Most workplace AI projects are stuck at the concept stage – so what can be done to improve this?

• Agent AI implementation is slow, but the technology itself is not to blame
• Privacy, compliance, governance and lack of skills all create obstacles
• Dynatrace says the way forward is to redefine ROI and focus on human-machine collaboration

A new Dynatrace report has claimed that around half of agent AI initiatives are still in proof-of-concept or pilot stages, showing how organizations are struggling to move from experimentation to full implementation, holding them back from the ROI they are aiming for.

But AI’s value isn’t the factor being questioned—rather, it’s barriers like governance and security that cause delays. Furthermore, one in three cites a lack of a clear business case as a barrier to progression.