Summary Bullets:

• Agentic orchestration has been dominating vendor narratives as growing complexity in the CX space is driving the need for a sophisticated coordination system.

• CX agentic orchestration remains fragmented as there are still infrastructural gaps, and trust remains a major issue for customers.

Agentic orchestration has been increasingly occupying CX vendor narratives in recent years, so it is time to take a deep dive into what orchestration is and why it is becoming relevant. Simply put, orchestration is bringing people, workflows, and technology together to achieve an outcome. Depending on how far we stretch the concept, orchestration is not new, and even a deterministic call routing can be described as orchestration. The question is, what is new today and why is orchestration increasingly dominating vendor narratives?

There are two factors bringing orchestration to the forefront: connected customer experience and agentic AI. Connected customer experience came into play with the advent of cloud platforms, making it possible to consolidate the CX IT stack and create a unified system. This enabled data to flow seamlessly between applications and systems for contextual insights before, during, and after customer engagement, reducing fragmented customer experiences. The second factor is the rise of agentic AI, which, as the name implies, involves agents that can autonomously perform tasks on behalf of humans. This autonomy stems from the ability to understand, reason, decide, and act. The projection is that agentic AI will proliferate, acting as digital assistants for workers and customers. While interactions today are still between agentic AI and humans, it is predicted that in the future, they will also take place directly between agentic AI.

What we are witnessing is the rise of a complex CX landscape with numerous actors, including AI, people, and applications, making it necessary to have a sophisticated coordination system. This is where orchestration becomes relevant and even critical. What makes orchestration compelling today is its agentic capabilities with a cognitive layer in the backend. It is probabilistic in nature and can understand, reason, decide, and act on its own, unlike the pre-programmed, deterministic models used for reaching a goal. Agentic orchestration is designed to autonomously map the most optimal customer journey depending on customer intent and stage, as well as determine the most suitable resources, insights, and tools to achieve a given goal. The challenge, however, is that the term is broadly used, and the underlying question is: How agentic are the orchestration layers today?

Agentic orchestration as it stands falls under two classifications: routing orchestration and workflow orchestration. Routing orchestration is when AI autonomously routes interactions to the most suitable agent, while workflow orchestration is when AI determines the optimal customer journey to accomplish an outcome. We see some instances where routing orchestration is deployed and some instances where workflow orchestration is deployed, but in most cases, orchestration remains fragmented across bits and pieces of specific, outcome-driven workflows.

In a true agentic era, orchestration will take place end-to-end across all use cases, but there are infrastructural gaps before such a state could be achieved: LLMs are still reasoning and decision tools and cannot really act although tool calling are used for driving actions; data requires more work, including quality improvement, aggregation, and the development of data graphs; and latency/real-time technology needs to be faster. The most critical factor, however, is trust. Today, AI autonomy and trust work in reverse order. It is a journey vendors are taking as they gradually fill the gaps, but there is still some distance to go before we see a pure state of agentic orchestration.