Why AI Adoption Fails Without Governance, Data, and Training

AI agents are moving from experiment to enterprise reality, but the real challenge is not adoption - it's control. Antonio Vieira Santos and Luis Lancos sit down with Derek Hickey to unpack what it actually takes for organizations to move from assisted AI to agentic AI without losing governance, trust, or direction. Derek brings more than 30 years of technology leadership across financial services and now the charity sector, giving him a rare view into how two highly regulated industries are approaching AI very differently. From data readiness and budget pressure to regulatory risk and release cycles that are shrinking from weeks to days, he shows why some organizations are accelerating fast while others are still getting their foundations in place.

You’ll discover:

• Why charities are adopting AI more slowly, even though the technology is now cheaper and more accessible than ever 
• How financial services are choosing platforms, defining strategy, and building use cases before scaling 
• The rise of AI Review and Design Authorities, and why guardrails matter more than speed 
• How roles are changing as engineers become reviewers and principal developers shift into approval and oversight 
• Why “AI literacy” and internal champions are becoming essential for safe, effective adoption 

Antonio pushes on the practical realities of governance, board expectations, and the gap between talking about AI and actually making it useful. 

Luis frames the big strategic question: are organizations building a new layer on top of legacy systems, or fundamentally changing how the organization thinks?

Derek’s answer is clear: if you want AI to work, you need clean data, trained people, transparent board conversations, and a strategy that leads technology instead of letting fear drive it. He makes the case that AI is no longer optional, but it only creates value when it is introduced with the right controls, use cases, and accountability.

This conversation is essential listening for leaders, technology teams, board members, and anyone responsible for turning AI ambition into something safe, measurable, and real. If you’re about to approve your first AI agent, this episode gives you the questions you need to ask before you say yes.

Transcript

Copy Transcript Copied

License to Act
Episode 1 transcript

Hosts: Antonio Santos and Luis Lancos
Guest: Derek Hickey

The speaker's name appears at the start of each turn.


Antonio Santos: Hello everyone. My name is Antonio Santos, and I'm one of the co-hosts of a new podcast called License to Act, a podcast about how organisations are using artificial intelligence agents today. With me is my co-host, Luis Lancos. Luis, welcome. Can you explain a little further what this podcast is about?

Luis Lancos: Absolutely. Thank you so much, Antonio. I'm really excited about this first episode. We've been talking about it for a while, so seeing it become a reality is awesome.

I think this podcast is the result of exciting times. As they used to say, what a time to be alive, given what's happening across the market with artificial intelligence technologies. We are moving away from assisted artificial intelligence. We are now in acting artificial intelligence. With what they call agentic artificial intelligence, we are moving towards what we call the thinking organisation, or the agentic organisation.

That means the organisation starts to use agentic artificial intelligence to map processes, to act, and to understand context. That raises a lot of questions, from the technology side, the regulatory side and the people side. What does it mean for the organisation? This podcast explores that, and talks with organisations that these technologies affect, especially in the regulatory space. What does it mean for them, and what is the best way to approach it?

Antonio Santos: To kick off our first episode, we have Derek. We're happy to have Derek today. Why did you decide to invite Derek to join us?

Luis Lancos: Derek is a fantastic guest and a fantastic person. I know him personally. We've crossed paths within financial services. Derek has more than 30 years of leadership in technology. He has been driving emerging technologies and innovation in financial services, and he has a proven track record. I met him when he was leading technology at a bank. Currently he is doing important work in the charity sector, and I believe what he's doing will be really impactful. So I thought it was an awesome start to have this conversation and look at what's happening there.

Antonio Santos: Welcome, Derek. Nice to have you. I heard that you are also based here in Ireland, correct?

Derek Hickey: Good morning. Indeed, yes. I'm based in a place called Dunboyne, just at the border between Meath and Dublin, although I'm Dublin born and bred. On a daily basis I usually end up somewhere in Dublin, either at our head office in Park West or at a conference somewhere around the Dublin area. Great to see you both.

