July 21, 2017
How to ensure your technology stack is fit for machine learning
By: Matthew O'Kane

Machine Learning is the new frontier for businesses. Smart, data-driven algorithms capable of learning unassisted are transforming the customer experience, creating always-on, real-time intelligent conversations with customers while cutting costs. As I have discussed previously, businesses today must start looking at how they can implement large scale Machine Learning applications across their operations to ensure they can deliver a competitive, “living” customer experience.

I believe there are three core considerations businesses need to bear in mind when scaling their Machine Learning capabilities: the workforce, data processes and technology. Today, I want to look at the third of these, and explain which technologies are key to realising a successful Machine Learning environment.

When it comes to Machine Learning, businesses must throw the old technology rulebook out the window. Machine Learning, and advanced analytics in general, requires a very different technology stack to the MI/BI implementations of old. If Machine Learning is to deliver new insights rapidly to support a real-time service, then businesses need to think again. Broadly speaking, there are three core considerations to bear in mind:

  1. Machine Learning analytics is cloud analytics. Machine Learning thrives on vast amounts of data being processed rapidly and at scale. The cloud is by far the most agile and cost-effective delivery method for the compute and storage capabilities required to do this. What’s more, Machine Learning algorithms are exceptionally easy to run in parallel; so, providing bursts of thousands of compute units, when required, through the cloud ensures better answers and insights can be delivered in shorter timeframes.

  2. Use the right language. For machine learning, the statistical programming language R is key. This is the language most commonly used at universities and is well known by the data scientist community. What’s more, as an open-source framework it receives most new algorithms developed by the research community, and so its use will help your business stay at the cutting-edge.

  3. Ensure real-time outputs. If Machine Learning is to help you deliver a true intelligent, automated conversation with customers, then real-time responses are key. Increasingly, analytics outputs automate decision making in real time for services—for example, price comparison sites, which aggregate quotes to provide to consumers. Near-real time responses to high-volume requests can be delivered through APIs calling the analytics models on a cloud‑based analytics technology platform.

By having the right people, processes and technology, Machine Learning will transform your business. The end result will be a finely tuned automated system that runs across the customer journey, continually experimenting and learning to understand in real-time how customers will react to individual treatment strategies and prescribe the best course of action every time.

This is the Customer Journey 2.0: a machine-enabled experience that is always-on, flexible and laser-focused on giving the customer exactly what they want, as they want it. You can be sure that your competitors are already looking at implementing this new approach to transforming customer conversations. Make sure you can remain competitive by doing the same.

Popular Tags

    More blogs on this topic