The AI & Machine Learning Tipping Point: Practical Application To Scale Across the Enterprise
It will be another year before companies truly start to scale their machine learning initiatives across the enterprise and move beyond pilot projects. We are still on our way to effectively using AI to deliver the types of 1:1 experience that are now possible, writes George Corugedo, Redpoint CTO and co-founder.
It will be another year before companies truly start to scale their machine learning initiatives across the enterprise and move beyond pilot projects. We are still on our way to effectively using AI to deliver the types of 1:1 experience that are now possible, writes George Corugedo, Redpoint CTO and co-founder.
Utilized properly and focused on practical applications, AI, and machine learning have the power to fundamentally change the way brands engage with customers by vastly improving customer experiences and removing friction in a customer journey. These advanced technologies improve the customer experience (CX) by delivering hyper-personalized, relevant interactions at a scale far beyond the capabilities of manual processes – allowing marketing teams to deliver personalized experiences that are in the context and cadence of uniquely individual journeys.
Personalized experiences powered by AI and machine learning drive revenue. Consider a recent Dynata survey where 70% of consumers said that during this holiday season, they would be highly likely to exclusively purchase from brands that personally understand them. Because brands are increasingly competing on CX, I foresee 2021 as a tipping point for AI and machine learning to finally receive their just dues as instrumental for creating relevant, hyper-personalized experiences that customers value even ahead of price and product.
Until now, many AI and machine learning use cases focused on improving customer experience have been surface-level – think of a smart chatbot, face recognition software, or interactive display boards. But as we move past the hype and marketing teams begin to see data-driven companies moving the needle, consensus will build that incorporating advanced technologies into the Martech stack to build a deep understanding of individual customers changes the playing field.
Learn More: How AI Can Help You Map the Buyer Journey in the COVID-19 Economy
It Starts With Customer Data
When one-off use cases are discarded in favor of practical applications based on a deep customer understanding, organizations will place a greater value on customer data – all customer data. Transactional data, in other words, barely scratches the surface for a single customer view that is required for data-driven marketing teams to deliver a personalized CX throughout an omnichannel journey. First-party, second-party, and third-party data, combined with structured, unstructured, and semi-structured data, begin to form a comprehensive and accurate single customer view. When the data that forms this unified profile is updated and made accessible in real-time, marketers have what is called a golden record – a complete profile that is always up-to-date.
Prioritizing customer data will become sacrosanct for ambitious marketing teams when they begin to see that decisions and predictions based on old, inaccurate, or incomplete data cannot possibly be intelligent. Worse, basing a decision on old data has a strong likelihood of introducing friction into an otherwise seamless CX, derailing a customer journey.
Applying automated machine learning (AML) to a customer golden record results in intelligent decisions – at scale – that are always in the context of a customer journey. With AML, marketers are able to have fleets of smart, self-learning models running simultaneously. Embedded, in-line models truly put the power of machine learning in the hands of marketers instead of data scientists because these types of models never go stale – meaning they never have to be taken off-line and rebuilt whenever there is new data or a new business goal or metric to chase. Rather, multiple in-line analytic models incrementally adjust and find opportunities for growth – they continually chase a pre-defined metric until there is an undisputed winner that will optimize a customer experience for a given moment in a journey.
Developing, updating, and applying in-line, self-learning models in real-time on an ongoing testing basis provide businesses with the means to truly capitalize on the moment of interaction with a customer – a critical need for personalized, real-time customer experiences.
Avoid Friction With Real-Time Relevance
Let us examine a typical retail customer experience to see the power of AML in action. Say you return home from purchasing a sweatshirt at a retail outlet. A few minutes later, you check your email and see a message from the retailer – a $10 coupon off any sweater! That is known as friction – it is a disconnect that introduces frustration and confusion. Can I apply the discount to the item I already purchased? Or do I have to buy another sweater? For the customer, it sounds like a lot of bureaucratic red tape to try to redeem a coupon that may or may not be applicable to their situation. Now, picture instead the same customer returns home to find an email thanking them for their purchase, with a $10 coupon off a matching scarf or another complementary item and a link to redeem online. This is relevant, personalized, and displays a connection with and understanding of the customer journey.
Automated machine learning makes the latter experience possible at scale for an endless number of touchpoints or permutations of a customer journey. It delivers the next-best-action that is optimized for a specific journey at the moment in time. And that is the key difference between using AML to compete on CX and using AI and machine learning to nominally improve a one-off experience.
Learn More: Why B2B Personalization Fails – and How to Change That With the Help of Well-Informed AI
Leave the Old Mindset Behind
The question, then, becomes why is the enterprise – until now – hesitant or reluctant to embrace AI and machine learning if the tools indeed provide such demonstrable benefit? It is not so much a failure to appreciate the power of scale – in this case, “scale”, meaning the ability to perform analysis on larger and larger data sets. It is more the scale of the application of AI across an enterprise to solve a multitude of problems.
Breaking AI and machine learning out of the rut, if you will, of the R&D or POC stage and into full-fledged deployment across the enterprise state of maturity is another tipping point that will come to pass in 2021. Two reasons explain why this will happen. First, enterprise strategic plans will finally reveal the fact that advanced technologies can be effectively applied to driving enterprise metrics, ROI, conversion rates – or any relevant business metric. These KPIs typically are not as prevalent across the enterprise as one might imagine, but this will change because, in the wake of COVID-19 and changing customer behaviors, businesses will have to re-examine their value propositions – which will be an enterprise effort.
The second reason AI and machine learning have yet to scale across many enterprises is that, for the most part, advanced technologies have become hostage to highly skilled data scientists who found their niche during the big data revolution, developing complex coded models to tackle big data problems. That entrenched mindset is being phased out. While the specialized skill of coding certainly has its time and place in an organization, it has never become the “common language” spoken across all types of businesses and industries. When the first wall comes tumbling down – the recognition that AI and machine learning can drive enterprise metrics – the second wall that until now ceded advanced technologies to data scientists will also fall.
Instead, we will begin to see a different approach that recognizes that scaling AI across an enterprise will not require an army of coders. Tools that embed AML in digital experience platforms will break through because they put the power to deliver personalized experiences at scale in the hands of operational marketers, not data scientists.
I believe that in 2021 we will begin to see AI and machine learning truly embraced as a standard set of applications in the business environment, which will unleash a torrent of innovation as the advanced technologies scale across an enterprise to deliver value across a multitude of functions.
