Conversational AI: use cases and benefits

featured image

Imagine if 80% of all your customer contacts could be handled automatically by machines. Human agents would then be able to give all their attention to the remaining 20% that either have complex issues or create genuine value for the business and customer.

The limitations to automating the routine queries and transactions that take up the bulk of the contact centre’s time have been: 1) AI’s ability to understand customer intentions, 2) your operational ability to design and build intelligent systems, and 3) users’ ability to interact with automated systems.

Conversational AI platforms have come a long way in the last few years and have now overcome all of these issues. Unlike stand-alone chatbots which at best manage part of an interaction, Conversational AI platforms are now handling entire customer interactions and processes end-to-end with little to no human intervention.

How it works

Conversational AI, as the name suggests, is a technology that allows customers and staff to query a system for information by posing questions in natural language, whether written or spoken.

The technology is powered by recent advances in AI technologies such as Natural Language Processing, speech-to-text, and text-to-speech. It is further enriched by big data capabilities like sentiment and intent analysis that are enabled by powerful cloud computing and storage.

The result is that it is possible to develop virtual agents that can understand natural language, interpret intentions, and respond in much the same way a human would – only faster, more accurately, and for a much lower cost per interaction, particularly as they can handle almost unlimited numbers of interactions simultaneously.

The reason this is a game changer for automation is that it allows people to interact with AI systems through the most natural user interface imaginable – speech (either spoken or written). With a website, app, or other type of self-service software there is always an interface that users have to learn. Naturally many of them abandon this in favour of calling the contact centre. This may be one reason why 50% of companies say they will spend more on enterprise chatbots solutions than mobile apps in 2021 (Gartner).

Why its time has come

The technologies behind Conversational AI are now proven and reliable. They are able to handle a wide range of pre-determined interactions very well, and actually automate many complete processes end-to-end allowing customers to get things done quickly.

Secondly, consumers are getting used to interacting with technologies through conversational interfaces – whether these are chatbots on a company’s website or the Google Home and Alexa machines in their homes, as well as Messenger on their phones.

There has also been a fundamental shift in the way companies and customers are approaching customer service. Customers want more choice, instant fixes for their problems, and quick interactions. They have become frustrated during the pandemic with the service offered as companies have had difficulties coping with the increased volumes. Waiting five minutes to talk to a human agent for a problem that is perceived as quick and easy to fix only adds to the frustration.

For their part, businesses have been under huge pressure to reduce costs but not at the expense of service levels, as customer experience is now an important differentiator. 64% of businesses believe that chatbots allow them to provide a more personalised service experience for customers (Statista).

Finally, the modern Conversational AI platforms now available include sophisticated tools that allow operational managers to build and configure bots themselves without having to code or lean on the IT department.

For all these reasons, many organisations are turning to customer service automation using AI technology.

Gartner expects 20% of all customer service interactions to be handled by AI by 2022.

Uses for virtual agents

A bot powered by Conversational AI can be deployed on any channel. Whether it’s interacting via speech or text, and whether the user is on IVR, webchat, messaging, social media or an internal self-help system or knowledge base, the underlying conversational engine remains the same.

Once integrated via APIs with the company’s CRM system and transactional databases, the AI is able to look up customer information and respond appropriately. It can also use that data for insight to predict likely customer behaviour even in the middle of an interaction.

These types of systems aren’t just customer-facing but can be used internally as well to provide agents with a natural way to search for information and receive recommended courses of action that help them solve problems more swiftly and personalise service and offers.

It needs to be enterprise-grade

Until recently most organisations have dabbled in deploying chatbots that operate in siloes. These often have limited functionality and are capable of responding to just a handful of customer or agent queries.

What’s needed instead is a reliable and robust automation platform, integrated with the organisation’s other business critical systems and channels, to provide an orchestrated response to customers.

Imagine a whole workforce of virtual agents, working together and pooling information to provide each customer with the most appropriate response as quickly as possible. The key difference between an AI platform and isolated bots is that a platform is capable of automating entire processes for customers, including in the back office, whereas standalone bots can only automate single steps.

Ensure the platform you select has built-in ‘low code/no code’ tools for creating end-to-end interactions and integrations with other applications. Without intuitive tools that operational team members can use developing even a basic use case such as “change of address” could take several months, and be difficult to change as and when needed. With a modern platform it should be possible to build such a process in just 1 to 2 weeks, enabling a much faster proof of concept and ROI.

An AI-first approach

To cope with the increased volume of interactions most contact centres have seen in the last year, many are turning to an AI-first approach. In this scenario, customers are initially answered by an AI no matter what channel they use.

Routine and simple queries – reset a password, get an account balance – or even more complex ones that can be systematised – transferring money, applying for a loan – can be entirely handled by the bot, or passed to a specialist bot.

Once it performs its initial ID verification and screening, the bot can hand over to humans only those queries that require it, either because they are too complex or emotive in nature, or need deep investigation. Rather than just dump the customer in a queue, the bot can pass the query to the most appropriate human agent, ideally using whichever channel the customer prefers – or even schedule a call back so as not to waste the customer’s time.

A major benefit of this approach is that all customer interactions get handled immediately, so no more wait times. Even if the query isn’t resolved immediately, the customer has been heard and knows that the company is taking action.

In this way, organisations can also cope with peaks in demand without having to increase headcount for short periods of time drastically. Take away the need for a human to answer every call within 30 seconds, and the contact centre is back in control. Instead of frantic spikes of activity between stretches of not very much, operations can smooth out those peaks and troughs, handle priority queries first and get back to everyone else.

Customers appreciate this approach too as they don’t like being put on hold while agents look into something. And if their query is one that can be handled entirely by the AI, the chances are it will be managed more quickly and accurately than a human could do it.

A win-win for customer and business

On an overall cost-to-serve basis, AI wins out over human-assisted service every day. Bots are tireless workers that don’t make mistakes. Neither do they need breaks, rostering, salaries, or expensive office space. They can also handle hundreds of interactions at once if need be, whereas humans can manage two to three. Conversational AI can therefore play a huge role in helping an organisation meet its cost reduction goals.

Some users report contact deflection rates of up to 80%, with the average being 20% to 30%. Handling times of those interactions still managed by human agents are also reduced by up to 40% when agents also have access to AI systems that assist them during calls. The use of AI to manage routine interactions – balance queries, password resets, and so on – has been shown to cut costs by up to 30%.

As for the quality of service, AIs score highly on most traditional contact centre KPIs such as time to answer, handling time, and first contact resolution. It should also not be forgotten that most of today’s consumers actually want to use digital, asynchronous channels (SMS, chat, messaging) to better manage their time – and AI works best on those channels.

For these reasons, most companies that have deployed automated customer service solutions actually see an uptick in their customer satisfaction, NPS, and customer effort scores, as well as cost savings.

Request a conversational AI demo

Related Content

2021 Guide to Technology for Contact Centres and Customer Experience

Last updated on: April 4, 2023