Luis Lancos: Great to see you too, Derek, and thank you for joining us.

Antonio Santos: I'd like to know where artificial intelligence agents are today in the charity sector, and how that compares with financial services, because I know these two areas are close to your heart and relate to the work you are doing.

Derek Hickey: From a charity perspective, it's probably a slower burn in take-up and adoption. There are a couple of reasons. The first is cost. Artificial intelligence is more accessible today than it has ever been. You can buy a licence for 20 euros a month, which is reasonably cheap. But you need to have your data in the right format. You need it to be accessible, and you need to be able to use it well.

Much of the charity sector is focused on cybersecurity, resilience, uptime and operations, without a doubt. But they weren't always in a position to be ready for this artificial intelligence onslaught and take-up. So it's slower in the charity sector. Because artificial intelligence is so accessible, and a lot cheaper than newer technologies were in the past, adoption will get easier. I think you'll see greater adoption, even hyper adoption, in the charity sector over the next while.

Banking and financial services have, over the last 20 years, embraced new technologies for a number of reasons. When the internet kicked off, financial services used it as another sales avenue. They didn't take it as a place to innovate. Over time they started to innovate a lot more, to look at technologies and where things were going, and to push. Now that artificial intelligence has come in, I think financial services have really adopted it. There are a couple of risks and areas of concern too. Financial services is a highly regulated industry, as is the charity sector, let me say. So there's caution, but there's also the want to adopt.

We're seeing more organisations choose the platform they want to build their artificial intelligence on. They're starting to understand the strategy they want, and they're pulling in a lot of use cases. Without those three pieces, it's very difficult for any organisation or industry to adopt a new technology or a new way of working.

Financial services has put itself in a really good position. It has a lot of data, and in some ways, because of regulation, that data is in a much better state than in other industries. That gives them a great starting point. They're used to adopting new technologies since the rise of the internet. Many already have an in-house innovation service, so they're keen to try new things. They have quick frameworks for bringing something in, trying it out, and either failing fast or adopting and scaling quickly. And they have the budgets, which is a big help. They can spend extra money on the right controls and guardrails while also investing to accelerate adoption, both in their own organisation and across the industry. So these are two very different industries, but both want to adopt artificial intelligence at a quick pace where they can.

Luis Lancos: Thank you for that, Derek. A couple of things come to mind. As you said very well, both your experience in financial services and in the charity sector are in highly regulated industries where artificial intelligence will have an impact at various levels.

When we look at the charity you are running, what are the key goals? And how do you see assistive artificial intelligence, so artificial intelligence that assists people, and autonomous artificial intelligence, artificial intelligence that can act and drive entire processes and then escalate to humans? That question relates to the discussions I'm seeing about the agentic organisation. McKinsey itself started to float this idea.

It raises a relevant question. When we move towards the agentic organisation, or as I call it the thinking organisation, are we talking about a layer on top of the old infrastructure, or a new way of thinking about organisations? I'm trying to get a sense of where we are, what this means, how we approach it, and how you see it transforming the charity sector.

Derek Hickey: That's a really good question. Let's start with the obvious part of the answer. Many charity organisations are software as a service model organisations, so they outsource quite a bit. They do that because of cost, team size, skill sets and knowledge, among other reasons. Over time I think you'll see less need to outsource everything. That will matter for a number of reasons. You get more control of your data, more control of your systems, more control over how quickly the systems update, and more control over how you can touch that data and do things with it. So charities will use artificial intelligence to drive down some costs, to find areas where they can move quicker, and to streamline processes. That's the basic answer.

Then, particularly in the sector I'm in, which includes health and also education, they'll look for better insights into their data. If I look at the health side for a moment, they want to know, with all this data, what's happening in the sector, what's working and what's not. Then they'll add and adapt more modelling on top of that. What is all this telling us? What happens if, in the next 3 years, we see more symptoms in one area or more specialisation in another? They'll use the data to understand what skill sets and what knowledge they will need, and to model where they've seen growth today and where they might see it tomorrow.

