Applying AI (Artificial Intelligence) and ML (Machine Learning) to customer and marketing analytics has the potential to drive substantial value and benefits. Yet the success of applying these sophisticated tools, so far, has been somewhat mixed. The gulf between those companies who have been successful to those who have been disappointed in the results is likely to widen as the technology and its deployment matures.

The holy grail of CX is to obtain a “360 Degree View of the Customer”. The 360-degree customer view is the idea that companies can know everything about their customers by aggregating data from every customer touchpoint. The application of AI and ML would appear, at first glance, a natural fit.

Cindy Maike, VP Industry Solutions for Cloudera, points out, however, this pursuit to know everything about the customer has led organisations astray in their AI and ML implementation. “The 360 degree view of the customer is a very large and complex use case. For AI to have a positive impact, the 360 degree view needs to be broken down into much smaller and manageable use cases and segments, instead of trying to look at everything involving customers.”

“Think, first and foremost, where are the problems. How can we actually look at a smaller use case that has a positive impact? And break the 360 case down into smaller use cases that we can go after and achieve”.

Why AI for CX and marketing analytics

The main drivers for applying AI and ML in CX to customer data centre on customer retention and growth. Nick Hoskins, ANZ country manager for Cloudera, says, “By applying AI to customer data and analytics we can identify the likelihood of a customer churning and the reasons why to prevent it from happening. Red flags can be set up to notify if a customer is at risk of churning.”

“Based on previous buying behaviours we can also create very personalised customer journeys, which allows us to up sell or cross sell to our customers. AI can help us identify how much an individual customer is worth to us, in actual dollar terms, and what we can do to increase that. AI can also be used measure and analyse the sentiment of customers and what they feel towards the brand.”

CX Decision makers don’t understand the technology

A recent study Artificial Intelligence: A Framework to Identify Challenges and Guide Successful Outcomes,” revealed that decision makers often do not fully understand the technology and have not thought through the true costs of implementing AI in their businesses.

Expectations of this new technology often exceed what can actually be achieved. Jeff Catlin comments in Forbes, “In some ways, AI is its own enemy. Sure, it has the potential to help solve the biggest problems we face. But potential isn’t the same thing as achievement. As high as our hopes are for AI, we need to temper our expectations a little. AI may get there one day, but it isn’t there yet.

Nick Hoskins, ANZ country manager for Cloudera, advises, “Organisations who have been very ambitious in this space and not necessarily getting what they want in time – have done all the work in defining what they want yet have failed to include IT and the technical teams early enough in scoping the project. IT has almost been an after-thought, where they are simply told to build something rather than being involved in defining and scoping the project”.

“Companies who are doing it well are building the right teams that represent the various business units, technical teams and stakeholders that need to be involved from the initial concept of the project”.

Moving forward

There are a number of factors you need to consider before you decide to invest in AI and ML to solve any CX related problem. “First and foremost”, advises Maike, “you need to clearly define and understand the business outcomes you want to achieve or specific problems you are looking to solve. Are you looking to increase customer retention or life time value? Or are you looking to provide greater personalisation to increase share of wallet? Whatever it is it needs to be specific and achievable.”

AI and ML require lots of very accurate data to do their job effectively. Maike says, “A data discovery exercise is an essential step to determine the ability of AI and ML to deliver the outcomes you are looking to achieve. If you don’t have enough data or if that data is significantly inaccurate then it is impossible to create a workable model that can be used to analyse the data and make predictions.

Don’t get caught up in the hype and the excitement

There is a tendency to get caught up with all the hype and promises. AI and ML are tools, they are not magical panaceas that can solve all our problems. Given the right data set AI can greatly expand our capacity to predict customer behaviours and assess the level of sentiment they have towards the brand. It can help us to make better informed and considered decisions.