AI executives remain hesitant towards AI agents, potentially forfeiting substantial financial benefits
As more businesses explore the potential of AI agents, a new report by Capgemini sheds light on the key challenges and benefits associated with scaling their deployment.
The report reveals that only 14% of organizations have begun implementation of agentic AI, with fewer than a quarter having launched pilot programs. However, the proportion of agents functioning independently is expected to rise to a quarter by 2028. This growth is driven by the belief that scaling AI agents over the next 12 months will provide a significant competitive edge, with 93% of business leaders sharing this view.
Despite the potential benefits, the report identifies several challenges that must be addressed. Trust and confidence issues have emerged as a significant concern, with confidence in fully autonomous AI agents declining from 43% to 27%. This decline is due to privacy and ethical concerns, limiting broader trust and scalability.
Another challenge is the scaling barriers, as very few organizations (only 2%) have fully scaled AI agent deployment. The report suggests that many organizations shy away from open-source AI models due to concerns over security vulnerabilities, complexity, and lack of enterprise-grade support, favoring trusted proprietary solutions instead.
A solid data infrastructure is also essential for successful AI adoption, but many organizations lack this foundation. Effective integration often requires ongoing human oversight and collaboration, as purely autonomous AI agents are still not fully trusted or reliable in all contexts. Deploying and scaling AI agents often requires significant technical skills, which organizations may struggle to source or develop.
However, the report also highlights the numerous benefits of scaling AI agents. Organizations report an average return on investment of nearly 1.7 times from AI agents, with broader adoption expected to accelerate as these benefits become clearer. Agentic AI is projected to deliver up to $450 billion in economic value by 2028 through revenue gains and cost savings.
AI agents contribute to substantial improvements in key business functions, such as reducing case handling times and improving productivity. Human-agent collaboration can increase engagement in high-value tasks by about 65%, leveraging AI as an augmentation tool rather than just automation.
The real value lies in reimagining how work is done with AI, moving towards value- and insight-driven enterprises rather than solely focusing on cost reduction. A shift towards human-agent teams is also expected to result in a 53% rise in creativity and a 49% boost in employee satisfaction.
To address the challenges and realize the benefits, Capgemini recommends a pragmatic and integrated approach centered around trust and data. Prioritizing human-AI collaboration is crucial to maximize benefits while mitigating risks. Starting with pilot projects and platform investments can help build capabilities and identify impactful use cases before scaling.
Greverie, a Capgemini executive, emphasized the importance of reshaping organizations to support effective human-AI chemistry. Central to this transformation is the need to build trust in AI by ensuring it is developed responsibly, with ethics and safety baked in from the outset.
In conclusion, while the deployment of AI agents promises significant economic and operational benefits, scaling them requires addressing trust, data, and technical challenges with a pragmatic, human-centric strategy. Organizations that succeed in this transformation are projected to generate substantial returns over the next three years, with an average of $76 million for those that have scaled deployment and $382 million for those operating at semi-autonomous to fully autonomous levels.
Cybersecurity plays a vital role in the scales of AI agent deployment as the report suggests that many organizations shun open-source AI models due to concerns over security vulnerabilities, complexity, and lack of enterprise-grade support, favoring trusted proprietary solutions instead.
The integration of artificial-intelligence technology is projected to significantly impact various business functions, such as reducing case handling times, improving productivity, and increasing engagement in high-value tasks.
