A global survey of 501 IT professionals published today finds that while nearly all (97%) have multiple orchestration tools, they are encountering operational friction that can be traced back to interoperability issues, with 80% noting their orchestration tools make it harder, not easier, to gain clear end‑to‑end visibility across their IT environments.

Conducted by Broadcom, the survey finds that 60% work for organizations that have deployed four or more orchestration tools.

Aline Gerew, head of automation for the Agile Operations Division of Broadcom, said many of those organizations are, as a result, going to, by definition, encounter challenges meeting service level agreements (SLAs).

For example, even though 84% of respondents said they are confident they can meet SLAs, 57% admit missing orchestration deadlines every month. A full 89% also report experiencing audit issues related to orchestration, with only 54% having proper audit trails.

Just as challenging, less than half (46%) track the return on investment (ROI) of orchestration initiatives, and 53% do not allocate or show orchestration costs back to the business.

All of these issues are going to only become more challenging to address in the age of agentic artificial intelligence (AI), noted Gerew. Each orchestration engine will eventually expose a Model Context Protocol (MCP) server to AI agents, but over time there will be thousands of AI agents in enterprise IT environments that, at any given time, might be spinning up additional subagents, she added.

The issue is that each of these AI agents is likely, at some point, to add yet another orchestration framework to an already diverse mix of tools and platforms that IT teams already find difficult to holistically manage, said Gerew.

Each IT team will need to determine to what degree it might make sense to rationalize existing orchestration platforms versus finding some way to unify visibility across them. Regardless of approach, the one thing that is certain is IT environments themselves continue to become more complex. Hopefully, IT teams will be able to make use of AI agents to manage the next era of IT at scale, but it’s more than likely deployments of AI agents will occur faster than internal IT teams will initially be able to consistently manage.

Eventually, however, responsibility for managing AI applications and platforms will rest with IT teams. In the short term, there will be any number of experimental projects and initiatives that may be led by a particular business unit, but the daily responsibilities of managing those applications will inevitably get shifted back to a centralized IT team to ensure operational efficiency.

In the meantime, IT teams should focus more time and effort on streamlining existing IT environments in anticipation of having to manage more applications than ever running at unprecedented scale. After all, the one thing that has become all too apparent in the age of AI is how quickly agents expose weaknesses and flaws in IT operations that humans may have either been unaware of or, alternatively, simply hoped no one other than the internal IT team would ever need to have known about.