How to take a strategic approach with agentic AI for CX

TechTarget and Informa Tech’s Digital Business Combine.

Together, we power an unparalleled network of 220+ online properties covering 10,000+ granular topics, serving an audience of 50+ million professionals with original, objective content from trusted sources. We help you gain critical insights and make more informed decisions across your business priorities.

ICMI Logo
 
Advertisement

How to take a strategic approach with agentic AI for CX

By Jon Arnold | Date Published: July 07, 2026 - Last Updated July 07, 2026 | Comments

AI continues to evolve faster than anyone can follow, and in the contact center space, no trend is more important right now than agentic AI. This is the first of a two-part series, looking at two core use cases, first internal, then customer-facing. With the overall focus of my ICMI articles being on business strategy for CX leaders, each requires a different approach, and I’ll begin here with the former.

What is Agentic AI?

Before doing that, with AI being complex for everyone, a basic definition is in order. The concept of agentic AI builds on previous AI advancements, especially conversational AI and ChatGPT. These iterations made AI more accessible for going beyond search queries and low-level task automation. Agentic AI is next-level automation, where tasks can be completed autonomously, without human intervention.

The breakthrough comes from these AI agents applying human-like reasoning and judgment to make decisions for taking a task to completion. By enabling AI agents to have this degree of “agency” to manage a task end-to-end, this opens up new possibilities for all types of workers to use automation to boost productivity.

Of course, this requires AI agents to perform at a certain level of proficiency, and for human workers to trust agentic AI – topics to be addressed another time – but that aside, CX leaders need to consider the potential benefits at face value.

Core Internal Use Cases

The starting point would be internal use cases, where many CX workflows are disjointed, with limited degrees of automation. This isn’t surprising since legacy technology remains widely used in many contact centers. Given that C-level management tends to view AI for cost reduction, this use case for agentic AI will be clear. Also, since AI is not an out-of-the-box technology, there is less risk fine-tuning agentic AI internally before developing customer-facing applications.

  • Call routing – agentic AI can make more precise decisions about routing incoming calls to the best-fit agent available in real time

  • Escalations – can determine at what point during a call that an escalation is needed, and then to initiate moving the call to the next-level human agent

  • Account updates – automating processes for updates as they occur during a call – address changes, renewals, payment methods, refund requests, etc.

  • Scheduling – can optimize overall operations by dynamically managing scheduling requests from agents to match anticipated call volumes

  • Call summary – going beyond automated call summaries, agentic AI can operationalize action items coming out of those calls – updating supervisors, training to address skills gaps, additions to knowledge base or CRM, etc.

 

These are just a few examples where agentic AI can provide new and/or improved levels of automation to tasks and workflows for internal operations. The intent here is to minimize manual work that takes human agents away from engaging with customers. 

Aside from freeing up agents, agentic AI can perform these tasks faster, more completely and more accurately – and that’s where the cost savings can materialize. With agentic AI being so new, these outcomes are far from guaranteed, but the potential is clearly there, and CX leaders need to recognize that the effectiveness of AI agents will improve over time.

Let’s be strategic

Adopting new technology always comes with challenges, but especially for agentic AI, which can have a transformational impact on automation to improve operations. The same holds for the next article in this series, which will focus on customer-facing use cases.

Being the ideal starting point for agentic AI, internal use cases need to be carefully managed, as this will likely determine the viability for future applications. This calls for a strategic approach, where the value of agentic AI needs to be tied to specific outcomes that have a tangible impact on the business.

To that end, here are three guiding principals to support your agentic AI strategy for CX. Setting this in place now will serve you well, as it applies as much to internal use cases here, as for customer-facing use cases, which will be addressed in my next article.

Start small and scale. AI may seem like a silver bullet to address many CX shortcomings, but it’s still evolving, and with agentic AI in particular, a high level of trust will be needed before it can be entrusted to make decisions autonomously. Risk mitigation is critical with this being so new, and small-scale pilots with basic workflows is the way to begin, then build on your successes.

Focus on data quality. AI is inherently data-centric, and the CX outputs will only be as good as the inputs. Initial agentic AI deployments should be with workflows where the data inputs are current and reliable. Data quality can be highly variable, especially in legacy-based contact centers, and agentic AI should serve as a catalyst to uplevel data quality across the operation. This reflects the fact that the root cause for poor results is more likely due more to lax practices around data management than the AI applications themselves.

Measure everything. This is an extension of the above point, but is an important principal for best practices to get the most from agentic AI. Sustainable support for AI initiatives will be driven by ROI, and metrics will be needed to show that. For this reason, initial pilots should focus on internal use cases that can be measured. The more datasets the better, especially for agentic AI, which must rely on metrics for autonomous decision-making.

headshot
Jon Arnold is Principal of J Arnold & Associates, an independent analyst practice providing thought leadership and go-to-market counsel with a focus on the business-level impact of digital transformation on the future of work. Core areas of expertise include unified communications, collaboration, cloud platforms (UCaaS, CCaaS and CPaaS), artificial intelligence, speech technologies, 5G, Internet of Things, contact centers, and customer experience.

 

Topics: AI, Customer Experience

Recent Articles

Subscribe now to ICMI Contact Center Insider!