They'll also use artificial intelligence in the day-to-day. Not necessarily within Rehab, but we're already seeing this in hospitals and in the Health Service Executive, where they've started to use artificial intelligence to look at X-rays and magnetic resonance imaging scans and feed information up. That doesn't mean they're taking the person out of the loop, and I don't think they ever will. But they're giving those individuals a better starting point. So I think you'll see artificial intelligence adopted slowly, and then you'll see an acceleration of the different use cases.

That will be shaped somewhat by regulation, by risk appetite, by the budget available, by the outcomes they see, and by the regulations coming down the road, whether that's the European Union Artificial Intelligence Act or others. So adoption and use in the charity sector will be slightly different from financial services. How organisations set themselves up will be quite different too.

I'm talking to a few people in the charity sector, in financial services and in other industries about how they're setting themselves up for good use of artificial intelligence. In some cases their people have started to do different work. You might have had engineers whose job was coding, maintaining and building the systems. Now they've taken a step back. They've created agents, and those agents do a lot of the bulk work. The principal engineers, who in the past might have been making architectural decisions or finding solutions to bigger problems, have stepped right back. They look at the code the agents produce, and they sign it off. Their role has changed. So it's not only a change in how we do things. It's a change in how we set up and teach our people.

One group in particular has about 300 agents doing all the coding. They have about 30 principal developers who used to develop day to day. They don't do that anymore. They review the code, they are responsible for signing it off, and then it is delivered into production. Where releases used to run on a weekly, two-week or even monthly cycle, we're now seeing a daily cycle. For certain industries, that's amazing.

So artificial intelligence is being used in all different ways as we move forward. Over the next 6 to 12 months, I think it will be very interesting to see whether organisations settle on a particular strategy, because up until now the strategy has changed every 6 months. It used to be one agent doing one thing, then one agent doing multiple things, then multiple agents doing multiple things. It keeps changing. Until organisations settle on what works best for them, we'll see that continue.

The other piece a lot of organisations now look at is cost. How are we spending the money on tokens? What are these costs? How can we control and manage them better? That will be a huge indicator of how artificial intelligence is brought forward.

Antonio Santos: Derek, a recent study found that Ireland is among the countries in Europe adopting artificial intelligence fastest. But adopting is different from making something useful out of it. The charity sector and the banking sector have common points, but they also have many differences in scale and in internal governance. In charities, you can give everyone licences to use software, but how do you make sure you have the governance in place and are actually building something useful? It's very easy to spend tokens. It's more difficult to make something useful out of them.

Derek Hickey: That's a really important piece, Antonio. I've been speaking to a few chief information officers and chief technology officers, and a few of them have done a really good job on this. Many technology leaders have noticed that it's far easier today to create legacy software and shadow information technology, as we used to call it, than it was two, three or four years ago. Back then it might take several years to build up shadow information technology or legacy technology. Now, with artificial intelligence, you can do it in 6 months. You can probably do it in 6 weeks.

So you have to be very careful, and having the right frameworks and guardrails is really important. What I've seen work really well, and my own organisation, Rehab, included, is what we call the Artificial Intelligence Review and Design Authority. It's a board, and everything to do with artificial intelligence comes through it. If we want to purchase new licences for a type of artificial intelligence, if someone has a use case, if someone wants to build an agent, or if we want to bring in a new tool that has artificial intelligence inside it, it all comes through that board before anything goes to production. It's like a change advisory board for artificial intelligence. It's reviewed, then it's declined or approved, then it's rolled out, documented and monitored.

If we do a proof of concept for 3 months, we sign off those 3 months, and if anything changes in that proof of concept it has to come back to the board. At the end of the 3 months it comes back to the board, and they explain what they did and what they found. If the results are what they hoped for or better, we work with them to scale it out.

