How traditional industries will reap the benefits of machine learning

The hype surrounding machine learning and artificial intelligence (AI) is higher than ever. Nevertheless, despite much anticipation, many businesses are still hesitant about the technologies’ ability to change their business.

This attitude is far from surprising. With the media and technology experts focusing on the most visual consumer-focused applications of the technology (AI assistants, chatbots, self-driving cars, etc.), there is little doubt as to why traditional organisations believe that disruption will only happen within the “new” economy, leaving them unaffected by the new arrivals.

>See also: What is machine learning?

But that is far from the reality. These new technologies are not only building new markets, they are also transforming traditional ones but in a somewhat unusual and less visible way.

Disruptive innovations tend to be risky endeavours due to high associated investments and the possibility of failure. With AI, it is not the case. While the end result for older industries is indeed revolutionary at scale, their path to success is quite well-paved. Contrary to popular belief, they are in fact most ready to benefit from AI and machine learning in short-term.

Gentle revolution

Traditional industries have the luxury of being able to improve things, before or alongside, rushing into the innovation madness of reinventing everything from scratch. Many older industries have stiff, repetitive processes that will not be changed much even at this next industrial leap. These are physical laws that guide how steel, glass or plastic is produced; retail goods continue being sold and transported at the requested quantities; and banking loans are being issued.

And that is where the businesses should look first: the cornerstone processes that have been here for decades, even centuries, and continue to be. AI can help improve those – without disrupting – through better and more precise decision-making, automatically delivered based on the analysis of the past data. In short, this is optimisation.

Traditional industries possess a wealth of historical data due to the established operational processes already in place. Presence of this historical in data is a key element in AI-based optimisation. And so, despite perceptions suggesting that older industries are relatively low on the “digital transformation” scale, the sheer volumes of historical data at hand actually puts these industries in an advantageous position when it comes to machine learning and AI projects.

>See also: 4 industries that will be transformed by machine learning in 2017

Even those verticals that have long adhered to continuous improvements – such as industrial manufacturing with lean and six sigma methodologies – possess the tremendous potential to squeeze additional efficiency with AI.

Especially given that they have tried and tested all of the previously available improvement options. While traditional methods of optimisation would require the introduction of new equipment or technologies, often leading to significant expenditure with no or little guaranteed return on investment for several years, with machine learning companies are able to increase efficiency in just a matter of months and with no capital investment.

How does this happen? Such applications are far more subdued than the grand expectations of tomorrow and instead of reinventing existing processes, will seek to enhance their precision and efficiency.

Where to look

When AI is used for optimisation, it is crucial to recognise how it delivers value – through dealing with uncertainty better than all previous means. As such, the companies should look at those processes that are most complex and involve a lot of factors that influence the outcome and may not always be known with precision.

For example, those may be predicting demand in a retail chain or managing an operation of an important machine at the shop floor that is heavily influenced by the fluctuations in raw material quality.

>See also: Machine learning: The saviour of cyber security?

If many parameters are unknown, the decisions made are often far from optimal. They can be based on generalised statistics, or expert’s opinion, but nothing can be as efficient as a machine learning algorithm that can analyse all the past data about the process and its factors, and recognise the patterns unnoticeable for a human eye and neglected by previous statistical approaches. Through this ability to deliver more precise predictions and recommend the best options for routine, repetitive processes, AI brings tangible improvements.

Another important factor is the scale of such effect. When choosing the right process, companies should keep in mind the effect on the bottom line that a specific improvement may have.

It makes sense to choose those processes that will bring a change of significant magnitude even iteratively improved by few per cents. For example, in retail this may be a replenishment process, which, if not precise enough, leads to significant overstock and under-stock losses.

For businesses that hire thousands of employees per year, AI may help in the form of candidate screening, where it could help to dodge recruiting costs. On production lines, AI could be real-time tuning of process parameters to improve output at the same cost, and so on.

>See also: Machine learning firmly integrated in the insurance industry

The bottom line

It is almost guaranteed that machine learning and AI will revolutionise day-to-day lives. Self-driving cars and AI assistants will eventually become the norm as society embraces machine learning advancements and move with fervent strides towards complete automation. But as we await this revolution – one bound to take several decades and several billion in financial investment to completely reveal its full potential – businesses should focus on another AI applications which can come to benefit in a matter of months.

Through its ability to improve what is already there, machine learning will make its impact most felt on the bottom line. While futuristic and almost “outlandish” use cases can be enticing, improving production levels or demand forecast with real-life machine learning applications is what is most beneficial today.

 

Sourced by Jane Zavalishina, CEO of Yandex Data Factory

 

The UK’s largest conference for tech leadership, TechLeaders Summit, returns on 14 September with 40+ top execs signed up to speak about the challenges and opportunities surrounding the most disruptive innovations facing the enterprise today. Secure your place at this prestigious summit byregistering here

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...