GenAI, Process Intelligence, and NHS case study – with IBM

The launch of ChatGPT in 2023 captured the business world’s attention. Very quickly, everyone looked for ways to make the promise of GenAI real. Many of them did just that – using AI chatbots to deal with quick, daily admin tasks as well as coming up with smart solutions to problematic use cases.

So with enterprises of all sizes, sectors, and industries on the hunt for AI best practice, IBM’s AI and Analystics Practice Leader Stephan Bloehdorn recently spoke to leaders from Reckitt, Carrefour, and a leading logistics firm about unleashing GenAI. Scroll down a little further to read about the NHS use case: 'Improving NHS healthcare services with GenAI and process mining.'

Where Reckitt, Carrefour, DHL Supply Chain started with GenAI

The discussion began with covering how the three leaders started using GenAI where they work. For Reckitt’s Kuldeep Dudeja, initial uses included content summarization and generation. Speaking in November 2023 he mentioned Reckitt is going purposefully slow and creating responsible AI principles to guide any further rollouts.

Aside from using it personally for daily tasks like taking meeting minutes, Florain Tué from Carrefour said the multinational retail and wholesaling firm uses it to compare supplier quotes – and has seen a huge gain already in terms of productivity.

"We realized comparing three quotes would take around 30 minutes for a buyer to do it manually," he said. "If we do it with ChatGPT and the POC that we've been running, it takes only 10 minutes. So the 20 minutes productivity gain is huge for our organization, and that's why we'll be pushing it so much.Bastian Schäffer shared how DHL Supply Chain is building use cases on its own platform, including using GenAI to inform the process of designing a warehouse.

• Watch ‘Generative AI Unleashed’ panel discussion here

What GenAI has to do with Process Intelligence

GenAI cannot operate in isolation from the rest of the tech stack. By feeding AI the intelligence it needs with Process Intelligence, AI can get processes to really work. Bastian neatly summed up the connection between AI and Process Intelligence, speaking about how AI can help prepare the data for process mining – a crucial part of Process Intelligene – as well as assist with the mapping, the modelling, and the analysis.

A ‘tremendous opportunity’ exists, said Kuldeep, to have GenAI connect and improve the business frontend and the backend. The former, he said, refes to enabling users to make decisions based on more reliable data. While the backend refers to how GenAI can enable more process mining capabilities by easily ingesting unstructured data, therefore improving data accuracy and completeness.

How GenAI is creating value with Process Intelligence

Everyone is learning on the job when it comes to GenAI. For Florian, the key is to focus on creating 'productivity hacks' by using the technology. Bastian agreed and told the audience it’s also about adopting a mindset that embraces AI and its ability to do some aspects of work better than people.

He shared how DHL Supply Chain is using GenAI to monitor contract obligations. The logistics company deploys GenAI to compare current company performance against contract agreements, and asks where risks to meeting contract obligations could occur – and gets the answers.

How GenAI creates value with Process Intelligence, said Kuldeep, should begin with teams asking three crucial questions: ‘Can we do it?’ ‘Should we do it?’ ‘And what if we do it?’ The most important, he added, is the second one as tight attention must be paid to data security. Reckitt so far has used GenAI to go from three-way to four-way match in P2P. Kuldeep explained how the multinational consumer goods company uses GenAI to extract data from thousands of pages of contracts and compares that information with what’s in the master data, the PO, and the invoice.

• Celosphere 2023: Everything you need to know news, product and process intelligence innovation

Improving NHS healthcare services with GenAI and process mining

The GenAI panel discussion was followed later that day at Celosphere by a presentation on using AI and process mining to alleviate the pressure on patients and practitioners alike.

"We use Celonis to pinpoint and identify the areas we can harness technology like GenAI,” began Lucy Marsden (IBM), before handing over to Dan Hayes (University Hospitals Coventry and Warwickshire) who explained an outpatient appointments use case.

The team thought communications between the General Practitioner and the patient were straightforward but “we found patients had thousands of touchpoints", he said. "There was a lot of waste in the processes. And we identified areas of waste that we probably wouldn’t have looked at without process mining."

NHS reduced ‘did not attend’ appointment rate by six points

With 650,000 patients not turning up to appointments per month across the NHS, Dan wanted to take a small step at his local trust to create big results. By sending clear, consistent communications about upcoming appointments, the team has done just that. “We found texting people on day 14 [days before the appointment] was most impactful,” said Dan. “This meant they could cancel earlier and we had the opportunity to re-book the appointment. By making a slight adjustment we massively improved performance… AI gives us a far bigger picture of our patient demographics, rather than just booking chronologically.”

The initial trial saw the ‘did not attend’ rate fall from 10% of appointments to 4% in just four weeks – meaning 900 more patients were seen each week. Dan and the team then began to look into grouping appointments for people needing treatment for more than one issue (roughly 17,000 at University Hospitals Coventry and Warwickshire) and for parents/guardians and their children.

By making it possible for partners/guardians and their children to be seen at similar times in the same location, it reduces inconveniences like the need to take extra time off work. As a direct consequence, this could lead to even fewer cancelled appointments. At the time of writing, this trial is ongoing.

The partnership between the NHS, IBM, and Celonis teams has so far been a force for good, said Dan. “It didn’t feel like three different teams operating, it felt like one.”

Edward Baggaley

Content Marketing Lead

Edward writes about Celonis, its customers, partners, and product. He creates blogs - perhaps the one you’re reading - as well as ads, ebooks, keynotes, and advertorials. Newsweek, The Times, Time, and many B2B magazines have published his work.