Chat is more popular (and more complicated) than ever before. Here’s our advice on chatbots in contact centres and meeting customer expectations using the three main chat channels – live, automated and asynchronous.
Chat is a no-brainer for most customer-obsessed businesses.
It helps you achieve two huge contact centre goals – improve the customer experience (through instant, convenient, two-way interactions) and reduce agent workload (by automating part of the customer journey).
But whilst the opportunity is huge, seizing it can be complicated.
Not that long ago, live chat would have satisfied most customer demands. But today expectations are soaring, thanks to the rise of Whatsapp, Facebook Messenger and Drift. The job is no longer enabling some kind of chat service, it’s refining and optimising the chat experience across multiple channels – live, automated and asynchronous.
This post looks at what customers expect from chat today and how you can meet these demands.
Three goals, three challenges
Every great customer experience is convenient, personal and human. This is true regardless of channel and applies to every type of chat.
Convenient means customers can contact you on channels that suit them, when it suits them.
Personal means that the chat conversation builds on what you know about the customer. So if they’ve shared their phone number with you in the past, you don’t ask for it again.
Human means connecting with customers through conversations. Messages are natural, empathetic and helpful.
So far, so obvious. But achieving these goals can be far from simple.
Convenience is a slippery concept. It’s subjective and often defined in the moment.
That means a channel can be convenient in one instance and not in another.
Chatbots are good for round the clock enquiries but struggle to deal with complex questions. Live chat is good for urgent questions that AI can’t resolve, but it isn’t designed for protracted conversations. Asynchronous chat meets this need, but it can’t be applied to every interaction – it asks too much of the agent.
The only solution is to offer all of these channels.
That way, when a customer wants to ask a quick question about bank opening times, they can get the information they need through a chatbot.
And when they want to discuss a new credit account they can speak to an agent through live chat.
If they want to ask a more complicated question about their ongoing mortgage application, they can reach an agent through asynchronous chat.
The challenge with personalisation is delivering a similar standard of experience across all chat channels. You don’t really want live chat experience to be dramatically more personal than your IM experience, or vice versa.
The key to getting this right is to create a single source of truth about customer interactions – an ongoing record of the data they’ve shared, the questions they’ve asked and your responses. That way agents and bots can draw from the same well when responding to queries.
This can only be achieved by integrating chat channels with all other contact centre channels – voice, SMS, email etc.
The other thing to consider here is data from outside the contact centre. It’s tempting to integrate your chat solutions with all sources of customer data but this is likely to be impractical in the short term. Instead focus on knowledge bases that are linked to specific use cases.
So if you’re using a chatbot for answering FAQ’s, it makes sense to integrate this with your web analytics to determine which users ask which questions. You can use this information to inform the chatbot’s response and where to guide the user to after the question has been answered.
Delivering the human touch is the hardest challenge of all and demands a deep understanding of the user experience as much as it calls for cutting-edge technology.
The goal here is to understand what your customer wants to achieve and apply this to everything – internal processes, channel choice and message content.
Often businesses do the opposite and try and apply one type of chat to any and all use cases. Even worse, they ignore the customer’s perspective entirely and apply chat for their own ends. For example, it’s common for chatbots to steer users to the sales department regardless of their enquiry. This is irritating for customers and means you miss out on a golden opportunity to demonstrate a bit of empathy.
It makes a lot more sense to accept the limitations of certain chat channels and build workflows accordingly. So instead of a chatbot automatically driving someone to the sales department, it picks up customer data and uses this to qualify prospects. Then, once you know you’re dealing with a warm lead, you can get an agent involved via live chat.
Imagine a user wants to upgrade their cellphone contract. They visit the provider’s website to see if they’re eligible. A chatbot can quickly determine if they are by asking when the contract started. If they are, it can direct them to an agent, or upsell them on other deals if they’re not.
Location data collected by the website can be used by the chatbot to determine the lead’s suitability and determine the conversation. An online store with limited delivery radius, for example, will be able to tell from the user’s location whether the chatbot needs to connect the user with an agent or get them to sign up for future email notifications.
Less friction, more flow
The promise of chat is still huge – it can help you eliminate call queues, liberate agents and reduce costs. Even better, it can help you deliver a more personal, human and convenient customer experience.
But chat is getting more complicated. There are more channels and modes to consider and customer expectations are always shifting.
The only way to handle this complexity is to become more open and flexible. You need to experiment with new channels, integrate relentlessly and eliminate data silos.
Then you can sit back and watch conversations flow.
Liberty Connect is an omnichannel messaging solution that connects to almost any messaging platform that matches your customers’ needs. The asynchronous chat works alongside chat bots you can build yourself, single thread ticketing for improved agent performance and web assistant so customers feel in touch at every step.