I should say that this is not about slowing down artificial intelligence, or about not wanting to use it. The board is set up to help with adoption, but to adopt it correctly, so that in 6 or 12 months we don't end up with a lot of artificial intelligence out there, a lot of high cost, and a lot of stuff that is redundant. We want high adoption, managed costs, a consistent artificial intelligence architecture, and people trained to get the most out of it. I'm seeing more and more of these frameworks come up, and it's absolutely key.

Then there is the regulation that has come down and is still coming down. You absolutely need this in place. You need somebody dedicated, your artificial intelligence person within the organisation. If you're building artificial intelligence, not just consuming it, you need it documented, so you know what you've done. If you change how you've built your agents, or the artificial intelligence itself, you must document that too. First, you have to comply with the regulation. But you can also be audited, and you could put your organisation at risk in certain areas. So you have to put the right guardrails and frameworks around it.

I truly believe that if you do this right and you do it early, it accelerates adoption, and it accelerates it in the right way. Areas like technology, risk and finance will build much more confidence in your organisation, and you'll get more support as you go on. I've seen this happen many times with emerging technologies, not just artificial intelligence. There's a fear factor, and understandably so. People wonder what risk it carries, what cost it carries, and how to make sure we don't breach any risk or become non-compliant. So it's really important to bring these people along, so that they understand what's happening and can clearly see the guardrails in place. If you demonstrate that, it really helps you grow and roll out artificial intelligence. And it needs to happen. Over the last 10 or 15 years, it's nearly the one technology that is non-negotiable at this stage.

We were all around when blockchain came out. It looked like a very exciting technology, and it's absolutely useful for a lot of things. People thought it would be a lot more prevalent in organisations, and over time they found the right uses for it. With artificial intelligence, it's non-negotiable. You've got to use it. It helps people in their day-to-day work. It helps organisations move faster, understand their data better, and make better and quicker decisions. It really helps organisations to streamline. Once we can do all of that confidently, within the right guardrails, it becomes one of the most needed and wanted technologies in the last number of years, if not in the last 20 years.

Luis Lancos: I think this is an incredibly important topic. It's a fantastic discussion, and to tell the truth, I could spend the next 2 hours going through some of these topics. You touched on really relevant points, and governance is definitely one of them.

On that, I'd like to hear your thoughts on what I call the distribution of intelligence across the organisation. Let me frame it. Artificial intelligence and these agents are incredibly powerful. The last thing we want is, imagine, to give Claude or ChatGPT or any other agent unlimited access to the resources of the organisation. You would be lucky if you still had your data or your customer relationship management system a week later. So the relevant question is no longer how powerful this technology is. The relevant question is how we distribute that intelligence across the organisation with purpose, controls and governance. I would love to hear how you're addressing that.

Derek Hickey: It comes back to one of my earlier points about the data. You have to understand your data: where it is, what's in it, and who has access to it. That's a core principle. If you don't understand that, your artificial intelligence, or even the people in your organisation, may have access to something they shouldn't, or use something in a way it shouldn't be used. So your data has to be set up correctly.

You also have to make sure your teams are trained. They have to understand the basics, the do's and don'ts of using artificial intelligence. You don't even have to have artificial intelligence in your organisation today. Someone can take a document or a spreadsheet, go to the ChatGPT web address, write a prompt and put it in. If they don't understand the basic do's and don'ts, that may be a data breach straight away. So educating people is number one.

Then there is training people on how to use it. You can buy as many Copilot licences or Claude licences as you can afford, but if people don't know how to get the best out of them, you won't help your adoption. So you have to look at training your organisation. That might be a couple of key people. You might have artificial intelligence champions. One thing I've found really useful is identifying key people across the organisation, whether in technology, finance, operations or elsewhere, and training them on how to build agents and write key prompts, and then how to roll that back out into their business group. That makes it really simple.

