AI Agents Could Unlock $450B by 2028, But Businesses Aren’t Ready Yet

AI agents are evolving from just tools into collaborative teammates, reshaping the way businesses operate and compete. Yet despite their potential, trust issues are holding organizations back, with only 2 percent having fully scaled their deployment of Agentic AI.

According to a new report from the Capgemini Research Institute, AI agents are emerging as a transformative force in enterprise operations, with the potential to unlock $450 billion in economic value by 2028. This includes gains from both revenue growth and cost reduction, especially in IT, customer service, and sales. 93 percent of surveyed leaders believe that organizations that scale AI agents within the next year are expected to gain a significant competitive edge.

Despite growing adoption, trust in fully autonomous AI agents is declining, dropping from 43% to 27% in just one year. As it turns out, ethical concerns (such as privacy, bias, and explainability) and lack of transparency are major barriers. The report suggests that only 14% of organizations have fully integrated ethical AI principles.

“Trust is a major hurdle. Businesses need confidence in AI systems before granting them any level of autonomy. One way to build trust is starting with what I call ‘read-only’ AI implementations, where AI provides recommendations, but humans make all final decisions,” Dr. Walter Sun, SVP, Global Head of AI, SAP.

Adoption is rising, but autonomy remains low

The report indicates that 14 percent of organizations have implemented AI agents at scale or partially, but most agents still operate at low autonomy levels. Currently, only a small share of business processes are managed by agents with Level 3 or higher autonomy. This is expected to rise to just 25 percent by 2028, which reflects a cautious approach to full automation.

A major barrier to scaling AI is the lack of readiness in data and technology infrastructure. Most organizations report low maturity in areas like data integration, governance, and AI infrastructure. Moreover, over 60% of organizations expect to have human-agent teams within a year, where agents support or augment human roles. This change is expected to enhance productivity, creativity, and employee satisfaction.

“The economic potential of AI agents is significant, but realizing this value depends on more than just the technology, it requires a comprehensive and strategic transformation across people, processes and systems,” Franck Greverie, chief portfolio & technology officer at Capgemini.

Lastly, more than half of organizations believe AI agents will eliminate more jobs than they create, and 61% report employee anxiety about their future roles. However, only some of them are taking proactive steps to reskill or restructure their workforce.

Strategic priorities for IT leaders to unlock the full potential of Agentic AI

Here are the key recommendations for IT leaders from the Capgemini report:

1. Redesign processes around AI Agents

IT leaders should rethink workflows from the ground up and integrate AI agents as core components. This involves orchestrating AI agents alongside Gen AI, traditional AI, and automation tools to manage end-to-end processes.

2. Transform workforce and organizational structures

AI agents should be treated as team members, not just tools. Organizations must redefine roles, create new positions like AI agent supervisors, and embed agents into team structures.

3. Balance autonomy with human oversight

IT leaders should determine the right level of autonomy for each process, and decisions should be categorized based on risk, reversibility, ethical implications, and compliance needs. They must build human oversight into important points, with clear escalation paths and override capabilities.

4. Strengthen data and technology foundations

IT leaders should ensure high-quality, integrated, and secure data systems and invest in scalable computing resources. It’s important to ensure interoperability across platforms and adopt modern architectures like vector databases and real-time pipelines.

5. Choose the right technology strategy

Organizations must decide whether to build, buy, or adopt a hybrid approach for AI agents. This decision should be based on integration ease, customization needs, cost control, and privacy requirements.