go on in that space, like many organisations. We're trying out a platform that's going to help us put some control and governance around that.”

Putting AI principles in place

With so much work ongoing, we wondered about the frameworks or governance models The OU is guided by in their approach to responsible AI. As they’re also registered in Ireland, the EUAI Act is relevant to them and, David points out, forms a very good basis for what they should do as an organisation.

He tells us, “We've established a responsible AI policy which is signed off across the organisation. We're establishing guidance for staff members and students that's been published. And we're in the middle now of operationalising governance models that will essentially allow us to do a risk assessment of that.

Working with Kainos, our digital partner, we have trialled an approach they had created to run an early and ongoing risk assessment for AI solutions that puts the emphasis on early engagement and practical action to address risks.

“So, if someone's got an idea they want to do something with AI – they want to build an application or something –  we'll do an early risk assessment using our responsible AI policy and the guidelines.

“The idea of the early risk assessment is so we say yes or no to this case upfront? And if we do, how do we help the owner of that case to fully understand the risk of the things that they are developing so they can mitigate them? Then what we will do is establish a process whereby people can illustrate how they've mitigated the risks associated with the AI.

“There’s always going to be trade-offs and there's going to be challenges. There's a lot of grey, so there are some things that are just a no. There are some things that are going to be an easy yes. But most things are going to be in the middle of those two things. So, the management of risk all the way through is really important.”

While they’ve been working through their approach, as a learning piece The OU has gone live with a generative AI solution to support staff with responding to student emails – and it’s been teaching them a lot.

We’re really proud of the fact that we have managed to get a Generative AI tool into production, especially when you hear the often-quoted statistic from MIT that only 5% of GenAI solutions make it to production.  It was a real team effort, with great support and inspiration from our external partner Innova, working with teams across the OU to be user driven from the start, and always with the human in the loop.  We know it is not perfect and we are embracing that because we also know that without going into production we would have missed a huge amount of learning about both what it takes to get a solution built and deployed and what it means to run a safe solution in production.

David shares, “It took us longer to create and go live with a solution for that because the governance wasn't in place, because the policy wasn't established, because we hadn't gone through that process. And therefore, we did lots of risk assessments, lots of pausing and reflecting, and then we added in lots and lots of checkpoints just to make sure it was OK.

“We're very hopeful that by having a clearer responsible policy and a clearer framework for going live with it, actually it’s going to speed things up in two ways: we'll just not do some things, which is good, and then we'll be able to move forward faster in a clear path with the stuff we’re running.”

Looking ahead at what’s possible

While there’s already a lot happening, there’s more to come for The Open University – with the priorities for embedding data and AI capabilities in the coming years focused on what they can do to improve students chances of succeeding.

The OU wants to embed AI in what they teach, how they teach and how they operate, David tells us, “We'll try to drive up efficiency, improve how we teach and improve what we teach using AI. Those will all sit as priorities for us.

“If you jumped to the three outcomes we're looking for, we really want to improve student outcomes, we want to give students a higher chance of succeeding and we want to try and narrow the gaps that still exist between different areas of society. There's a lot of thought that AI might widen those gaps because of digital divides, however I think we take the view that AI could help us to narrow those gaps, because it's easier to reach the student directly if you can give them easy-to-use tools.

“Obviously we also want to drive up the efficiency of the university – spend more of our hard earned money on teaching people rather than administering things. And we also believe that AI can continue to support us to maintain really high standards of education, so high quality research could be supported using AI, with AI, and about AI. We think that there's a lot we can do in that space.”

The future of AI in HE?

Looking more broadly at how David sees data and AI may shape the future of Higher Education – particularly for part-time, remote learners – he thinks that significant change is ahead, “It's definitely going to reshape the sector. It's definitely reshaping learning. Students are already using AI in their learning and that's, I think, broadly a good thing. What we need to be good at is teaching them how to do that safely and critically, so they are equipped for a confident future with AI. I think it can help us reach more students and help them to be more successful, and I think it can help us be more efficient in the way we do things.”

When it comes to personalisation with AI, he sees that as being dependent on an organisation having the ability to get good data from the student that they’re happy to provide because they trust the organisation with it, “We've got to be really good at collecting and being trusted with data. And then we've got to be really good at how we use students' data transparently and ethically back with them as well. I think that's going to be important for all universities across the sector.”

He thinks a wider trend of wanting flexibility is another relevant factor, “Society is consistently looking for more ease of learning – learning when and where I want it, hop on, hop off, make it relevant to me now.”

Add to this the importance of having really up-to-date content for students and that creates a challenge, “AI could really help in that space. I think it's quite possibly transformative. But it's transformative because society is transforming to be much more flexible and we need to be as ahead of that as we possibly can to provide what students really need.”

We wondered how hard it is for organisations of this scale and in this sector to plan for the future when technology is changing fast? It doesn’t seem to faze David, “I find that really exciting. The idea that we would be in a static world: I find that would be a little bit dull. I like the change. I like the provocation that AI brings. I like the opportunity.

“The possibilities of what we can do with data and now with AI are huge. The challenge is getting that right and making it work. And, you know, we've had a pretty good few years trying to do things with data and AI. I'm pretty confident we can do a lot more as well.