The other piece is keeping your artificial intelligence architecture consistent. There is so much choice out there today. How do you keep it consistent? Some of that adoption rolls off or reduces over time, so you have to look at that. And you have to have real use cases. What are the use cases you want to pursue as an organisation to drive adoption, to drive return and to show the benefits of artificial intelligence? I think everybody should be using artificial intelligence today at a minimum, whether that's to produce a document or write an email, something simple that helps in your day-to-day. Most people seem to do that today.

But getting the real benefits means introducing all those elements, and they all have to move at a similar pace. If you introduce new artificial intelligence features or licences and people don't know how to use them, you'll never get adoption. If you've trained people but have no real use cases, you won't get adoption either. You have to look at all the factors that support the use of artificial intelligence.

That's a challenge, because the business wants to move really quickly, and charities are included. We want to move quickly because we want to offer the best services we can, and to reduce our costs where possible, while making sure the people who use our services get the benefits without putting them or anybody else at risk. Doing all the right things and setting yourself up in the best possible way will give you the benefits much quicker.

Antonio Santos: Training is an important piece. But a large number of people have moved into this space of training, so organisations are probably flooded with offers: we can train your organisation on artificial intelligence, please don't miss out. What I want to ask is how important it is that the organisation can also do this kind of training internally, in a collaborative way. You already have your own experts. They are going to start using artificial intelligence, and some people already know it really well. How can the organisation, from the inside, work collaboratively so that people can help each other and upskill in the context of what the organisation needs?

Derek Hickey: I think it's about being selective in some ways, and rolling out widely in others. Awareness training is really important, and it's for everybody. What is artificial intelligence, how do I use it, what are the key do's and don'ts, what is the regulation out there, how does it affect me. That gets everybody to a certain level and gets the general use of artificial intelligence out there, so people see the benefit. That's number one.

The second piece is the artificial intelligence champions. You find people across the organisation who have a little more technical aptitude. Artificial intelligence is not only for people who have been developers for the last 20 years. Most people can learn how to code with artificial intelligence or write good prompts. But you want a number of key people across the organisation, for a couple of reasons. One, they can build agents that support the organisation. Two, they understand the organisation and the business, and that's really important. As with any process or new technology, to get the right output you need experts feeding in to build the right thing.

Then you have a funnel where people bring ideas and use cases that get reviewed, approved and agreed very quickly, so they can be rolled out. Training is one aspect, and then it's about maximising the output from that training. You capture use cases, review them, approve them, develop them, roll them out, and measure afterwards. As with any good innovation, you baseline before you do anything, and you baseline after, and you see what benefits you got. If the benefits you set out to achieve are there, brilliant, and you might scale up or roll it out further. If they weren't achieved, you might decide that it doesn't work, kill it and move on to the next one. You have to have all those gateways and doors available for people, to get the most out of it.

So start by getting everybody to the same knowledge base, and then start to grow some of the innovators. Some organisations have done this really quickly and exceptionally well. Others are slower on the adoption of artificial intelligence, so it will take them a little longer.

Luis Lancos: Derek, on that topic, because people is probably one of the most important subjects to drive this. We all agree that we can't put the lid back on. These technologies are out. The can of worms is open, so it can't go back. The question is what it means from now on and how we address it.

There is a lot of doom around artificial intelligence: that it will kill all the jobs, that it will make people irrelevant, and even the latest conversation about deceleration, and all the rest. The charity sector deals with people, so it's particularly sensitive in this area. When we look at this doom, the perception that is mostly driven by fear, what does it mean? How should technology leaders, and not just technology leaders but chief executives as well, look at artificial intelligence? What does it mean for the organisation, for the people it serves and for the people inside it?

We already understood, and you touched on this with the engineers, that there is a change of roles. I think artificial intelligence will change the level of abstraction within the organisation, so we are going to reshape roles. I would really like to look deeper into the people aspect, in light of this doom conversation.

