Machine learning has been around for decades, but for the most of that time, organisations only used a few models, which needed arduous, time-consuming labour from PhDs and machine learning professionals. Machine learning has exploded in popularity in recent years, due to the widespread availability of standardised, cloud-based machine learning platforms.

Across all industries, organisations are now deploying millions of machine learning models across different business lines. Intuit, a tax and financial software company, began with a machine learning model to assist clients in maximising tax deductions; today, machine learning is used in almost every aspect of the company. Intuit has expanded the number of models deployed throughout its platform by more than 50% in the last year.

Lyft, for example, collects huge quantities of data in real time from the mobile applications of over two million drivers and 30 million riders. Millions of machine learning models are used by the firm to correctly detect abnormalities in route usage or driving habits that might indicate problems that need to be addressed right away.

But this is only the start. Machine learning's next step will bring about what scientists could only dream of: industrialising and democratising machine learning. We're on the verge of a big transformation, with purpose-built machine learning platforms and tools that can systematise and automate deploying machine learning models at scale, allowing all enterprises—not just global Fortune 50 companies—to leverage this transformative technology and become genuinely disruptive.

The road to industrialisation of machine learning

Automation is being used to industrialise processes and enable mass deployment in machine learning, which is a common trend across sectors. For example, the earliest automobiles were created by boutique manufacturers such as Duryea and Packard, who constructed fantastical luxury cars in restricted quantities due to the time and effort necessary. Ford Motor Company flipped this concept on its head by standardising vehicle design and manufacturing techniques to establish an assembly line, allowing mass consumption of the automobile and permanently altering transportation and commerce.

Nine decades later, the software industry underwent a similar shift, transitioning from a collection of exquisite, bespoke programmes built by a few skilled coders to a methodical engineering discipline that is now widely accessible.

Integrated development environments, debuggers, profilers, and continuous integration and continuous deployment (CI/CD) technologies enable programmers at all levels to construct strong programmes today. The capacity to mass-produce programmes has resulted in widespread software consumption, making software an important part of how we live and work.

Machine learning is through a similar period of industrialisation. To succeed, we must not be persuaded that creative and imaginative machine learning demonstrations—such as composing poetry and creating smart dialogue in video games—are the norm or the way ahead for machine learning.

These boutique, "proof of concept" displays, like future concept automobiles that delight onlookers at auto shows, have captivated imaginations and inspired excitement, but they cannot be readily duplicated or scaled. Not only that, but they're also prohibitively costly and add little to the bottom line.

Machine learning platforms and analytics suites are now capable of doing what computer scientists envisioned in the 1950s when artificial intelligence algorithms were first proposed. We are seeing a new business shift in almost every area because to advances in machine learning infrastructure and tools.

(This is a slightly modified version of an article originally published in Info World. The original article can be found at