Lloyds Bank plc is a major British retail and commercial bank with a significant presence across England and Wales. It has traditionally been regarded as one of the “Big Four” clearing banks. With around 68,000 employees, it decided to deploy Workday HCM in 2016.
Enterprise Times spoke to Stuart Martin, Global People Operations and Services leader at Lloyds Banking Group and Carolyn Horne, SVP at Workday, about the deployment and how, more recently, it has been rolling out the AI capabilities within Workday.
Lloyds went live on Workday HCM in 2018. Two years later it deployed the Workday Skills Cloud. It continues to implement the new functionality that Workday has added. It is currently rolling out Workday Payroll for its UK operations.
Since 2018, Workday has leveraged AI within its HCM platform. One example is the ability to create customised learning journeys for staff. The platform suggests specific training based on the flow of an employees work. It has also used AI to surface trend analysis from an HR operational perspective.
Martin noted, “We’ve always had embedded AI built into the tool and used it quite selectively over the last few years and continue to do so.”
The emergence of Gen AI
More recently, it decided to deploy Gen AI in a safe and secure way. It currently has 300 employees using the Gen AI capabilities. Lloyd has a cautious approach with a very rigorous stage gate process.
I asked Martin how Lloyds are using it, he explained, “Our first use case is around the management of policy queries. We’ve selected a few policies that are high volume for our colleagues. For example, understanding leave and the multiple options around leave from a wellbeing perspective.”
Martin has other use cases lined up, but this pilot will enable Lloyds to use AI to provide clear, up-to-date, relevant policy information to employees.
Initially, Lloyds will focus on query management. At a later stage, they will look at transactional and more complex use cases. However, implementing Gen AI is more than just switching it on. Lloyds are sensibly ensuring that the data the Gen AI leverages is carefully curated.
It has created a multi-disciplinary team to create, update and curate the policies. Those policies have gone through changes recently to ensure that the policy information is both up to date with any new ways of working, digitally compliant and adheres to all the regulatory requirements.
Rather than just rely on text, the team approached the task using design thinking. It has broken the policies into bite-sized chunks and made better use of iconography. Everything is held internally and securely within the bank.
Martin added, “We currently use SharePoint, and that’s probably what we’ll continue to use. So, it’s a very closed, safe environment for our colleagues to operate in.”
The benefits of using Gen AI
What are the benefits you’re hoping to achieve? Quantifiable benefits?
Lloyds is already seeing quantifiable benefits from the pilot. About 12-15% of the policy enquiries have been answered by Gen AI. Overall, Martin is looking to achieve a 1% efficiency gain through enquiries being answered by Gen AI. He sees that as a huge efficiency gain.
He commented, “The plan is to give that efficiency gain back into the organization and to allow them to focus on other work. But we are quite confident and focused on making sure that any gains we make is ploughed back into the organization to use in a different way.”
This is just a pilot, and Martin made his objectives clear, saying, “Truthfully, right now, we want to make sure the tool works. We want to make sure our colleagues understand how to access it.”
Governance
As one might expect, Lloyds has an impressive multi-layered approach to the governance of AI. Martin shared that from an HR perspective, they ensure they meet UK regulatory requirements and align with EU regulations. They also heed the advice from Workday who have a much wider view of what enterprises are doing.
At a high level, Martin explained the Lloyds approach, “We entirely reviewed our governance processes and the way we introduce IT solutions into the business to ensure that we have these guardrails in place. We do that continuously so that we are proactive in that space.”
Internally it has set up multi-disciplinary forums that act as a gatekeeper for any new AI or Gen AI initiatives. These forums include people from across the organisation, including legal, risk and HR. So far, only two use cases of Gen AI have been approved across the whole organisation. The second was for HR.
On the ethics side, it was the first UK bank to hire a head of AI ethics into the organization. They serve on several of the AI governance forums.
I asked Martin about the risk of bias in the data. He replied, “From an HR perspective we are particularly focused on making sure we take out any implicit bias, any unconscious bias, out of the data.”
