2017: the year of real applications

Companies and researchers unveiled potential use cases of machine learning and its benefits were discussed and debated at length.

Nevertheless, as organisations were opened to the realm of possibilities capable with machine learning solutions, hesitation grew and few made the leap towards a data-driven journey.

But as the necessity of machine learning becomes more evident, 2017 will see trepidation and hesitation replaced by yearning ambition.

More than providing a competitive advantage, enterprises will come to see the technology as an essential part of their “survival kit” in an ever competitive market.

>See also: 10 cyber security trends to look out for in 2017

The rise and growth of IoT throughout 2016 has provided fertile ground for organisations wanting to adopt machine learning solutions. With most now possessing enough data to monetise their assets, the enterprises of 2017 will do more than merely toy with the idea of machine learning.

Building on the foundations

The machine learning discussions of today have evolved from those of early 2016. More than mere “talk” about its suspected ability to optimise business processes and increase ROI, machine learning has transformed into the technology enabling organisations to strive and survive in an ever unpredictable market.

Its adoption is seen as a crucial way of staying ahead of the curve as businesses begin to delve into their data to realise complete business potential.

This shift in attitudes towards machine learning will see various industries transform the way they work and achieve ROI in 2017. Existing business processes will change as companies adopt experimental approaches to new machine learning solutions.

The manufacturing, retail and telecoms industry are three industries set to be positively disrupted by the real life applications of machine learning.


The factory floor isn’t alien to change. The rise of Industry 4.0 has dominated manufacturing talk during 2016 and this coming year will see the fourth industrial revolution make way for real machine learning applications on the factory floor and across supply chains.

Up until now, manufacturers have approached data analytics as a way to organise the masses of data generated by the increasing number of connected machines.

While this has laid the foundations, the manufacturer of 2017 will move beyond looking at data analytics from an organisational point of view.

Instead, machine learning solutions will become part of routine decision-making. Indeed, it will prove pivotal in applying cost-saving optimisations to manufacturing processes.

For example, by using recommender models to determine the optimal make-up or design process of a product, manufacturers will see drastic reductions in cost.

>See also: Top 3 telecom trends for 2017

Predictive maintenance is another way in which manufacturers will benefit from real-life machine learning applications.

By identifying patterns in historical data sets extracted from machines’ logs, machine learning is able to anticipate faults and breakdowns before they happen. In turn, this will reduce or prevent machine downtime and cut down on cost of repairs.


The retail industry has also been subject to great change over the last year and today, the personalisation of products, offers and more has almost become a common aspect of the retailers’ differentiation strategy.

Whilst this approach has seen revenue increase and helped boost sales, it comes with limitations. Retailers must seek to roll-out differentiated marketing techniques and offers with the help of comprehensive machine learning solutions.

In an industry such as retail, staying ahead of the curve is key, and if retailers are to survive in the face of competitors’ innovative offerings, they must begin to use data differently.

2017 will therefore see retailers’ approach to machine learning change – more than a support for business strategy, it will become an integral part of operational decision making.

Being able to accurately forecast product demand during sales periods according to shopper behaviour patterns found in data sets, or optimising pricing policies according to shopper demand, will see retailers deliver more accurate offers and therefore increase revenue and ROI.


When it comes to data, the telecoms industry is far from short of it. Most vendors and operators recognise the value that lies in their data with many venturing into in-house analytics projects in a bid to extract this business potential.

However, while this approach is the right step towards implementing a machine learning strategy, many vendors and operators fail to realise the value in hiring the correct kind of team to carry out these projects.

>See also: The most disruptive enterprise technology trends of 2017

Indeed, many of the existing teams which move into in-house data analytics and machine learning projects mirror the mentality of the industry, one only slowly emerging out of the era of clunky legacy systems and unfit for the innovation needed for these projects.

If industry is to see this skillset gap reconciled come 2017, vendors and operators must realise that machine learning projects require innovative teams of CIOs and data scientists, capable of envisioning the roll-out of quick time to market offers and services, alongside telcos’ existing legacy systems.

These teams will also be able to embrace experimentation in an industry that is often driven by definitive results.

What lies ahead?

In 2017, these industries will see their processes significantly improved by machine learning advances. The potential of the technology will be replaced by real-life use cases.

Ultimately, manufacturing, retail and telecoms are just three of the industries set to be enhanced and optimised thanks to the various applications of machine learning. But while innovation and business prowess is in sight, the road is still long and winding.

As industries are faced with the arduous task of survival, adaptations will have to be made to existing processes to ensure organisations use machine learning to their advantage and see valuable results from their investments.


Sourced by Alexander Khaytin, COO of Yandex Data Factory

Avatar photo

Nick Ismail

Nick Ismail is a former editor for Information Age (from 2018 to 2022) before moving on to become Global Head of Brand Journalism at HCLTech. He has a particular interest in smart technologies, AI and...