“Thank you for calling your bank this morning, please tell me in a few words why you have called.”
I outlined my query in a few words.
“Please tell me in a few words why you have called. “
This went on until I was finally given the option to speak with a live person.
The computerised answering system could only recognise specific text associated with simple requests like balance enquiries, ordering a new card or applying for a credit facility.
Conversational AI that understands human speech with all its inflections, dialects, accents etc is one of the hard AI challenges. Automated conversations are a goal on many customer experience leaders’ agendas, however there is no quick fix.
Most systems are based on Natural Language Processing (NLP) which provides a statistical output; this is useful for classifying a body of text, so can identify if it contains key words or phrases and provide useful outputs for sentiment analysis and question answering. These algorithms may simulate a level of human understanding but without the ability to discern context, their application to actual customer engagements is limited to menu driven conversations.
McDonalds recently announced the acquisition of Apprente, a voice AI company; they plan to deploy the technology to enable faster and more accurate order taking at the Drive Thru counters. Technology is particularly suited to this environment as the conversation is entirely menu driven and the Apprente technology offers multi-lingual support and filters out background noise.
McDonalds is a good example of using Conversational AI in a setting where it is likely to perform well; these are typically handling routine queries or other functions that are repetitive.
The major technology companies are continuously developing their virtual assistants using the great volumes of data they extract every day from all our interactions with them.
Consequently, the breadth of application of these virtual assistants continues to increase. However, without empathy, critical thinking and creativity they will be unable to replace people for more complex interactions.
NLP based tools can augment these more complex human interactions by providing real-time insight into how the customer is feeling about the dialogue with the customer agent, whether the agent is acting in a compliant and ethical manner etc.
I think it is important to have a holistic strategy for the use of conversational AI; this will inevitably be bound by the capabilities and limitations of NLP. I advise clients to consider how technology can enhance each aspect of the customer conversation, be it the actual engagement, the content or the tone and develop the strategy to make best use of the technology capabilities available.
For practical advice on integrating AI into your customer engagement strategy please get in touch.