Agentic AI raises concerns over enterprise risks & value

A service provider has voiced concerns over the practical benefits and risks associated with Agentic AI as spending increases on automated systems promoted as capable of acting autonomously in large enterprise settings.

Dan French, Co-Founder and Chief Executive Officer of Consider Solutions, has observed a growing pressure among Chief Financial Officers and Chief Information Officers who feel compelled to invest in new software advertised as able to operate independently across business processes.

French said: "The jury is still out on the true profit and loss benefits of GenAI, although the initial frenetic excitement has slowed. The next big thing, Agentic AI, has its own challenges and, as always, should be tested in safe spaces to protect organisations from tangible operational, financial and reputational risks."

He raised concerns that while some organisations may have the capacity to undertake internal proof-of-concept projects for Agentic AI, deploying these systems at scale within an enterprise presents complexities that go beyond automation. "If you have the funds and resources for an internal Agentic AI proof of concept that's great, but integrating complex tasks in an enterprise is much more than just an automation problem - it's a process problem, a human problem and an accountability problem. Not every task can be treated as a probabilistic or statistical problem, and failing to grasp this can run the risk of frustrating internal and external customers," French explained.

As businesses explore the use of AI 'agents' that act without human input, French noted a rise in scepticism about the suitability of off-the-shelf Agentic AI systems, particularly when integration with existing enterprise applications is not fully understood.

He referenced previous experiences with digital bots, highlighting ongoing questions about oversight and accountability for agent-driven decisions. The suggestion that such agents may one day function as digital employees is tempered by the recognition that human judgment plays a vital role in critical organisational decisions. French stated that, "Important decisions require a human to remain in the loop because critical decision making remains a nuanced human speciality – supported by digitised data among other factors."

French advised leaders to prioritise internal conditions necessary for successful AI adoption, including the readiness of technology infrastructure, the organisation of data, and clear structures for human accountability. He cited practices at technology-driven companies where multiple review cycles are implemented to check AI outputs for errors in critical processes.

"Agentic AI is a nascent technology, still very much in its infancy and we're a long way off from confidently adopting a hands-off approach and it being able to deliver real business value. Plus, like Analytical and Generative AI, it is not a silver bullet that will replace human interaction. Augmentation is much more likely to be the order of the day," French said.

He pointed to the underlying barriers to productivity within enterprises, noting that these must be tackled before AI can deliver on its potential. "We first need to focus on the real barriers to enterprise productivity, such as end-to-end process complexity – a critical enabler for business streamlining and optimisation – as well as performance measures and the human behavioural challenges, habits and cognitive biases," he said.

French also underscored the importance of developing capabilities that artificial intelligence cannot replicate, such as empathy, practical expertise, collaboration, and creative thinking. He mentioned that these skills are vital, along with an understanding of how digital technology can be used more broadly to drive efficiencies, productivity, customer engagement, and cash flow improvements. "We also need to be nurturing skills that AI can't replicate - whether that's empathy, practical expertise, collaboration, or the ability to think creatively beyond data – and the broader potential of digital technology and how that can drive progress in efficiencies, productivity gains, customer engagement and improved cash flow," French concluded.