Alan Petrie Head of Corporate Data and Analytics at Heathrow gives his thoughts on the role of AI in analytics and Oracle’s new AI Data Platform
Alan Petrie is a veteran analytics leader with 30 years’ experience of the industry. He was recently promoted to Head of Corporate Data and Analytics at Heathrow, where he leads the team that recently won the EMEA Oracle Excellence Award for Data Analytics and AI.
We caught up with Alan on the day the award was announced to ask his opinion of the emerging role of AI in data analytics, his plans to exploit AI at Heathrow, and his expectations for the newly launched Oracle AI Data Platform.
“The adoption in Fusion Applications is going well at Heathrow, with AI agents being embedded in many business processes to easily answer questions like ‘How many days’ leave have I got left?’, ‘what’s the maternity policy at Heathrow?’. I expect to see the applications team spin up many more of these agents, moving forward. It transforms the experience of using the applications.
Material benefits from automating Accounts Payable
“We are also working on cutting edge use of AI in the Corporate Data & Analytics team. In the Accounts Payable team previously, we manually matched invoices to POs to ensure we are paying out the correct amount, including VAT where appropriate. At best we were able to check around 3% of all invoices. Now, by using AI and displaying the results in a Fusion Data Intelligence (FDI) dashboard, we can check 100% — around 4,000 each month. The time and money saved is very significant.
“First, we use AI—Oracle OCI Document Understanding—to make the invoices completely machine readable, even multi-page documents. Then we can check, line item by line item, to make sure the invoice matches our corresponding PO. Any anomalies are instantly and automatically identified and flagged. Then we check that the VAT number is accurate on the invoice and automatically check back to Companies House, the UK corporate regulator, to ensure that the supplier is registered correctly. We can ensure that what we are paying, including VAT, is the correct amount versus the PO. All the details are displayed on our FDI dashboard, so we can instantly see any errors on any of the 4,000 invoices and investigate. We have already identified several other similar applications in other departments for this type of application of AI at Heathrow.
Understanding the human impact of absenteeism
“We are using AI to track trends on our Absence FDI dashboard, blending data from multiple systems. The absence system is on Databricks, the rostering system which is legacy, and other HR data coming from Fusion. The dashboard itself gives us a complete picture of who’s been absent, for how long, how many shifts have been lost. Basic stuff. We can also enable individual departments to see how their absence levels are fluctuating and why, and help them make comparisons between their department and others. For example, we employ over 4,000 security staff at Heathrow, and having a clear picture of absenteeism in that department is important because it affects the operations of every area of the airport.
How AI makes Finance faster
“In Finance, we’re looking at how we can use AI to manage procedures based on a wide variety of policy documents and emails to answer questions like, ‘In this particular situation, what should I do?’ So far, results have been very good—not always 100% accurate, but close. It saves an incredible amount of time, giving answers in 20 seconds that used to take hours, manually searching all the documents. We’ve tried it in Finance, and all departments have procedures that must be followed. HR policy documents are an obvious example.
Transforming work with conversational analytics
“We are very focused on how AI can help us make the experience of using FDI more conversational. When someone has a business problem to solve, it can be time consuming to have to go and search through dashboards to find the specific visualization to provide answers. They must know where to look. With AI and conversational interfaces, users should be able to say or type ‘show me the average salaries of my department over the past three years’, and the correct visualization is created for them on the fly.
“It’s going to be a huge transformation to the way people work. And it will tie into the Fusion Applications’ AI agents I mentioned earlier. Business users want to have just one interface where they can answer all their questions, from operational ones—‘How do I approve Alan’s requested time off next Tuesday?’ to understanding trends, ‘Who has the highest absence rate in my department over the past six months, and why?’ They’ll speak to this concierge AI that takes the request and passes it off to whichever agents are needed, potentially multiple agents that can talk to each other. Essentially, it’s like creating a new role in the company.
Managing AI initiatives requires a platform
“Such a broad approach will require a platform. The data will come from all sorts of disparate sources, in addition to Oracle Fusion. And there will be a need to manage all the different agents, to ensure they are working properly and that the right one is presented for each use case. The Oracle AI Data Platform plays an important role in this. Initially, we will use it to manage all the agents used with FDI and its data, and to make our agents available for use elsewhere in the company.
“With AI, there’s been 10 different ways to do something, and that’s resulted in a phase of fragmentation, with stuff everywhere and lots of different tools. I think it’s starting to come back together a little bit. And I’m looking forward to seeing what we can do with the tooling we have. And AI Data Platform is one of the big parts of that.”
Learn More
You can read more about Heathrow’s journey in their previous post, Europe’s largest airport makes use of its data to make better business decisions, and visit Oracle.com for more information about Oracle Analytics, AI, and AIDP.