August 22, 2018
AI, Chatbots, and the Future of Customer Service Assistants
Have you ever interacted with a ‘chatbot’?
What was that experience like?
Chatbots have been around for some time now and to date, the experience they have offered has been less than optimal. Following the lines of ‘monkey see, monkey do’, they have been fairly primitive, spitting out whichever answer they are programmed to based on a predefined rule that allows them to recognise what you’re asking. But if they don’t recognise your query, well things can go south pretty quickly! This leads to customers becoming frustrated and that resentment is often directed towards the brand, not the tech!
But! Times are changing and there are new developments in chatbots and Artificial Intelligence (AI) happening almost every week!
While this technology has undoubtedly been around in some capacity for years now, the advancements in capabilities and functionalities means enormous opportunities are being created to shake up customer service delivery. Bots are smart enough now that they can deliver more human-like experiences incorporating natural language. Customers can get the answers they want and need, and customer service teams can spend their time on more meaningful tasks (rather than answering basic questions over and over again).
What do I mean? Well, let’s have a look at some of the ways that a ‘bot’ can interact with customers now…
- Fun Fact #1 – Using the right platform, your ‘bot’ can actually use sentiment analysis to identify if a customer is becoming frustrated or angry, and can initiate handover to a human consultant as soon as this is flagged. Forget talking to a bot and getting absolutely nowhere… if the bot can’t help, it won’t be long before it passes the customer on to someone who can. But what about when the office is closed you ask? Well, the bot can be smart enough to know what time of day it is and have a select protocol (i.e. a call back system) for when a human is not immediately available.
- Fun Fact #2 – Smart bots learn. You can have a bot that if it doesn’t get the answer right the first time, it will take note and learn for next time. If a customer asks that question again later on, the bot will have the correct answer sorted this time around!
- Fun Fact #3 – If a customer digresses from the current logic flow, a modern chatbot can handle this seamlessly. For example, if the customer is having a conversation with a bot to book a table for 2 on Wednesday night, then just before the booking is made they say, ‘actually, are you open on Thursdays?’, the bot can respond with, ‘Yes we are. Would you like me to change the booking to Thursday?’.
- Fun Fact #4 – Chatbots can recognise when a customer asks questions which contain multiple trains of thought and can address both in a way that appears natural and seamless. For example, if I was to ask a bot, ‘I want to know the store opening hours and can you please let me know if the monkey socks are in stock?’, then the bot would be able to recognise the two different requests and address both without missing a beat!
- Fun Fact #5 – Chatbots can draw in information and data from multiple data sources. For example, it might draw from your CMS, your CRM, and Google My Business, to be able to present accurate and relevant responses to customer queries.
- Fun Fact #6 – Chatbots can recognise natural language identifiers (i.e. one we were playing with recently recognised that ‘Sahlia’ was a name, now that’s impressive!). They will have preset identifiers but in no way does that limit you! You can teach the bot to identify any entities you want! Have a bunch of obscure product codes that people will be asking for? You can teach them to the bot so it will recognise which type of product customers are asking about.
Are you as excited about this as we are yet?
A quick word to the wise, however… not all bots are created equal and having good tech doesn’t automatically mean you’ll have a good bot!
Some Key Things to Consider Before Building a Bot!
There are a lot of tech options out there, many with similar capabilities, some a cut above the rest… but it depends on what you need the bot to do!
We see time and time again that a tech solution is chosen first, and then the business tries to make that fit its needs and the needs of its customers. This almost never works!
Before you even look at the tech you’ll use to build your bot, you need to consider the use case:
- Who is the bot for/who will be interacting with it?
- What will they want and expect from the bot?
- What value will this bot offer our audience?
- Can internal processes be optimised or streamlined using the bot?
- What functionalities and capabilities are needed to deliver an optimal customer experience?
Our number one tip for delivering an outstanding customer experience using AI? Plan first! Ask your stakeholders, ask your customers, and take the time to plan out the strategy behind the bot. After all, artificial intelligence is only as intelligent as you make it.
One Final Note
I have referred to chatbots a lot in this post. However, I actually am not a big fan of the term. Why? Because it hints at the primitive experiences that we have all encountered and doesn’t really embody all the possibilities that are now open to us through recent and ongoing advancements.
So what term do we use then, you ask? Well, we like to refer to them as Customer Service Assistants. When done properly, it really is assisting – specifically, a) your customers as they progress along the customer journey – and b) your staff, freeing them to work on more meaningful work.