Human-AI Trust to drive USD 450B agentic AI surge

A new report by the Capgemini Research Institute projects that agentic AI could generate up to USD 450 billion in economic value by 2028. However, this potential remains largely untapped, as only 2% of organisations have achieved full-scale implementation. The study emphasises that building trust and fostering human-AI collaboration will be critical to unlock this transformative value.

Despite growing investments, confidence in fully autonomous AI agents has fallen sharply, from 43% to 27% in just one year, amid rising concerns over ethics, data privacy, and explainability. Yet, AI is increasingly being viewed as a collaborative partner, with 90% of executives stating that human oversight is either positive or cost-neutral. Companies implementing human-AI teams report significant benefits, including a 65% increase in human engagement in high-value work, and improved creativity and job satisfaction.

The report, titled “Rise of Agentic AI: How Trust is the Key to Human-AI Collaboration”, finds that although agentic AI is gaining momentum, most firms are still in the planning or pilot phase. Only 14% have moved to implementation. Nevertheless, 93% of business leaders believe that scaling AI agents within the next 12 months will offer a competitive edge. However, nearly half lack a coherent strategy to make that a reality.

Transparency, ethical safeguards, and clarity in decision-making are now central to building trust in AI systems. Organisations already implementing AI agents report higher trust levels, 47% compared to 37% in exploratory phases. Businesses are also rethinking structure: 70% say AI agents will drive organisational redesign.

The most common use cases are currently found in customer service, IT, and sales, with broader expansion expected in R&D, operations, and marketing over the next three years. Yet, autonomy remains limited; only 15% of processes are semi or fully autonomous today, expected to rise to 25% by 2028.

A lack of AI infrastructure, poor data readiness, and persistent ethical concerns continue to hold back adoption. For example, although 51% cite privacy as a top issue, only 34% are actively addressing it.

To capitalise on agentic AI, the report urges firms to prioritise strategic, human-centric AI deployment over experimentation, balancing autonomy with trust and oversight. 

Read More : 

OutSystems launches agent workbench for enterprise AI

Securonix, LA Tech team up for AI-powered security

C-suite takes charge of OT cybersecurity: Fortinet

Okta and Palo Alto unite to secure digital identity