eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Utilized properly and focused on practical applications, AI and machine learning can be transformational in improving customer experience. Yet too often the technology is stuck at R&D or POC stages rather than full-fledged development across an enterprise state of maturity.
I see two main reasons for the disconnect:
- KPIs for strategic planning and enterprise measurement are not prevalent. Organizations must adopt strategy and in-line analytic reporting to be able to thoroughly and comprehensively apply AI and/or machine learning across an enterprise.
- AI has become the hostage of highly skilled data scientists who found their niche during the big data revolution, where they developed an approach to AI based on coding. While important, coding is not – and never will be – the common language spoken across all industries and all types of businesses. There needs to be a different approach available (versus an army of coders) for the scaling of AI across an enterprise.
Scaling AI and machine learning across the enterprise begins with a solid data foundation. High-quality, robust data that is structured at ingest is a requirement for models to analyze, learn and perform – especially when the end goal is real-time personalization.
For real-time personalization to be effective, marketers must have a unified customer profile, that unifies data from every source and of every type – structured, semi-structured, unstructured, etc. The customer profile must be updated in real time, to ensure that new data is analyzed in the context of previous data. Having a data strategy is a key requirement to enable marketers to execute contextually relevant campaigns that are always in the cadence of an always-on, connected customer throughout an omnichannel journey.
Personalization with Persistent and Accurate Customer Data
With a unified customer profile, the next step is to combine it with automated machine learning (AML). Here, we see why a reliance on data scientists for building offline models is an antiquated notion. If we accept that real time is indispensable for the delivery of a hyper-personalized CX, building offline models simply cannot keep pace. Customer journeys – and business goals – change at a moment’s notice, and models go stale – particularly because their construct fails to address the influx of data in the time it takes to operationalize a model.
By contrast, self-training, online AML models that evolve over time without human input ensure that a model ties directly to a metric a marketer is trying to push during development. Like natural selection, the model is tuned to optimize the metric, which guarantees that the model will be highly relevant and effective in moving the metrics a marketer intends to move, all without relying on human judgment.
With the infusion of machine learning with algorithmic optimization, marketers are freed to run dozens or hundreds of models, with optimization done against fleets of models simultaneously, all continuously assessing if the models are reducing error against the perfect solution for a particular metric. This type of machine learning paired with the experience of human marketers is critical to creating an authentic, personalized customer experience.
Break Through by Creating Digital Experiences
With an appreciation of the true power of AI and machine learning, it’s easy to see that one-off use case – such as a chatbot – are short-sighted. Unfortunately, despite AI and ML being capable of delivering such powerful experiences, they are not being applied consistently at the enterprise level.
McKinsey research shows that adoption and impact are often the byproduct of having executive support. The “State of AI in 2020” report concludes that what “separates the best from the rest” (read: highest bottom-line impact) are businesses that exhibit overall organizational strength and engage in a set of core best practices. That is, those that recognize the value of a solid data foundation and the need to think beyond implementing AI and machine learning for a one-off simple use case.
Collaboration and Utilization are Key to Success
Beyond putting the right technology in place, how can organizations begin to transform CX with AI that spans the enterprise? First off, there must be an acceptance that marketing is a mission-critical department with the potential to be the top revenue driver for the enterprise. There should be no doubt that achieving true segment-of-one marketing with hyper-personalized experiences is worth the time, effort and resources.
Technology, strategy, process and change management overhauls are all part of the equation. Organizations may be hesitant to make these investments, but satisfying customers with sustained omnichannel relevance is worth the effort.
The end result is having the consistent capability to land the perfect message, offer or content at the precise moment of interaction – whether it’s a customer landing on a website, a mobile app or email at the moment it’s opened. With the right analytics, it is not beyond the scope of AI to develop the optimal sequence and various alternatives that may be relevant based on a customer’s behavior. Importantly, the sequence and alternatives are continuously updated to match a unique customer journey.
Citizen marketers should look to build and deploy their own machine learning models with little manual effort or statistical skills – and thus reap enterprise-level benefits that consistently drive new revenue. Once these types of tools are truly embraced as a standard set of applications in the business environment, we will begin to see AI/ML scale across the enterprise to deliver value across a multitude of functions. We anticipate that day arrives sooner rather than later.
About the author:
George Corugedo is the CTO of Redpoint