How industries are pairing AI and process mining to drive transformation

The emergence of generative AI (gen AI) and large-language models (LLMs) such as ChatGPT, means the stakes have gotten higher and the race to find AI use cases that will deliver business value has intensified. The latest McKinsey global survey on the state of AI revealed that, less than a year after the emergence of gen AI, one third of organizations were already using it in at least one business function. What’s more, 40% planned to increase their overall investment in artificial intelligence because of advances in gen AI.

But there are well publicized drawbacks to artificial intelligence as a whole, and gen AI in particular. As gen AI has already reached the peak of inflated expectations on the Gartner AI hype cycle, it will be interesting to see how quickly it descends into the trough of disillusionment.

To drive truly valuable enterprise transformation, AI needs a partner. And process intelligence is the perfect contender.

“With Process Intelligence, processes don’t just run, they work for you,” wrote Alex Rinke, Co-founder and co-CEO of Celonis, in an open letter on AI to the process mining community. “It’s the layer that knows how your business flows. How processes interact and impact each other across every department, every system.”

AI as accelerator, Process Intelligence as enabler

AI offers tremendous potential. To reach that potential, it can’t be used in a vacuum.

Reliability (or a lack of it) is one of the biggest concerns plaguing artificial intelligence. Gen AI has brought these reliability concerns to the fore as it is frequently seen to state incorrect information as fact in what’s become known as a ‘hallucination’.

Timeliness is another potential issue. An LLM can take months to train and deploy and, although models such as ChatGPT have now been upgraded to run on real-time data, it is still difficult to be sure they are using the latest insights. Finally, security is a major worry. If businesses train AI and LLMs using customer data there’s a risk of this information being leaked to the model's users.

When it is paired with process mining, artificial intelligence suddenly becomes truly beneficial for business transformation as these issues can be overcome. Process mining generates accurate and up-to-date process data — also known as process intelligence — which can be used to feed AI. Process intelligence grounds the AI’s responses in the process context, is based on the latest information to ensure responses are accurate and current, and uses event log data rather than sensitive information such as customer data, which helps avoid privacy and security issues.

Process intelligence allows AI to understand any business process so it can speak the language of that business. It provides an enablement layer to power sophisticated AI solutions and replace simpler rules-based automation systems such as robotic process automation (RPA), or those using business intelligence. When process intelligence feeds an LLM, the model’s responses get smarter, faster and more accurate, so enterprises can make decisions and take action knowing what value it will bring.

How verticals are using AI and process mining

AI is already being used across a wide range of industry verticals, and the addition of process mining technology (and process intelligence), will make virtually every use case more effective.

According to respondents to the recent KPMG Generative AI Survey, March 2023, 83 percent said their finance and accounting functions were exploring generative AI use in financial forecasting and budgeting. The possibilities include integrating predictive models, creating scenarios, and generating insights on potential financial outcomes.

In a statement to Celonis, KPMG, LLP said:

“Moving forward, the focus of LLMs would be to develop “domain-specific” and “use case-specific models” that help unlock efficiency and productivity in new ways. Process mining could complement the use of generative AI by identifying true use cases to help generative AI capture real efficiencies.”

Let’s take a look at how AI is being used in retail, healthcare and the oil and gas verticals.

AI in retail and CPG

From fashion retailers and CPG brands to leading ecommerce platforms, AI is being used for demand forecasting, helping the retail industry accurately predict which products specific groups of customers will want, as well as where and when they will want them. Retailers are also using AI and ML for automated inventory management, product recommendations and customer sentiment analysis.

AI in life science and healthcare

Life sciences and healthcare organizations already rely heavily on AI. It is used in diagnostics, and particularly in early detection, for example to find patterns in patient data that could indicate the presence of a disease. Natural language processing (NLP) is increasingly being used to understand input from both patients and medical professionals. AI is also used in precision medicine, predicting the drugs that will work best for individual patients and tailoring treatments to their specific needs to improve outcomes. In addition, it is used to streamline the drug discovery process and enhance existing treatments.

AI in the oil and gas industry

As with retail, the oil and gas industry makes use of AI for demand forecasting, using vast volumes of data to understand what will be needed, where and when. This allows enterprises to strategically allocate resources, optimize operations, streamline supply chains, avoid unnecessary downtime and maximize profitability. AI is also used for oil and gas exploration, as well as to continually monitor critical infrastructure to improve safety and to enable predictive maintenance.

Functional use cases for AI and process mining

AI is also being used across a variety of business functions such as finance, supply chain and shared services. Once again, any of these use cases could be improved by feeding the technology with process intelligence that allows the AI to speak the business’ language. Business functions can use tools like the Machine Learning Workbench to make the most of process intelligence by building custom models and driving smarter decision making.

Finance AI use cases

Many finance departments are already using automation for time consuming, repetitive tasks such as inputting invoices, tracking receivables and logging transactions. But combining a process mining tool with AI can make the whole department smarter and more efficient.

In accounts receivable, for instance the two technologies can be used to centralize data across fragmented source systems and intelligently prioritize collector actions. In accounts payable they can be used to intelligently prioritize critical invoices and resolve payment blocks. This ensures invoices are processed and paid at the right time to avoid penalties and take advantage of discounts, without unnecessarily reducing DPO (days payable outstanding).

Supply chain AI use cases

A key use case for AI in supply chain functions is the employment of predictive analytics to analyze patterns and trends, allowing it to forecast downstream demand or upstream shortages. In the short term AI can be used to flag events it recognizes as precursors to a supply chain failure so action can be taken to minimize its impact. In the longer term it can be used for proactive planning, process improvement and risk mitigation. AI can also be used to track materials, components and products throughout the supply chain, from procurement and manufacturing to distribution and delivery.

Shared services AI use cases

When enterprises consolidate functions such as customer service, HR, marketing and legal into a shared services center, AI is frequently used to automate tasks, increase efficiency and improve the customer experience.

For customer services this could mean using chatbots powered by gen AI to supplement human agents. For marketing it could mean using AI to generate customized content and experiences for individual customers. HR departments in shared services frequently use AI to streamline the recruitment process, perhaps by generating job descriptions and scheduling interviews, while legal teams use gen AI for document drafting.

The process mining artificial intelligence partnership in action

To see how the combination of process mining and AI works in the real world, take a look at these customer stories, which span the retail, insurance and telco industries.

Ocado Group puts AI and process mining to work

UK retailer Ocado Group puts technology to work in everything they do. At the heart of their operations is their Ocado Smart Platform (OSP) — an end-to-end commerce, fulfillment and logistics platform that combines robotics, AI and machine learning to streamline the process from order placed to goods delivered. The group operates fully automated fulfillment centers around the world, each of which can generate 75 billion data events every day.

Discover how Celonis’ process intelligence complements Ocado’s high-tech portfolio, enabling the company to drive their partner's success, deliver process optimization and unlock cash value — with a projected hundredfold return on investment.

ERGO increases its use of process intelligence

ERGO is one of the major insurance groups in Germany and Europe, and uses a combination of process mining and AI to make its processes more customer-friendly, for instance by reducing claim settlement times.

Find out how ERGO intends to use process intelligence to optimize core areas and further drive the group’s digital transformation, while opening up new use cases for applications that have already been successfully implemented, such as voice, artificial intelligence and robotics. This will increase customer satisfaction, quality, efficiency and productivity.