While the vast majority of contact centres are equipped with an IVR, these systems generate dissatisfaction among more than 80% of callers. Considered too complex, too rigid or too directive, DTMF IVRs persist despite everything, for lack of a viable alternative. However, recent advances in Artificial Intelligence now allow us to consider an innovative and more efficient solution: IVR in natural language.
In 2019, IVRs still generates dissatisfaction among 82% of callers
According to the “State of the IVRs” report published at the end of 2018, nearly three-quarters of contact centre interactions pass through an IVR (Interactive Voice Server). One would think that such a widespread solution is necessarily the best. And yet…. IVRs creates dissatisfaction in 82% of callers!
The statistics are clear: on average, 5% of callers hang up immediately (average dropout rate). But most importantly, nearly 50% of them bypass the IVS, in order to talk to an agent as quickly as possible, which leads to an explosion in the internal transfer rate. As a result, the cost of processing these calls explodes (x12). This reinforces the (alas frequent) perception of the contact centre as a cost centre. Thus, although attractive due to their low implementation cost, traditional DTMF IVRs suffer from many hidden costs.
Natural language is emerging as an alternative to traditional IVR
Why do IVS generate dissatisfaction for callers? The system suffers from four main defects:
- Too complex and opaque: the caller does not always know how to choose the option that corresponds to him among the choices offered, and does not know the different levels of the IVS
- Too rigid: if no options match the request, the caller does not know where to go to talk to an counsellor
- Too long: IVS increases the duration of calls and degrades the experience because it forces the caller to listen to many options that do not concern him/her
- Too directive: until the caller has answered all the questions on the IVS, he or she cannot even talk to someone who can solve the problem
Faced with these irritants, offering the caller a natural language discussion is an innovative solution to improve their satisfaction, while reducing routing errors within the contact centre.
While speech-based IVRs, operating by keywords, have been around for a few years, the technology was too immature to ensure exhibitor satisfaction: often the tool did not understand the caller’s words or did not capture the subject of the request. Artificial intelligence has changed the game. Now, the caller can speak with a conversational AI using words and sentences from his or her daily life: it is the natural language IVR.
Four factors explaining the development of IVS in natural language
Because it’s much more comfortable for the caller
No need to press keys, no need to listen and remember the sometimes numerous choices while analyzing whether they correspond to our request: natural language allows the caller to speak to the conversational agent as he would to a human.
Thanks to an average 30% faster treatment, the experience is much more fluid and pleasant, and increases satisfaction while reducing the dropout rate.
Because in 2019, talking to robots is part of our lives
The era of smartphones and connected speakers has brought robots into our lives. Their names are Siri, Alexa or Google Home, and more and more of us talk to them every day. Artificial intelligence provides us all with a personal assistant, and humans now trust a machine as much (or even more) to deal with their problems.
According to a Pindrop study, 85% of companies plan to use conversational technologies in 2019. The market is changing very quickly, and it is an opportunity for each of them to innovate in their customer relations and modernize their image.
Natural language makes it possible to humanize the customer/company relationship, and to increase the personalization granted to the customer.
Because data is an opportunity to derive the IVR
Conversational IVRs make it possible to collect a lot of data: recordings, transcribing conversations, detecting intentions… The computing power of artificial intelligence then makes it possible to identify new and extremely broad applications. For example, intention mapping can identify the caller’s needs, and identify redundant or missing branches at the current IVR level. Thus, a more detailed understanding of the caller’s need allows us to direct them to the best resource capable of processing their request.
In the longer term, artificial intelligence therefore makes it possible to rationalise and optimise the processing of incoming calls in the call centre, in particular by eliminating organisational inefficiencies.
Because each request can be qualified even before the conversation begins
While IVS only provides information on the person’s journey before they reach the counsellor, IVS in natural language uses the first few seconds of the conversation to ask the caller for the reason for the call.
This allows the agent, even before starting the communication, to be able to identify the customer or to benefit from a verbatim allowing him to locate his request.
In conclusion, the democratization of conversational technologies and the reduction of their cost seems to promise good days to IVS in natural language!