Enterprises are engineering autonomous service desks using agentic AI to scale IT operations. This approach involves deploying autonomous agents capable of understanding operational intent, sensing real-time conditions, and executing coordinated actions across multiple domains. The goal is to move beyond predefined automation workflows toward more adaptive, self-optimizing systems that can handle complex, dynamic environments.

The adoption of agentic AI in enterprise IT is still in its early stages. Most implementations are currently at Level 2–3 autonomy, as defined by TM Forum’s autonomous networks taxonomy. At these levels, automation is limited to executing predefined solutions within specific network domains. True autonomy—reaching Levels 4–5—requires agents that can research, plan, and adapt to new situations without human intervention.

The transition to higher levels of autonomy depends on the development of a shared platform that integrates telecom-domain models, policy controls, and digital twins. This infrastructure allows agents to discover and validate new operational methods, rather than simply executing existing ones. The complexity of building such a platform is a major barrier to achieving full autonomy in enterprise IT.

The consequences of adopting agentic AI in enterprise IT include significant shifts in operational cost structures, increased dependency on vendor-specific platforms, and new governance challenges. Organizations must navigate the trade-offs between automation efficiency and the complexity of managing autonomous systems. Market reactions have been mixed, with some early adopters reporting improved efficiency while others face integration and scalability hurdles.

As the technology matures, enterprises will need to balance the benefits of autonomy with the risks of vendor lock-in and governance complexity. The development of standardized platforms and open frameworks will be critical to enabling broader adoption. This shift is expected to reshape how IT operations are managed, with agentic AI playing an increasingly central role in enterprise infrastructure.