At Shopworks we have been busy with an exciting project in the last few months; we are taking the first steps towards adding artificial intelligence (AI) to our cloud based staff-scheduling engine. As we have been exploring different ways to take advantage of this highly disruptive technology with various suppliers it has made me think about other software services that are likely to be disrupted by AI and machine learning.
One area that I know a fair bit about is CRM, having implemented a few solutions in online gaming over the years. And last week I found myself discussing AI and technology strategy with the CTO of a large UK land and online gaming operator and he was telling me about his plans to rebuild his company’s in house CRM system.
At first glance, that seems to be right on trend for tier 1 UK operators. The vogue in online gaming in recent years has been for operators to reduce their reliance on the big gaming specific software suppliers by building their own customer wallet, CRM system and user interfaces, whilst still buying in bonus engines, sports book platforms and games engines. The aim has been to increase speed to market and own the core IPR and differentiators of the business. CRM in particular, is seen as a core competence by most online gaming companies. Some operators have built several hundred different CRM cycles all backed up by propensity models, which are trying to predict when a customer is most likely to stop or start interacting and then triggering a relevant CRM action. The big online operators are spending over £100m per year in marketing and a quick look at William Hill’s annual report shows they recruit around 1 million new customers per year and have around 2.5 million active users, a number that grew by around 200,000 customers between 2014 and 2015. That means that they have had around 800,000 customers lapse in a year and I am sure the other tier 1 operators spend a similar amount on marketing and have similar issues. As you can see with these sort of numbers, CRM matters to these companies and so it is no surprise that over the years online gaming has been at the forefront of CRM technology and operators see their huge library of CRM cycles and finely tuned incentives engines as core IPR and something they don’t want to delegate.
However, I would argue that AI and machine learning are about to change all of that and if it does, building an in house system could put these tier 1 operators at a disadvantage to smaller operators. Tier 2 and 3 operators don’t have the R&D budget to build so they go with cloud-based operators who are already adding AI to their offering and may use this to overtake their larger competitors. The main reason I think in house systems will fall behind cloud-based providers is because of Artificial Intelligence. AI will extract insight from CRM and other data to tell the full customer story, by customer. Machine learning optimises how an organisation understands and engages with its customers—both at the individual customer level and across an entire customer base. Leading to a highly personalized customer experiences on a much larger scale than previously possible. Even if operators try to add third party AI to their in house systems they wont benefit from the huge data sets and shared customer profiles that AI requires and because global cloud-based CRM providers are buying up AI companies at a rate that will make them hard to compete with.
We are all using AI in our day to day lives; Google search, Apple’s SIRI and Facebook’s picture tagging all rely on the input of huge amounts of data in the cloud to feed the machine learning platform. AI needs a lot of data and although William Hill’s 2.5m active customer base is impressive, Salesforce, a well know cloud based CRM provider has over 150,000 businesses using its software. The growth in Software as a Service in the Cloud is providing the huge amounts of data that machine learning needs to add real value.
AI is predicted to dominate in several marketing functions. These include marketing automation, which is another way of describing the CRM cycles that online gaming operators have worked so hard to develop. AI will also see the evolution of many other marketing tools such as chatbots and speech recognition.
Global CRM suppliers such as Salesforce, Hubspot and Sugar CRM are all investing heavily in AI. Saleforce recently launched “Salesforce Einstein”, having bought three AI companies in 2016. They recently purchased Implisit Insights Ltd, a Tel Aviv-based sales forecasting and intelligence technology provider to add to their recent acquisitions of PredictionIO, a machine learning platform provider, and MetaMind, an artificial intelligence software provider. This sort of investment will be difficult for any single operator to match.
With Einstein, Salesforce claim they “are giving marketers the ability to shift away from using analytics that only look at past behaviour to analytics that predict the optimal timing, channel, content and audience for any marketing message.” The product employs tools such as:
• Predictive Scoring: which enables marketers to judge how likely it is that customers will perform a given action such as deposit or when they are likely to lapse.
• Predictive Audiences: which will allow use Predictive Scores to build audience segments of people showing multiple predicted behaviours in common. Marketers are then able to take this insight to incentivize the customer to complete the predicted action.
• Automated Send-time Optimization: this helps marketers to maximize email marketing ROI by automatically delivering their messages exactly when subscribers are most likely to engage.
These are only the first wave of AI tools for CRM and marketing and Einstein only launched in Q3 of 2016. If you look at how far AI has come with Google Search and Apple Siri, I predict we will be seeing a lot more tools in the next few years. The build time of a new CRM platform for a tier one operator in probably over 18 months, which is a long time in computing; I believe a company should be comparing what their end system will look like and comparing it with the likely state of the cloud based alternatives. AI also works best when humans train the software to think for itself, the implementation of several years worth of tested CRM cycles would act as a powerful training tool for an AI and may be a better use of tech resource than building a new system.
I know of one Tier 2 UK operator that has recently opted for Salesforce for its CRM tool and another cloud-based tool for its customer support. Support is another area that will see AI being added to cloud-based solutions to disrupt current processes. I shall be watching their progress with interest.
Despite all the progress I think it will be a few years before a cloud based AI solution overtakes the best in house systems, however with the time frame involved in building a new CRM system, I certainly recommended to my friend that they look at AI now. What is certain is the build v buy debate will be spiced up a lot in the next few years by AI and the cloud and at Shopworks we believe that for those applications that benefit from access to big data the era of the build is coming to an end and the era of cloud based, software as a service (SAAS) using AI is just beginning.
By Ian Hogg, of Shopworks who is also Chairman of GamCrowd