Conversational AI can be defined as technology that enables humans to interact with applications in the same way they would engage with another human. This mainly involves the capacity to use natural language in a written or spoken conversation with an application.
The growth of conversational AI is accelerating—with Gartner predicting it to reduce labour costs by $80 billion in 2026, and with spending on conversational AI reaching $2 billion in 2022.
The early days of conversational AI
Early enterprise deployments of conversational AI were in the form of “chatbots”. These bots were designed to answer simple questions and very basic requests. They were mostly deployed within websites and applications, with mostly written interactions. Chatbots also facilitated limited voice interactions that required the user to ask a question in a specific way.
While these early chatbots served a purpose, they were restricted in their capacity to understand beyond simple dialogue or specific means of asking questions. They often provided constrained answers, which resulted in varied customer experiences—the simplest cases were resolved, while most interactions still required human involvement.
The new generation of conversational AI
Today’s bots can engage in a more open dialogue, allowing users to have “normal” conversations across all digital channels. Callers can now use their voice and engage in more natural conversations across comparatively complex cases, which is driving increasing automation levels and delivering contact centre productivity savings.
Conversational AI has progressed from being able to answer fairly simple FAQs to handling a broader range of general queries, and is now evolving into areas where digital assistants are “specialists” in particular areas such as Insurance, Banking, or Technical Support. These digital assistants have been built with a specific use case in mind, for example, an insurance claims digital assistant. The AI is pre-programmed with all the processes, terminology, and other elements required for an insurance claims agent and comes ready to deploy. In this example, call or workflows would be pre-designed to represent the flow of an insurance claim process, with the typical intents prebuilt, such as "lodging a claim", "understanding the claim process" and "where is my claim". They will also usually include the kind of utterances, or words spoken during a claims process. The benefit is the capacity for the digital assistance to be deployed faster, at a lower cost, and delivering superior service. In many ways it’s like employing an experienced worker versus someone new to insurance claims.
However, enterprise clients are starting to struggle with multiple, disparate chatbot solutions from various vendors deployed across their business. The move is towards a more holistic and unified approach across voice and digital channels, with conversational AI applications and services deployed and composed from a single platform across the various areas of the business. This would include internal applications such as HR, support desk, policy, and education, in addition to customer-facing sales and support services. The capacity for a single platform that delivers a consistent and managed experience across every channel is compelling, especially in enterprise and government.
Ensuring success with conversational AI
A primary factor in the increasing success of conversational AI is the ability to undertake complex calls and handle “intents” (the reason for the customer’s call, e.g., to obtain their bank balance) and “context switching” (where there’s a change in the conversation, e.g., the customer calls to renew their car insurance and then switches to questions about home insurance).
Critical factors that will determine a successful deployment of conversational AI include conversational design, integration to CRMs and data sources for personalisation, and understanding of the utterances or way questions may be asked by users.
Is conversational AI for you?
The business case for conversational AI remains to simplify engagement and reduce the cost to serve, while increasing customer and employee experience. Primary drivers of the business case are the challenge of recruiting and retaining people across the contact centre, the need to be accessible 24X7 across every channel, and the capacity to personalise at scale.
Within the contact centre environment, the capacity for calls to be automated using conversational AI not only drives significant cost savings, but delivers a no-wait experience for callers, while freeing up contact centre agents for higher value tasks or to engage with those who cannot self-service with automation.
Across digital channels, the benefits lie in being able to resolve complex requests from customers, resulting in a positive customer experience and ensuring and encouraging an ongoing digital relationship through the channel of the customer’s choice, while delivering cost savings to the organisation.
Conversational AI has come a long way, with advances in the technology making it difficult for many to determine whether they are engaging with a human or a machine. The emerging debate is one of ethics in AI, and this includes ensuring customers understand that they are not engaging with a human. This is indicative of the advanced nature of engagements capable and the very real potential of the platform to drive the $80 billion of cost savings by 2026, as forecasted by Gartner.