He explained how, adding, “Mostly, it’s what we call human in the loop. We eyeball all of the content coming with a group of people who are pretty much focused on that. It is checked and continuously cross-checked by a group of people.”
Lloyds has enhanced its AI expertise
Lloyds also recognised that AI is developing quickly. This presents both a risk and an opportunity. It is a complex piece of technology. To manage this development, it is has doubled the number of data scientists within the organisation in the last 24 months.
Martin explained how this growing team will help HR. “We’ve established our own Artificial Intelligence Center of expertise within the bank to help us really understand. These are people with deep knowledge and expertise. We in HR will draw on that set of expertise because these folks who can help us through some of the more complex questions that we may need to answer.”
What about other employees?
While bringing in external experts may help build AI within guardrails, Martin acknowledges that training the internal teams on understanding and using AI is important.
All three hundred senior leaders have undergone training. These are in the form of master classes that build levels of understanding, confidence and capability in that space. This is also continuous as AI and other technologies evolve to ensure everyone is up to date.
It has also rolled out very specific learning modules for all employees. These range from basic courses about data, data archiving, and data ethics to some more complex ones.
The intent is to ensure that the workforce fully understands what AI means for the business. In addition different courses have been designed and rolled out by function. Ensuring that, as AI is rolled out to HR, employees understand and are prepared to govern and use it.
Martin explained, “For example, as an HR leader, there is an expectation and rightly so, that I’m in a position to provide guidance, support and input into decisions around our workforce, where Gen AI is going to have an impact. I’m in the position of knowledge and understanding to help with that capability.”
How has this made a difference? “The overall answer would be enhanced governance, enhanced capability of our leadership team and our workforce to make sure that they understand that. The final bit has been just increasing our communication associated with Gen AI, so that we build a collaborative trust in that space.”
Challenges
With such a cautious and well-planned approach, the challenges that Lloyds faced in rolling out AI were few. Communication and transparency were key. Martin explained that they spoke to the employees and their unions about the direction they were taking, as well as the guardrails and ethical considerations. This structured approach may have meant that things have progressed slowly compared to some, but it also should mean that adoption is high and the benefits are there.
He added, “We’ve gone on a very structured way of making sure before we implement anything, we increase the level of capability and understanding of what it is and how we tend to use it.”
Martin acknowledged that there were concerns, explaining how the bank has overcome these. He added, “We’ve got working groups together, and we’ve worked our way through that and spoken to team members or folks who have these concerns. The focus has been on being as open and transparent as possible so that people understand where we’re headed. Where there is curiosity, we’re curious together.”
Hindsight reveals a lesson for others.
Martin also shared what Lloyds might have changed with hindsight on the Gen AI project. There was only one surprise, and that was the new skills required at an operational level. He admitted, however, that they needed to learn that.
He commented, “We started to see the need for new skills that we perhaps had not anticipated. So, prompt engineering capability. We are having to think now about who will manage our prompt library, etc. Not big issues, not showstoppers at all, but probably would have given it a bit more thought.”
The wider lessons of AI deployment
I asked Horne, when customers inquire about the use of AI, what are the most important questions that customers should ask before they get to that deployment stage?
She answered, “AI is not new to us. It has been embedded in our products for several years. When it came along, generative AI raised the profile of it.”
Horne believes that first, they have to really consider the use cases of AI. Secondly organisations need to consider their data, and its sources, because the quality of your AI is based purely on the data. Horne explained, “At Workday, we’re sitting on top of 75 million employees within our solution. That ensures a level of quality of data.”
On vendors, she feels organisations should consider their philosophy on AI, their governance around that and their ethical approach.
I asked Horne how customers select the right use case. “Many HR leaders are still trying to get their heads around that, but we would say you have to embrace AI. That’s the first thing.
“It doesn’t have to be perfect. Try out some use cases if you want to do that in a smaller way. Stuart mentioned the example of the 300 people for the policies. Do a pilot first, and check that you’re getting the results and the benefits that you’re looking for before you go into the bigger scale. We are hearing from leaders that it’s not a question of “if”. It’s a question of “when”.”