Derek Hickey: It's a tough one to answer, but I'll give you my thoughts. It's something I've thought a lot about and had a lot of conversations on. First, artificial intelligence is a massive opportunity, because in some ways it nearly levels the playing field. It's so accessible. If you have a smartphone or a laptop, you pretty much have access to artificial intelligence in some shape or form. So I think it's a great opportunity from that side.

From a people perspective, roles will absolutely change. There's no doubt about it. Whether that means a net loss or a net gain in jobs, I don't know. Nobody knows the real answer to that. I've got my thoughts on it, but no one knows. What I'm pretty convinced about is that roles will change.

From a technology perspective, when technology started, our job was to look after the hardware and keep the lights on. We were usually on a different floor, and that's what we did. Over the years, technology and the business grew closer together. Over the last 5 years or so, and probably more, technology moved much more into the risk conversation and into the security conversation, and into the cost conversation. Over the last 18 months or so it has moved more and more into the operational conversation. That's relatively new for organisations, because operations was a place where the experts ran things and got on with it.

So I do think roles are going to change, and I've seen them change already. As I said earlier, some of your engineers are no longer developing. They're reviewing and approving. You're also seeing a crossover between technology, product and operations. Those groups are coming much closer together. You're no longer just a technologist who knows how things operate under the hood. Technology has always been good at this, but now you have to understand the product, the operations, the cost, and more and more everything around the business: the people, the customers, the day-to-day running. You have to care about it all. You have to understand that what you do here happens over there.

Artificial intelligence has really changed that profile within an organisation, and I think it has to continue to do so. The hard lines, the silos where you had operations, technology or finance, are evaporating, and it's becoming one. You have to get under the hood of everything now. It's not enough to say, I created this agent, now you go off and use it. That's not how it works. It's: how does this affect the organisation? If we do this, what's the next step? How can we automate this better? How can we use artificial intelligence to drive this better? If you don't understand the nuts and bolts of the organisation all the way across, you will struggle.

So from a net perspective, it's hard to know what it will do to roles and jobs. But certainly from a skills and expertise and day-to-day perspective, it is...

Antonio Santos: There are lots of expectations about artificial intelligence. Boards have expectations. How do we make sure that boards have the right information and are educated about artificial intelligence? I've sometimes had the impression that some executives make decisions from reading reports and talking with analysts and specialists, without being hands on with artificial intelligence themselves. How do we keep expectations at the right level while we support the board to make wise decisions?

Derek Hickey: It's a good question, and most boards today are, I won't say struggling with it, but they're certainly discussing it. The role of the board has changed over the last number of years, along with their responsibilities, and artificial intelligence has brought that to bear, given the potential risks and also the potential wins.

How I've managed it, with the board members I work with very closely, is to make sure the board is aware of the frameworks and controls in place. That is absolutely key. It matters for board reporting, and it matters to keep your organisation safe.

Be really transparent about where you are in your artificial intelligence journey. A lot of boards and chief executives got very excited, seeing this as a silver bullet. It isn't. In 99% of cases it's not a silver bullet, because, as I keep going back to, if the data is not in the right place, your artificial intelligence will not be as effective as you'd want it to be. So you have to be very transparent about where you really are.

The next piece is strategy. Work closely with your board on strategy, and on how you get from where you are to where everybody wants to go. Be very clear about the key steps. One of the worst things anyone can do is assume they're in the right place, or underestimate how long it will take to reach a position where they can use artificial intelligence in a much greater form.

Keep open lines of communication with the board. I meet members of our board on a regular basis, and artificial intelligence is always on the table for discussion. What are we doing? What's happening with it? What are the risks? What are the gains? We always keep an open mind. Sometimes it's a difficult conversation. Sometimes it's: we are not there yet, and we won't be there for 12 months, because we have to sort out all these other things first. If you keep those conversations going regularly, open and transparent, and flowing both ways, you're setting yourself up for success.

