Training machines instead of programming them is delivering all kinds of opportunities for automation and autonomous behavior. It is not only in telecom, service provision and security that machine learning is falling into use, but these are emerging as bellwether use cases.
The history of the software industry is one of automating repetitive human processes. It started in finance, and has spread more or less across all aspects of business. Business rules are encoded in software to automate those repetitive tasks. But what if those tasks vary, or what if you don’t even know how to perform the task, or what you’re looking for? This is where machine learning (ML) comes in. Facebook, Google, Microsoft, IBM and others have been using machine learning for years – think machine-driven targeted advertising, translation, facial recognition, self-driving cars or tailored news feeds. ML is also widely used in sales funnel forecasting, retail price optimization, predictive analytics, cloud infrastructure management, mobile customer engagement, IT monitoring and security.
However, most ML technology vendors are not doing more than $30m in annual revenue, and defining a new market is tough with so many incumbent vendors offering ‘slideware’ to describe their future plans to use ML. Firms such as arago, IPsoft and Blue Prism stand out as bringing ML to IT operations. Security firms such as Darktrace, Securonix, E8 Security, Distil Networks, RedOwl, LogRhythm, Splunk and Rapid7 are using ML techniques. This report examines some of the progress being made with ML techniques and what can be expected. It is not only in telecom, service provision and security – where real-time operation and preemption are key – that ML is falling into use, but these are emerging as bellwether use cases.