90% of companies are committed to agentic AI, but there is one detail that sets them apart

This is clear from the new Process Optimization 2026 report carried out by Celonis, in which more than 1,600 global business leaders have shared their situation when it comes to investing in business AI.

The results show how the majority of the companies surveyed agree with a future driven by AI. So much so that 85 percent of organizations claim to aim to become “an agentic company” within three years.

However, faced with this goal, 76 percent also admit that their current processes are limiting their plans, in part, because they need operational processes that work and an operational context that “entrenches AI in the reality of their business,” as the company specialized in process intelligence has explained.

Thus, even though 90 percent of organizations are already using or exploring multi-agent systems to automate complex decision making. 82 percent of decision makers surveyed believe that AI will not deliver a good return on investment (ROI) if it does not understand how the business operates.

In this sense, Celonis has explained that for AI agents to act autonomously and effectively, they need processes optimized and ready to use with this technology, in addition to the process data and operational context that “only process intelligence can provide.”

Without these two factors, agents “cannot understand how a company really works or know how to improve it,” the technology company has determined, while ensuring that the lack of knowledge and context “are the biggest obstacles” to the adoption of AI.

In fact, the report’s results indicate that a lack of knowledge within the organization is one of the main impediments in 47 percent of companies. Likewise, the difficulty of making AI understand the business context is another brake for 45 percent of organizations.

Following this line, another point that the data reflects is that silos, that is, isolated information repositories that prevent the exchange, access and unified vision of data, also block the effective deployment of AI.

This is because, according to 58 percent of process and operations leader respondents, their departments “still do not operate fluidly with each other.” This prevents the end-to-end visibility necessary for effective enterprise AI.

LOOK: Google workers ask that the Pentagon not use Gemini for classified work

Additionally, the report also reflects a competitive urgency to adopt AI, as 89 percent of business leaders consider AI to be “their biggest opportunity to compete in the market.”

Taking all this into account, to close this gap between the potential of AI and the preparation required for its use, Celonis has stressed that organizations “must go beyond isolated automation.”

This is confirmed by the report’s findings, which suggest that for AI to navigate integrated into the business, instead of executing isolated tasks, “it must be anchored in process intelligence.” This will act as the “common language” that will allow AI agents to understand how work flows between departments and systems, as well as identify points of friction and execute actions that generate results for the business.

As Celonis president Carsten Thoma explained, “by using process intelligence to give AI a shared understanding of how a business works and how to improve it, we are finally turning that potential into continuous, measurable value.”