- Agent AI implementation is slow, but the technology itself is not to blame
- Privacy, compliance, management and skills shortages are creating obstacles
- Dynatrace says the way forward is to redefine ROI and focus on human-machine collaboration
A new report from Dynatrace claims that around half of agent AI initiatives are still in pilot or proof-of-concept stages, showing how organizations are struggling to move from experimentation to full implementation, preventing them from achieving the return on investment (ROI) they are aiming for.
But the value of AI is not the factor in question; rather, it is barriers such as governance and security that are causing delays. Additionally, one in three cite the lack of a clear business case as a barrier to progress.
But companies are not discouraged: three-quarters (74%) expect to increase agent AI budgets next year.
These are the main barriers to agent AI and how to overcome them
The most important implementation areas today are IT operations and DevOps (72%), software engineering (56%) and customer support (51%), however the Dynatrace report reveals a disparity between investment approaches and where companies expect to see the greatest return on investment. Instead, the best returns are expected to come from IT operations and systems monitoring (44%), cybersecurity (27%), and data processing and reporting (25%).
The study details some of the most preventative barriers, including security, privacy and compliance concerns (shared by 52% of respondents), difficulty managing and monitoring agents at scale (51%), and shortages of qualified staff or training (44%).
Business leaders also highlighted the importance of human workers in an agent-driven world, predicting a 50:50 split for IT and routine support tasks. At the moment, about two-thirds (69%) of AI decisions are still verified by humans, and 87% are creating AI agents that require human supervision.
Another quarter (23%) prefer to rely solely on human-supervised agents.
Looking ahead, Dynatrace’s recommendations include reconsidering metrics and redefining ROI, establishing clear barriers to human-machine collaboration, and scaling slowly with intention rather than wasting large amounts of cash with varying degrees of success.
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds. Be sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form and receive regular updates from us on WhatsApp also.
