The role of AI in Customer Success 2.0
It is one of the most talked about subjects in technology – artificial intelligence. According to some, it is going to revolutionise the world. Kevin Kelly (of Wired fame) describes AI as the second industrial revolution in an insightful TED talk. This blog examines how AI will change the field of customer success; specifically the contribution it makes to the development of CS 2.0, where the product is at the heart of and the main delivery vehicle for customer success.
Let me begin with a grossly simplified view of the elements of AI.
Data capture and categorisation -> Machine learning -> Propensity modelling -> AI applications.
Data capture and categorisation
AI without rich data is an oxymoron. Tools for smart capture and categorisation of data are the foundation of AI. In the field of CS, a single customer view has always been; with the advent of AI is not optional. Tools are emerging that capture data from documents and others that use AI to rate the quality of data. And of course we are all aware of AI based intelligent assistants that recognise and respond to speech. These will be used increasingly in a business context.
Machine learning
Machine learning is the automation of pattern identification in large data sets. It answers questions like “What product usage data correlates with churn are the characteristics and activities that correlate with retention?” or “What customer characteristics lead to churn?”
Propensity models
Propensity models use the patterns identified via machine learning to predict outcomes.
AI applications
These are applications of AI to do specific tasks. These may be full automation of tasks or automation used to guide people in the completion of their work.
Here are a number of tasks where AI applications will contribute to customer success. Many of these are already in use, although most are still a minority sport.
- Conversational discovery. Natural language interactions to collect information on the customer’s goals, challenges and modus operandum.
- Customer ROI guidance. Delivered in the application itself, AI will identify the actions a customer should take to achieve their objectives/desired outcome.
- Personalised implementation plans. On-boarding tailored to a specific customer’s situation and goals
- Next best actions will replace playbooks. Playbooks are typically a company’s interpretation of what a customer should do next. Next best actions use more granular data patterns to understand context and suggest an action.
- Customer health/engagement scoring. AI driven health dashboards will improve the reliability of scoring and will self-adjust as continuous machine learning identifies changes in the underlying data.
- Feature targeting. Identify customers that can gain greatest benefit from new features and should therefore be targeted first.
- Sentiment analysis. Discerning the behaviour and intent from the content and tone of customer conversations.
- Upsell targeting. Listing the customers most receptive to additional purchases and why.
- Content curation. Identifying the content which will be positively received by which customers.
- Dynamic pricing. Suggesting the best price for up-sells and renewals.
Here and now
I am not suggesting that these capabilities will be widespread this year, not even next but I think many underestimate the maturity, sophistication and speed of development of the technology. Here’s a few things to take note of.
- Research from IDC into the use of AI in CRM suggests that 55% of companies expect to have implementations (not pilots) established next year (2018). This will generate additional revenue for the companies using AI of $1,100 billions by 2021.
- AlphaGo, a Google AI programme beat the best two Go players in the world in 2016. It was coached on how to play. Its’ successor, AlphaGo Zero was just given the rules of the game and learned how to play itself. It took AlphaGo Zero just three days to reach world beating standards using a fraction of the computing power. Professional players say it uses moves never seen before. This is what neural networks do and they do it far better and faster than people can. Imagine giving an AI tool a set of business principles and letting it learn how best to deliver customer success.
- AI tools are widely available and will become a utility within five years. The IDC report above suggests spending on AI will grow to $46 billions by 2021. That is more than the forecast market for CRM, itself one of the biggest technology markets. In November 17, Salesforce launched MyEinstein, a tool to allow administrators, not developers users to build their own AI applications. Almost all customer success software has, or have plans for using AI in their applications.
- Andreessen Horowitz, one of Silicon Valley’s leading technology investors believe that AI is a fundamental platform of the same order of importance as cloud and mobile.
AI and competitive advantage
Whilst the technology is vital, I don’t think it is the real source of competitive advantage. Given its ubiquity, how can it be? Sure, early movers will gain an advantage but technology that everyone has access to won’t sustain that. There are two factors that will create the winning edge in the use of AI in customer success.
The first is data – the richness of the picture that companies can build about their customers. Single view of the customer has always provided an edge. AI provides the means to extract meaning out of much larger and more diverse data sets. The more dots you have, the more patterns you can spot.
This is reliant on the second source of advantage; mutuality – a true customer first culture. Mutuality is not just a belief that customer satisfaction or even their success drives what a company does but about taking actions that are good for the customer and the company. It is about using the customer’s data to their benefit, not just yours. It is about doing the right thing for the customer, not just selling them anything you can. Customers will soon have the means to control access to their own data and will increasingly restrict access unless they can see something in it for them; unless they can see that the company is doing the right things for them. Remember, it doesn’t matter how good your AI applications are, they are useless without data.
Changing CS
AI will radically change the customer success landscape. Routine tasks will disappear: chatbots are already replacing the support team and that will extend into the simpler, routine tasks of customer success, particularly as more and more companies build this into their application. So what of the CSM? Well if they spend their time doing routine tasks like on-boarding, training and low-level process change; they will also go. If they are change mentors, guiding people through the human aspects of change, then they will stay. In proportion, we will need more of these. We already do it is just that we can’t afford them given everything else a CSM has to do.
Make no mistake, AI will improve productivity and that will impact jobs. In the early stages, AI will augment people but as confidence in systems grow, AI will take over some tasks. Whilst I have concerns about individuals who won’t or can’t adapt, I am not fearful of the overall impact. President John F Kennedy said “We believe that if men have the talent to invent new machines that put men out of work, they have the talent to put those men back to work.” Throughout history, new technologies have changed work and jobs but the overall effect has always been growth. People just do different things. Question is, if you are in CS, what are you going to be doing?