AI Agents Aren’t Delivering the Expected ROI Because Enterprises Don’t Understand Their Own Business, Says ARIS CEO
New thought leadership released ahead of AI4 argues operational context – not larger models – is the missing ingredient for enterprise-scale Agentic AI
LAS VEGAS, NV, 3 August 2026
As thousands of business and technology leaders gather at America’s largest AI conference, ARIS CEO Guillaume Bacuvier is challenging one of the biggest assumptions shaping enterprise AI adoption.
In a new thought leadership article, “Before AI Agents Can Transform Your Business, They Need to Understand It,” Bacuvier argues that the biggest obstacle to scaling AI isn’t model capability – it’s a lack of operational understanding.
“Large language models already possess extraordinary capabilities. What they don’t possess is an understanding of how your business actually works,” says Bacuvier. “Enterprises don’t need smarter AI. They need AI that understands their processes, policies, governance and decision-making.”
While much of the AI industry is focused on making agents more capable or adding guardrails after deployment, ARIS argues that organizations should instead first create a governed operational foundation before agents can reliably execute work across complex enterprises.
According to ARIS, enterprise AI initiatives often struggle because organizations underestimate the complexity hidden inside their own operations.
Global enterprises operate thousands of interconnected processes spanning business functions, IT systems, regulations, decision rights and governance policies. Without understanding those relationships, AI agents may perform well in controlled demonstrations but struggle when confronted with the realities of enterprise operations.
“Perhaps the biggest misconception about Agentic AI is that success can be measured by the number of agents deployed,” says Bacuvier. “This is a fallacy – the real question is whether business outcomes improve. Are customer journeys faster and costs lower? Has compliance improved and risk been reduced?”
His article argues that organizations creating the greatest return from AI share a common characteristic: they prepare their operational foundations before deploying autonomous agents.
That means:
- Creating end-to-end visibility across enterprise processes
- Defining decision rights and governance before automation
- Embedding compliance into operational workflows
- Measuring business outcomes – not simply AI adoption
This approach is already delivering measurable business results at leading enterprise companies. Boots UK built a connected process architecture covering more than 2,000 business processes, enabling one finance process to be redesigned from 220 steps to just 40, reducing execution time by 75% while establishing the operational foundations for enterpr
Bacuvier adds: “You can deploy AI agents without building a comprehensive operational model but that approach only works if your ambitions are limited or your business is relatively simple. Once you’re operating at enterprise scale, operational understanding and context stops being optional. It becomes the prerequisite for trusted, effective AI.”
To learn more about how ARIS process context platform enables Fortune 500 and G2000 companies to deliver the ROI on AI, visit us at Booth xx at Ai4.
About ARIS
ARIS is the process context foundation platform for enterprise AI deployment. Combining process mining, modelling, and analysis in a single unified platform, ARIS helps leading global organisations move from AI experimentation to scalable execution – driving efficiency, reducing risk, and delivering measurable business outcomes.
For more information, visit www.aris.com
Media Contact
Andrew Tongue
Head of Content, Media and PR
andrew.tongue@aris.com