I do know of some organisations where board members, as you said, Antonio, have read an article, or sat on another board and heard what someone else is doing, and come back and said, we want to do X, Y and Z with artificial intelligence within the next 3 months. You have people sitting there thinking, no, we can't do that. The board doesn't understand why, because they haven't had those conversations, and they don't necessarily understand where people are in the journey of artificial intelligence adoption. So it's really important to have that open conversation.

Luis Lancos: Derek, which takes us to an interesting question. Is artificial intelligence driving the technology roadmap, or is technology driving the artificial intelligence roadmap? Some time ago I would have thought we should create the roadmap for artificial intelligence. But sometimes I'm not sure, because the market is evolving so fast that I have the feeling artificial intelligence is driving the roadmap. Which one is it, in your view?

Derek Hickey: I think fear is driving a lot of roadmaps today. Fear of missing out, fear of not adopting quickly enough. That's driving some roadmaps and strategies at the minute, and that's never a good way to be. Strategy should always be driven at the right level with the right focus. Technology, with artificial intelligence as part of it, then becomes part of the solution that delivers the strategy. If people flip that, they could end up creating solutions that then need to find problems. They could go the completely wrong way strategically, and spend a lot of time, effort and money on something that won't give them the benefit.

Anybody putting artificial intelligence into their strategy, and that's probably everybody, let's be honest, should do it the right way. Where do you want to be as an organisation? What's the focus, and what's the goal? What are the key strategic things you want to do to get there? Then what can technology do to support that, and how? Some of that will be through artificial intelligence, and some through other things.

That's the organisational strategy. Then there is the technology strategy, and there might be a specific strategy for artificial intelligence. That shouldn't be driven by fear, or by people saying we have to use artificial intelligence more and more. It should be driven by the organisation saying: we know where we need to go and what we need to do, we know exactly how we want to use artificial intelligence, and now we build towards that. So you might have a specific technology and artificial intelligence strategy, but at the organisation level, artificial intelligence is part of the overall strategy, or at least it should be.

Antonio Santos: Time flies, and we're about to close our conversation. Before that, I'll ask a question that maybe both of you can answer. Who grants the license to act? And what is your piece of advice for an executive who is about to approve the first artificial intelligence agent to be deployed in the organisation?

Derek Hickey: Good question. The license to act should come from the top down. It's important that organisations act together, act as one, when they talk about artificial intelligence and what they want to achieve. They might talk about it at a strategy level, or at a solution level, though be careful not to get the two mixed up. It has to be something the organisation wants to do. If one person goes off on a single mission to bring artificial intelligence into an organisation, it will be difficult to achieve, and you might do a lot of wrong before you do some right. That's not always bad, by the way, but you have to be careful.

One piece of advice before we end, from an executive side: have honest conversations about artificial intelligence. Don't be afraid to hear people push back, because you might not be in a good place. Don't try to adopt artificial intelligence just because everybody else is doing it. Make sure you're doing it for the right reasons. Make sure you have the right foundations and the right controls in place. Not having those simple things, which you would expect for every other part of your organisation, can put you in a very precarious position.

So have the transparent conversations, have the tough conversations, and build in all the attributes you would expect for any other conversation outside of artificial intelligence. Really important.

Luis Lancos: Fantastic conversation. To tell the truth, it was a short conversation. You know that feeling that we could carry on? We have so many things to go through.

Derek Hickey: Too short.

Luis Lancos: It was fantastic to have you here. I really appreciate it. A big thank you to you, and I really enjoyed this whole conversation.

Derek Hickey: As did I. Guys, great seeing you. I really appreciate you having me on. Thanks so much.

Antonio Santos: Thank you, Derek. And to all of you out there who might be listening to us, please feel free to reach out to us on LinkedIn. I'm sure you will be able to find me, Derek and Luis there, in case you have any questions or just want to engage. Thank you so much, Derek, once more. Thank you, Luis, and have a great day.

Derek Hickey: Thanks, guys. See you. Bye-bye.

Luis Lancos: Thank you. Cheers. Bye.