Improving customer experience with AI & machine learning
By Craig Harper-Ashton, Multichannel Director, Salmon
This new article - the second by Craig to feature in IMRG - focuses on Artificial Intelligence and what it means for businesses.
Have you heard? This is the year of Artificial Intelligence (AI), with the rise of the machines, machines building machines, cyborgs, new realities and self-learning robots all coming to the fore. You don’t have to be a science-fiction fan to see the early signs: it could very quickly lead to a self-made leap in human evolution, with dramatic changes to the society in which we live.
We are already on the road, and although it’s looking like a very interesting future, there is plenty to be gained from the technology right now.
So how can companies today use this tech to improve their customer experience?
The Frictionless Concierge
Providing a good customer experience has always been about reducing friction, and thereby increasing convenience. Each hurdle we put in the customer’s way adds to their inconvenience, making their experience incrementally harder each time, and you don’t need many for it to become a bad experience rather quickly.
Automated “chat” windows (chatbots), AI-based telephone services, natural language interpreters and voice interfaces can all provide your business’s data and services at the touch of a button and provide a frictionless service. There are many services already like this that supply the abilities to provide order information, placement of an order, returns services, general content and of course household services with integration to lighting, heating and appliances.
The further you can embed this assortment of services into your business, the more value you can offer your customers. If a customer wants to check the stock position in a local store at 3am, then that’s good value and a good experience. But how about a “frictionless concierge” that had the ability to modify the entire supply chain at 3am to make sure the product was there for said customer when they wanted it? Now that’s a really good experience and in today’s world, this is driving customer loyalty.
Personalisation has been around for some time. However, it still lacks precision and falls short somewhat of creating a personal relationship with each and every customer, as if they’re your only customer. Even the mighty Amazon has failed to get this licked - just yet.
Although personalisation engines are mature and extremely powerful, how often do we see marketing based on something that we bought a year ago, or an item other users have been browsing that may be related to what I’m looking at in some way?
The tools themselves are capable of so much more but are under-utilised because they are still built fundamentally on rules that must be configured by humans. Laziness is not the prime prevention for improvement; it’s the complexity that restricts our ability to personalise to the level that we wish. Setting up user segmentation alone is not a trivial task for any sizeable organisation, and to then nest a significant number of overlapping rules makes the job so difficult that we return to fashion only the simplest interactions.
Artificial Intelligence changes this fundamentally. It’s an area where it demonstrates its value very clearly, and where machine-learning comes into its own. Not only does it configure complex rules, it creates a continual tuning cycle that learns from user interaction and tweaks the rules accordingly.
Prediction: Making Sense of it All
Machine learning is based, to a very large degree, on pattern matching. It is this area of AI that accounts for it being so pervasive today. AI is now at the level where it can “view” an image and actually make sense of its component parts, recognise objects and detect patterns within very complex data. Humans have always been exceptionally good at seeing patterns in data, but where AI excels is in the amount of data it can work on, and how quickly it can turn this data in to information.
Why is this pattern cognition ability so important? The understanding of patterns is only one small step to prediction, and all sorts of really interesting things start happening when you extrapolate a pattern in to a prediction. For instance, you can weight your search return accordingly if you can predict that Thursdays tend to inspire products from a particular category, and so you may wish to automatically promote some content if you can predict an appropriate interest level from historical data.
These however are not examples of the true power of prediction. In fact, we must think much harder to improve the customer’s experience in order to make it more meaningful. How about ordering a product for the customer without them buying it? Too “out there”? There’s a pretty well-known ecommerce giant that’s already thinking about doing just that. Or when it comes to stock management, effectively predicting what stock needs to be moved from distribution centres, and to where?
Or when it comes to stock management, how about predicting that one of your customers is unlikely to be at home on the day that they’ve ordered, so why not move them up to express delivery for free? This is what a great customer experience looks like.
The fact is that even if this all sounds well beyond your means currently, it’s very likely that you’ve got all the data you need already.
The Not Too Distant Future
AI is being applied to almost every category and every type of business. It won’t make a bad product sell and it won’t make a bad business succeed. It is, however, a powerful weapon in the arsenal to improve your relationship with your customers and offer them more convenience, and allow your business to provide products and services on your customers’ terms.
To understand how best to add value to your customer is actually the easy part. Think of any and all friction between you and your customer – the queue at the checkout, trying on clothes to see if they fit if you don’t have the time, waiting in for your delivery, the inconvenience of organising returns, not having your coffee ready when you would like it, having to dig out your payment card each time you make a purchase. The businesses that offer more choice to remove friction are the winners of the future. You can’t afford an army that looks after each customer but it’s becoming very clear that AI can assist in some amazing ways.
The customer of the very near future will not demand that businesses make them feel special and loved, they will expect it.