Artificial intelligence – AI – is a complex topic that continues to remake the boundaries and possibilities of the gaming industry, especially regarding security concerns and capabilities. A similarly complex and invigorating discussion on the topic, and in particular how to personalise AI to fit your client base, took place at ICE London on Wednesday. The discussion was moderated by Gutsy Ninja co-founder Pooja Capellos; panel members included Rank Group Head of Digital Experience and Intel Kerry Dawes, Wyzer co-founder and CEO Abdalla Kablan, and Khel Group CEO Naveen Goyal.
In her opening remarks, Dawes stressed the importance of using data in the context in which you have gathered it, to retain relevance to the customer.
“It’s a mistake to wait for data warehouses before we start working with data,” she said, noting that it’s generally better to start early and work with what you have. At the Rank Group, she said, they embark on a lot of testing which involves reaching out to customers, prototype testing, user studies, and early design stages studies, as well as digital focus groups and AB testing. Dawes also made the point that it is important to “share the win” between your product teams, your digital and CRM teams, and your contact centres.
“Nothing beats asking the customer,” Dawes said. “We are not all players, and it can be hard for us to judge what customers want.”
Dawes also explained how these approaches have aided The Rank Group in generating predictive models that allow the company to spot early moves towards problem gambling situation as they are developing.
Being proactive, softer and less intrusive in starting conversations with customers identified through the use of these methods have helped The Rank Group to use the right tools early on to pinpoint these problem gambling situations. Often, in these cases, the company highlighted the activity of concern and pointed it out to the user, and subsequently made tools available to them to limit or control their gambling. 10% of customers change activity simply based on that, according to The Rank Group’s figures.
Does personalised content encourage playing more? No, answered Dawes – “we’re not being intrusive, we are trying to make it easier to find what they want to play. We are changing the conversation.” In doing so, she said, the company is reducing player churn.
Wyzer’s Kablan talked about owning and managing proprietary data as a company, comparing the quality and reach of digital identity to that of personal health. We might have records from different clinics, he said, and there may not be one digital identity solution to cover it all; the structure should work, he said, more like a bank account.
Kablan mentioned Virtual Assistants, saying that they haven’t yet gained our trust as allies and that we continue to have a “human/slave” perception about them.
Khel Group CEO Goyal brought a sense of expertise on data distribution to the conversation. He spoke on the challenge of integration and use of data and gave as an example a situation in which a provider might have gathered customer data from conversations with a user in live chat, on the telephone, and from website views, but that the databases that hold these data might have conflicting formats and structures.
He also recommended extensive AB Testing and the use of control groups to validate the framework of the recommendation engine being developed by a firm and stressed the importance of including customer feedback in any data repository.
Goyal was also quite careful to point out that the intended purpose of personalisation is not, in his view, to make players play more or deposit more, but to feel actively cared for and understood. Doing so, he said, would cause customers to be retained for a longer period of time.
