3 Tips for Launching a Brand Chatbot
To chatbot, or not to chatbot?
It’s a big question many brands have wrestled with over the last 18 months. And fortunately for all of us who design and build them, we’ve mostly managed to steer the conversation through “should I” to “where do I start”. There are many contributing factors to this new resolve – a new budget cycle, advances in technology, a few good examples, and a growing realization of the return on investment and experience that make it easier than ever to start.
The other big contributing factor is that as we interact more and more with bots in our daily lives – “Hi Alexa”, “OK Google” – we are training ourselves on their value to our experience as customers. That exposure means we’re more accepting and expecting of bots in our transactional relationships with brands.
So yes, ‘tis the season for bot building – and now I’ll stop with the bad Elizabethan language.
For anyone who is still not sure where to start here are a few key considerations:
Bite off only what you can chew today
As with any new idea or tool, it pays to get a few test swings in before you scale. You’ll want the initial undertaking to be relatively contained to not cause disruption to typical day-to-day operations for your employees and customers.
You only want to “bite off what you can chew” and that means being honest with yourself and your CX, marketing and IT teams about the goals you want to achieve, and having the agility to alter those goals based on what is truly attainable. Think about your brand, your industry and your customer and identify one or two challenges you can solve with a bot interaction.
One of the big mistakes brands make is trying to do too much with their first foray into automation. Conversely, there is a risk of doing too little and launching a bot that adds no value. Think about frequent use cases and solve for those. For example, if you’re a retail brand, product availability is probably a good place to start. If you’re a bank or financial company, can checking account balance is probably a use case that you should think about. If you can perform those simple things well with a bot, that should have a positive impact on the customer experience.
Train your bot to speak the language
No one knows your customers better than you and your employees so when it comes to “onboarding” a bot, you need to invest in training – to make the bot sound natural in conversations and to accurately understand customer queries. Many brands outsource that work to development teams – who don’t know your brand and aren’t necessarily invested in getting the dialog right.
This idea of ‘conversational ai’ is one of the biggest priorities for automation today and also holds true for voice assistants like Siri, Google Assistant, Alexa and Cortana as they struggle to go from Q&A assistants to true conversational assistants..
Ensuring your bot is able to converse naturally and intelligently in your own brand language with your diverse customer base will lead to better and more consistent experiences with your brand.
Pivot, adjust and pivot again
The successful launch of a chatbot is only the end of the beginning. Ensuring its continued improvement means capturing feedback from consumer interactions and turning that insight into a continuous improvement loop – ideally in next-to-real-time.
Capturing metrics is critical, understanding how many interactions ended successfully (the consumer got the answer they were looking for), abandon rates of conversations (indicating that the bot couldn’t get to the source of an issue), if the customer was/is happy with the service provided by the bot (the consumer did not seek to speak with someone else) and more.
These parameters are key to understanding if your bot is making a positive contribution or not. If feedback isn’t what you were expecting, fix the issues and test again. Addressing the areas that need the most attention and improving them rapidly is key to a ever-improving chatbot experience.
As customer expectations continue to evolve and push brands to respond immediately with context and empathy, and as we interact more and more with bots in our everyday lives, the question should no longer be whether or not to build a chatbot, but how quickly can you start?
Author: Gordon White, General Manager, TSC
Bio:Gordon White is General Manager at TSC, (Sitel Group company). Gordon has spent most of his career at the intersection of digital disruption and consumer-centric innovation, working to develop and build deeper brand relationships through the creative application of technology, data and next-gen experiences.