ISG Software Research Analyst Perspectives | Conversational Intelligence

Conversational automation is one of those software segments that means something different depending on who you are or your role in an organization. According to my colleague Jeff Orr, the core of the idea is that conversational automation tools benefit from artificial intelligence (AI), allowing software agents, chatbots and virtual assistants to automate customer interactions and internal processes. This broad definition hits the mark, I think, because it identifies the core functions without putting too tight a straitjacket on the technology itself, which is developing very quickly. The software provider landscape is analyzed in the ISG Buyers Guide for Conversational Automation. I want to expand a little on Jeff’s analysis and put some customer experience (CX) context around conversational automation (or CA from here on out). The current generation of CA applications is distinguished by understanding natural language, sentiment and intent, generating relevant responses and executing actions based on user input. This would appear to include much of what is currently being sold by CX and contact center providers in AI portfolios, including what’s beginning to pass for agentic AI. (Not to be extra confusing, but the category is sometimes called conversational AI.) So, I think when we talk about the spectrum of technologies in this basket, we should view it less as a strict market segment and more as a basket of tools that—together—manage and automate customer interactions in real time across various channels. These applications interact with customers and human support representatives, providing information resources and guiding users through processes and transactions with as little human input as possible. They are autonomous enough to handle much of the traditional customer volume but are also capable of bringing a human rep into the interaction when it finds a problem too challenging. In the past few years, the contact center industry has reached a milestone never considered possible: the deflection of 50% or more of an enterprise’s interaction volume away from human support. For decades, the rule of thumb was that if you were capturing 30-35% of your volume within an automated system, usually an interactive voice response (IVR) application, you were doing pretty well. Today, completing half of the interactions within automation is not unusual. When we count the interactions that have a significant element of automation but still rely on a human to oversee, finish or approve, the portion goes much higher. ISG Research now asserts that by 2028, automated systems will be able to contain two-thirds of customer interactions within self-service, due to advancements in conversational AI and knowledge retrieval. This is already changing the way the industry approaches some long-standing, intractable problems like labor attrition. There is every reason to believe that we are already on the road to near-total automation at the front end and that the success enterprises are having with this transition is encouraging the next step: automating processes inside organizational CX, including the development of procedures that connect what happens in the contact center to other parts of the customer life cycle pre- and post-purchase. CA doesn’t necessarily require AI, which is a source of confusion and market muddling. Yes, AI is increasingly part of the mix in software products aimed at customer interactions, so it is difficult to disentangle which elements require AI and which are fine using old-school rules-based processes to automate. In the long run, those distinctions won’t matter as much as understanding the role automation plays in delivering successful outcomes for businesses. In CX, particularly in contact centers, you’re trying to automate for two purposes. One is to reduce costs by streamlining the human labor needed to handle increasing interaction volume. The second is to encourage the creation of new processes that support customers in ways that haven’t been possible before. That’s where the real excitement comes from—the idea that, instead of having a customer relationship that starts and stops with each call into a center, you can have an organizational relationship. That means customer success teams using automation to encourage product usage and help with product onboarding. It means automating the creation (and delivery) of offers and messages personalized to the needs and identity of each customer. It means automating the delivery of interaction data to analytics systems that reside outside the contact center, where marketers can orchestrate experiences based on targeted segments and audiences built on contextual cues like sentiment and advocacy. This suggests to me that, when we talk about conversational automation or AI, we’re taking too narrow an approach. While the “conversation” is the first and perhaps easiest part of the process to be automated, it’s just one prong of a whole-enterprise approach to CX that injects automation wherever it might make sense. Based on the features applied to CX and CC products over the past two years, I think conversational automation may be a term that’s unnecessarily limited to the communications element of CX. The same tools and technologies that automate conversations also provide an on-ramp to automating more complex, cross-departmental processes. It also makes it easier to see where a business has opportunities to influence and nudge customer behavior in positive directions. I completely agree with Jeff’s assertion that conversational automation will progressively integrate with process orchestration and management platforms in the coming years. CA should be viewed as part of a long-term project that gradually brings into the automation fold many teams that may not touch (or converse with) customers but whose efforts and analyses impact CX. With this in mind, I would urge enterprises to look beyond the “conversational” label and think about CX automation more broadly, not just as a way to save money on the front end of the interaction. When incorporating more automation into interaction handling, look at the software provider offerings to determine whether the conception of automation starts and stops with reducing human call center labor or whether automation extends to the back office, marketing analysis, customer success and even product development. Regards, Keith Dawson

Analyst Perspectives

Conversational Automation’s Impact on AI and CX

The Consequences of AI-Induced Head Count Reduction for Contact Centers

Our Analyst Perspective Policy

Our Analysts

David Menninger

Jeff Orr

Keith Dawson

Mark Smith

Matt Aslett

Matthew Brown

Robert Kugel

Stephen Hurrell

ISG Software Research

Subscribe to Email Updates

Posts by Month

Posts by Topic

Analyst Perspectives Archive