From proof of concept to scaling - the agentic reality check
Deloitte has released the 17th edition of Tech Trends, its annual survey that tracks the important changes in technology. This year’s survey focuses on five trends but for our purpose, we will focus on one—the agentic reality check.
The new workforce
This section starts with data from Gartner, which predicts that 15% of day-to-day work decisions will be made autonomously through agentic AI by 2028, up from none in 2024. And that 33% of enterprise software applications will include agentic AI by the same timeframe, compared with less than 1% today.
Nonetheless, businesses are encountering significant obstacles in translating agentic pilots into production-ready solutions. The diagram below shows that only 11% are actively using these systems in production.
From proof of concept to scaling
The new workforce
So why is this the case that only 11% are using this in production? Here are some the reasons:
- Legacy system integration. Most agents still rely on APIs and conventional data pipelines –this issue has come up in EVERY survey we have reviewed. And according to Gartner, over 40% of agentic AI projects will fail by 2027 because legacy systems can’t support modern AI execution demands.
- Data architecture constraints. Most organizational data isn’t positioned to be consumed by agents that need to understand business context and make decisions. Deloitte cites another survey it published, where 48% of organizations cited searchability of data and 47% cited reusability of data as challenges to their AI automation strategy.
- Governance and control frameworks. Businesses struggle to establish appropriate oversight for systems designed to operate autonomously as traditional IT governance models DON’T account for AI systems that make independent decisions and take actions. What is quite poignant is that many organizations attempt to automate current processes rather than reimagine workflows for an agentic environment—many so-called agentic initiatives are actually automation use cases in disguise.
The survey shows that some companies are solving for this problem by redesigning processes to be agent-native and not simply layering agents onto existing workflows—which is the solution many businesses use. However, how many of these have the resources to redesign processes, modernize architecture and adopt new governance frameworks like the likes of HPE and Intel are doing?
.