The actual rate at which it is imagined artificial intelligence (AI) will transform business processes is likely to exceed reality for several years to come, as the transition from predictive and generative to truly assistive AI takes more time than many today fully appreciate.
Camunda CEO Jakob Freund told attendees at the CamundaCon 2023 event today that early next year the company will leverage generative AI technologies to create co-pilots that will make it simpler to build process models that will make it easier to automate processes using a natural language interface. However, it will take significantly longer to automate workflows using assistive AI technologies that automatically fill out a form, he noted.
Generative AI technologies today can automatically create the form, but the level of automation that many business executives today envision in terms of being able to automatically populate that form with actual data is still a way off, says Freund.
In fact, it’s likely there will be a period of disillusionment with AI once business executives realize the limitations of generative AI, he says. “It’s dreams versus reality,” says Freund. “Today there’s a lot of fantasy.”
Even as more end users are exposed to generative AI platforms the hype continues to build, but there will inevitably be a trough of disillusionment, notes Freund. However, after that period there will be substantive advances that will lead to massive gains in business process efficiency, he adds.
As a provider of a platform for orchestrating business processes, the Camunda platform is already employed by 546 organizations, including CapitalOne, First American, Lowe’s, NatWest Bank, Sonexus Health, The Municipality of Munich, Urban Tech Hero and Walmart, to drive more than 5,000 business use cases.
Most of those use cases will inevitably be augmented by AI, but building and testing AI models for specific processes requires time and expertise. Organizations will also need to determine the degree to which they can employ a public AI platform versus needing to customize an existing one or build their own based on the sensitivity of the data used to train them. In addition, data privacy regulations may also require organizations to always retain control of the data used to train an AI model.
It’s also not clear whether those AI models will be deployed in the cloud and accessed via a managed service or deployed by an internal IT team in an on-premises IT environment where much of the data that large organizations store and analyze still resides.
Each organization will need to decide for themselves what type of AI to embrace when, based on what makes the most financial sense. The one thing that is for certain is that rather than being an event, AI will be more of a journey that organizations should prepare to be on for many years to come.