Smart Sourcing for the Customer Journey

Most web chat (or instant messaging / IM) sessions act by offering a live assistance option to the process of web browsing. Like email, it has been around for many years, but only very recently has started to grow volumes to the extent where it has become a mainstream channel for customer-business interactions. With the advert of AI / virtual agent technology, web chat is primed for strong growth.

Web chat has often been used as a ‘point of crisis’ channel, for example, to convert an online shopping basket into a sale by providing timely service, or if a browser is paused on a webpage too long, perhaps as they can’t find what they are looking for. In such cases, there are two main benefits to the business in providing web chat: revenue maximisation, and the avoidance of unnecessary calls.

Web chat can also act as a safety net for the customer if an online self-service attempt fails. An analogy can be made with voice self-service, where a failed session is often ended with the customer ‘zeroing-out’ - pressing zero to get in touch with an agent. Failed web self-service sessions may end with a phone call being made, but web chat can avoid a number of these, which is a cost saving for the business, and better for the customer as well.

The analyst firm ContactBabel reports that web chat is experiencing strong growth in its availability in the UK, although volumes on average are still only around 4-5% of all customer/business interactions. There is no reason why the user uptake of web chat will not continue: it works well for customers as providing an immediate response, and with multiple concurrent chat sessions per agent, it can be a lower cost channel than voice for the business to support, although cost differential between phone and web chat are not dramatically different, as so much of the web chat work carried out is still non-automated. Solution providers report that web chat is currently being trialled by numerous businesses, often at a limited, or departmental level so they can assess the suitability of the channel for a company-wide rollout, and understand what needs to be done to ensure full implementation is a success.

Chatbots and virtual agents

One form of value-added web chat functionality is the Virtual Agent, which may appear to a browsing website visitor to be a human agent, offering web chat. However, it is an automated piece of software which attempts to answer the customer’s request, including sending relevant links, directing them to the correct part of the website or accessing the correct part of the knowledge base. If the virtual agent cannot answer the request successfully, it may then seamlessly route the interaction to a live web chat agent who will take over.

Most virtual agents encourage the visitor to engage with them using natural language, rather than keywords. The virtual agent will parse, analyse and search for the answer which is deemed to be most suitable, returning this to the customer instantly. Many virtual agent applications will allow customers to give all sorts of information in any order, and either work with what it has been given, or ask the user for more detail about what they actually meant.

The virtual agent application is different from standard search functionality, ignoring bad punctuation or grammar, and using longer phrases rather than just searching on keywords. Sophisticated AI-enabled applications attempt to look for the actual intent behind the customer’s question, trying to deliver a single correct answer (or at least a relatively small number of possible answers), rather than a list of dozens of potential answers contained in documents which may happen to contain some of the keywords that the customer has used.

Through ‘listening’ to what the customers actually say - perhaps through a mixture of large quantities of audio and text – the initial set-up configuration can achieve a good accuracy rate, which really benefits over time as a positive feedback loop is established. Solutions that gather and differentiate customer requests and results from multiple channels, noting the difference between them, have an even better success rate.

When the virtual agent application has low confidence that it has returned the correct result, it is able to escalate the customers query seamlessly to a live chat agent, who then has access to the self-service session history, enabling a greater chance of a successful resolution without repetition. The eventual correct response can be fed back to the automated virtual agent and the knowledge base underlying it, which will make it more likely that future similar requests can be handled successfully through automated agents.

Web chat cost

The mean average cost of a web chat is £4.24, around the same as a phone call (£4.27) and a little higher than an email (£3.87), but we would expect to see a greater differential from a channel that can be at least partially-automated, and which offers the opportunity for multiple concurrent sessions.

How are web chats handled?

40% of respondents using web chat offer the option immediately to all website visitors, with 60% only doing so at some specifically-triggered point in the interaction.

Of these 60%, the most frequently used trigger for web chat was when a visitor went to a specific page, with other triggers being when a customer was on a page for a certain amount of time, and at the point of sale, although these latter options are much less frequent.

One of web chat’s traditional strengths is seen as the ability to have agents handle multiple chats concurrently (of course, it only seems this way to a customer, as the web chat agent uses the time that the customer is typing their response to handle other chats). Some vendors have stated in the past that agents could run five or six concurrent chat sessions: the reality seems to be that two sessions is a reasonably consistent average, with a peak of three or even four if required, but which is not possible on a long-term basis.

Respondents indicate that the typical wait for a web chat session is usually somewhat less than that of a phone call. 24% of contact centres have an average wait time for web chat of lower than 10 seconds, with a further 28% stating that the average wait time is 10-20 seconds. Maintaining this level of accessibility for customers will reinforce their positive experiences of web chat, and will encourage customers to keep using the channel, not only when contacting a specific business, but also in general.

Improving web chat

As the cost of web chat is similar to other channels such as email, voice and social media, there is considerable room for increasing efficiencies and lowering costs.

Whereas only 5% of web chats had any automation involved in 2015, this grew to 19% in 2018, mainly as a result of initial handling by automated chat bots which may then hand off to live agents where appropriate.

Further comparing the experience of web chats with telephone calls, 51% of web chats take longer than 3 minutes to complete fully, as agent multi-tasking and the time taken to type differs from the experience of handling a phone call.