Process Intelligence: Unlocking Artificial Intelligence for Efficient Enterprise

By Cong Yu, VP of Engineering at Celonis.

  • 1 year ago Posted in

There’s a lot of excitement around generative AI since it is a genuinely huge opportunity for business, but only when it can be supported by process intelligence, which provides the data and knowledge foundation for generative AI to provide trustworthy, fresh, and tailored results.

The AI renaissance over the past year has led many business leaders to hope that they can use AI techniques such as large language models (LLMs) to improve efficiency and productivity. The amazing conversational capabilities exhibited by recent LLMs have made it no longer a topic of IT and senior leadership alone. Department and functional leaders are now asking themselves how they can leverage LLMs, which AI solutions they can use to effectively automate business processes, and how to get better and more accurate recommendations.

The success of AI relies on data

Like any other technology, AI must be enabled to be successful. The right data and knowledge foundation must be in place or it will fail to live up to even a fraction of its potential when deployed. AI’s performance is only as good as the data that feeds it. LLMs without process intelligence don’t live up to their full potential. Process mining, analysing processes throughout an organisation, offers business leaders the best data possible to feed into AI in order to deliver results. LLMs and operational AI solutions both have enormous potential to transform enterprises, and both work best when paired with process intelligence.

For business leaders, the question of which data to feed the LLM is crucial. LLMs, which are trained on huge amounts of noisy data from the Web, can ‘hallucinate’, namely making up falsehoods or answering questions, confidently, even when they do not know the facts. They can also make more subtle errors, pulling stale information from older sources, or giving very generic answers. This is where process intelligence comes in - it ensures that LLMs have the solid foundation needed to be correct and fresh, empowering them to reach their full potential.

Microsoft CEO Satya Nadella recently said that process mining capabilities are helping organisations ‘build their AI advantage’. Using data and knowledge from process mining ‘grounds’ the LLM in factuality, and paves the way for organisations to unlock the full power of artificial intelligence. Put simply, generative AI is only as good as the data it’s fed on and the knowledge it’s working with.

Why process mining matters

Process mining acts like an ‘X-Ray’ of an organisation, allowing businesses to take an overview of processes from end-to-end, highlighting problems and bottlenecks. The data and knowledge delivered from process mining is the perfect tool to help business leaders to identify where to deploy AI systems to augment the abilities of human workers.

Object-centric process mining (OCPM) is even more powerful. It takes a ‘three-dimensional’ approach to process mining, offering an overview across objects within multiple different systems such as Enterprise Resource Planning (ERP) and Customer Relationship Management (CRM) software. OCPM unleashes the full power of process mining when it comes to AI. Working as a ‘digital twin’ of the whole enterprise, it allows business leaders to analyse the interactions of multiple complex systems. OCPM works as a ‘single source of truth’ across the whole organisation. This is incredibly important when dealing with AI techniques.

Not just for company leadership

AI is no longer simply a topic of interest for IT and senior leadership. The simplicity of conversational interface means that department heads in everything from supply chain to shared services can use LLMs directly to gain business insights. This also means that access to process intelligence is even more important. AI can truly democratise and disseminate knowledge and insights across the whole organisation, based on process intelligence as a crucial backbone. The ease-of-use of AI prompts means that employees no longer need to be trained in technical skills to run analyses of data, allowing more employees to find more insights. AI can also suggest jumping off points for further analysis, dramatically accelerating time-to-value.

Imagine being able to ask the LLM questions such as, ‘What caused a decrease in on-time payments over the last four weeks?’ or even, ‘How can I save £100,000 next month?’. Grounded in process intelligence, less technically literate users can extract business value rapidly, without waiting for help from a technical analyst. This in turn means that the organisation's analysts can increase their impact, configuring an LLM once, rather than building hundreds of custom dashboards over and over again. Process intelligence can also cut the time analysts spend in preparing data for use - which accounts for most of the effort in data science - by ‘steering’ the AI towards the right knowledge it needs to work well.

Making things work better, faster

Apart from delivering insights more rapidly, AI also has potential for businesses in using intelligent automation to perform tasks more efficiently. Process mining again has a role to play here. Operational AI solutions thrive when fed the full knowledge and context around how business processes run.

If you’re employing AI tools to augment Accounts Payable teams, for example, those tools need process intelligence to understand the team’s work in the context of the whole business, otherwise they can stumble when confronted with deviations such as duplicate invoices. Process mining allows business leaders to create higher quality data models much faster, enabling access to more mature tools. In turn, this significantly shortens time to value for new processes.

The optimal combination

LLMs without process intelligence are effectively just somewhat-easier-to-use business intelligence tools. But businesses that use process intelligence to power AI deployments, whether LLMs or operational AI, will see the bigger benefits faster, and have improved productivity compared to rivals that fail to do so. Process intelligence offers business leaders a way to offer AI tools the ‘full picture’ of the business. LLMs fed with process intelligence offer better insights, and require less effort to deliver relevant information.

AI systems can produce better results, faster, with process intelligence as the foundation. For any business leader hoping to use AI to improve efficiency and productivity in their organisation, process intelligence should be the foundational stone of their approach. Paired together, AI and process intelligence have the potential for huge impact on how business processes are run and improved.

By Danny Lopez, CEO of Glasswall.
By Oz Olivo, VP, Product Management at Inrupt.
By Jason Beckett, Head of Technical Sales, Hitachi Vantara.
By Thomas Kiessling, CTO Siemens Smart Infrastructure & Gerhard Kress, SVP Xcelerator Portfolio...
By Dael Williamson, Chief Technology Officer EMEA at Databricks.
By Ramzi Charif, VP Technical Operations, EMEA, VIRTUS Data Centres.
Companies are facing a Catch 22 when it comes to the need to invest in new forms of AI, whilst...