Crawl, Walk, Run … and focused on the Customer

Over the past 12 months we have delivered numerous engagements focused on extracting actionable insights based on what respective companies’ customers are actually saying during interactions. For these companies, the engagements have primarily focused on improving sales conversion, auditing and improving compliance, customer experience, service effectiveness or self-service optimisation.

However, we often see the Digital team at these same organisations are in parallel experimenting, implementing or operating chat bots and voice bots. The focus generally is around removing calls from the contact centre where the ultimate aim is cost reduction (by removing the people related costs).

Where this gets interesting with chat bots and voice bots, is where does one start from a customer centric perspective?

One of the biggest misconceptions is that you can just take a chat or a voice bot off the shelf, plug it in, let it interact with your customers, and then you can turn off Agent chat or reduce interactions to the Contact Centre (and therefore people costs) at the same time.

Up front, it is crucial to acknowledge that your bot will be new, have to learn, and there is a journey in setting and recalibrating the end customer expectations.
One of the best practices we see is to focus on a specific use case, and then begin to analyse how customers express their “intent” for the sales or service use case in question.

A great place to start, is to get a statistically relevant number of your call recordings analysed. Analysing this unstructured data with the right methods and expertise, will provide rich insights as to true customer intent (far more accurate than agent disposition or wrap up codes) and this then serves as a solid data driven foundation for understanding the way customers express their “intent”.

We often deliver engagements for customers utilising speech and text analytics to identify the true intent of customers which then helps to model improved agent competency models, design improved routing strategies, improve web self service and other improvement initiates. A natural extension has been to utilise this same analysis of customer intent to uncover a set of intents that are unique to a company and their customers, that can in turn be used as key inputs on how to implement a customer focused bot.

Whilst innovation can be crucial for competitive advantage, don’t go so quickly such that your customers take the brunt of your bot learning in the early stages as many will be turned off and not come back. And importantly, don’t just turn off other channels whilst this process is happening as you will not get the benefits you are looking for. You need a phased transition to test, learn and mature.

If there is one call to action I would recommend, it is that you start to build your Intent framework based on what your customers are trying to do with you right now. This will be a great base from which to keep building out your customer engagement channels with chat, chat bots, voice bots … the list goes on.
Would love to hear your thoughts, feedback, and experience on this ever evolving subject matter.