Enterprise firms all over the world have ploughed billions into their AI ambitions, a new study from Deloitte argues that businesses still stand at the “untapped edge” of the technology’s potential, with most stuck in pilot mode.

Based on a survey of over 3,000 executives worldwide, Deloitte’s latest State of AI in the Enterprise report found that only a quarter of firms have moved even 40% of their AI pilots into production.

This bottleneck is being driven by competing priorities, with companies wanting to invest in innovation to stay competitive, but tied to running their core operations on legacy stacks. As a result, just 30% of organisations have so far been able to redesign key processes around AI, while 37% report only using the tech at a “surface level”.

Deloitte found that those moving AI deployments past experiment mode were seeing tangible gains, with 25% of leaders reporting that the tech is having a transformative effect on their companies – more than double from a year ago.

The report found that the firms gaining the most momentum are those expanding AI access across their entire workforce. Staff use of AI tools jumped 50% last year, a shift that has helped double the number of companies expecting to move projects into production within six months.

However, while businesses are emphasising AI literacy, a broad skills gap is still viewed as the biggest barrier to integration, with 84% of firms saying they have not been able to redesign jobs around AI’s capabilities, yet fewer than half are making significant adjustments to their talent strategies.

“Organisations succeeding with AI aren’t just investing in automation and algorithms, they’re investing in their people,” said Jorg Schalekamp, Deloitte Nederland AI Lead. “By strengthening both the capabilities of their talent and their AI tools, teams can adopt new ways of working and create sustainable competitive advantage.”

Recommended reading

• New KPMG Report Demonstrates AI Scaling Challenges for Firms
• 95% of AI Pilots Deliver Zero ROI, Finds MIT Study
• Microsoft CEO Warns AI Risks Becoming a Speculative Bubble

Looking ahead, agentic AI looks poised for growth, with Deloitte finding close to three-quarters of companies planning to deploy agentic systems within two years, though only 21% of those firms report having a mature modelin place for agent governance.

Interest in physical AI is also on the rise, with more than half (58%) of companies reporting some use of these technologies, with the report predicting this to hit 80% within the next two years.

The Asia Pacific region is emerging as the frontrunner in this field, rolling out robotics, autonomous vehicles, and drones at a scale likely to shape the next wave of industrial automation.