Is AI ready to deliver more than Teenage Kicks?
Claire Sporton, SVP, Customer Experience Innovation, Confirmit
If all the hype is to believed, the robots are not only coming, they are the answer to many of the challenges we face as CX professionals. Although CX budgets are expected to either remain the same or decrease in 2018, it appears that many organisations expect to use customer experience to compete, and intend to turn to AI to provide the vital advantage. However, there is a real debate about whether AI is the silver bullet we are all looking for. Like any new kid on the block, is it ‘totally wicked’ or still a bit ‘dazed and confused’?
AI offers some fascinating and exciting possibilities, from driverless cars to drone-enabled delivery of our weekly groceries. We are learning to embrace virtual and augmented reality and are starting to get used to the idea that there can be people-literate technology as well technology-literate people. However as Gartner says, the focus is currently on ‘narrow AI’ or highly scoped machine-learning solutions for specific tasks.
In the customer experience arena, this means that most real-life applications tend to be limited to clever indexing and good search functionality, with either a pleasing avatar at the front end or a quirky name to make it more approachable. In other words, the AI that we’ve seen at the movies (I, Robot, Blade Runner, Star Wars) may be what people are excited about but it isn’t here yet.
Early adopters in the CX arena may have rushed to deploy AI in the form of chat bots but they have quickly being ‘found out’ by customers for what they are – pre-programmed, script-based automatons. Yes, they can answer simple questions to reduce the load on contact centre agents but they still have to refer more complex or contextually-heavy problems on to a human to handle. They are not yet the conversational interfaces that we might hope for and are incapable of hearing the frustration in your voice or seeing the frown on your face.
Sentiment analysis and emotion recognition technology will start to offer improvements on this front. However the ability of AI and IVR to successfully reduce the number of calls through to the contact centre does not equate to customer satisfaction. It’s no surprise therefore that The Temkin Group says that the short term hysteria for chat bots will subside.
Don’t misunderstand me, I’m no Luddite – I love new-fangled things; but AI is in danger of falling foul of the hype cycle – unrealistic excitement followed by poor experience and disappointment. I’m very aware that in many respects, AI is less Artificial Intelligence and more Adolescent Intelligence. It’s like a promising but inexperienced teenager that needs a lot more exposure to the real world before it is ready to fly the nest.
It still requires us to direct it to look for the insight we need by providing it with clear commands in the form of algorithms. It has to learn or train on a large supply of quality survey, operational and behavioural data which isn’t always available to CX professionals (and customers are not necessarily going to respond well to being practiced upon). It can take on boring or unpleasant tasks – crunching huge volumes of data faster and doing the ‘heavy lifting’ on our behalf – so long as we remember to supervise searches and apply the human brain to the results presented to us for consideration.
Yes, AI can optimise analysis, provide answers based on key criteria and we can use regression analysis to identify or prioritise the best course of action. But AI cannot possibly use the data it has gathered to take account of human issues such as preference or office politics. It can’t hypothesise outcomes based on unknown variables such as the weather or a customer’s undisclosed claustrophobia. It can reduce the cognitive load on CX professionals, but it cannot be expected to make the final decision.
So, although it is true to say that CX is at a cross roads, I don’t think that AI’s ability to identify patterns and offer predictions based on mathematical probability and trend analysis will usurp the role of the CX Professional as such. Too many decisions relating to CX cannot be taken in isolation, without the context, environment or the undefinable human factor. Instead, AI in the form of text, social and predictive analytics should been as a liberator, reducing labour-intensive spreadsheet analysis and leaving us free to increase our focus on nuanced decision making and culture change.
In this respect, I agree with Forrester’s view that we can expect to see CX practitioners evolving into CX ‘coaches’. They will no longer have to simply ‘measure CX’ but will be able to leverage the insight hidden in vast volumes of structured and unstructured data, to focus on the much harder strategy development and human side of CX, thanks to the AI tools that can track KPIs and present trends on their behalf.
CX professionals clearly need to adapt to this new AI-enabled way of working, learning to compete with technology, not against it. However, as we outlined in an earlier blog offering a ‘CX Careers Guide’, if we can blend skills from traditional research and operational management with human skills such as empathy and real world experience of how businesses are run and what makes our customers ‘tick’, we should be able to look out from behind the spreadsheet and use CX strategically to achieve key business objectives.
What we need to take on board going forward is that we still need to ask the right questions, or rather programme the right algorithms, in order to ensure that we have good quality data to interrogate and therefore sound insight on which to make decisions. We will need to respond to the shift away from traditional surveys (only deploying them when we are prepared to take action) to more non-scripted, conversational interfaces (speech recognition) and focus innovation on the customer journey. We will need to recruit and train the analytics experts that will be the key to making predictive, behavioural, text and social analytics an effective tool in the CX arsenal.
Fortunately, The Temkin Group describes 2018 as the ‘Year of Humanity’. It seems clear to me that the challenge is getting the balance right between what AI is currently capable of offering and the human support it requires to deliver. If, as predicted, companies renew their efforts to create a customer-centric culture and strive to deliver on their brand values, ‘Adolescent Intelligence’ could help to prove the strategic value of CX. Provided we don’t use this narrow form of AI ‘for the sake of it’ and don’t expect too much, too soon from the new kid on the block.