The advent of emergent technologies such as machine learning (ML) and artificial intelligence (AI) has sparked debate across industry and organisation alike. These technologies promise so much potential, and yet few companies can truly harness this potential and leverage it.

That’s because there are multiple variables that have to be addressed before the business can truly partner up with AI and ML.

“The business can’t really make a call on AI and ML until it has the right culture for it,” says Anton Herbst, Head of Strategy, Tarsus Technology Group. “Most of us are drowning in data and there is just so much more entering the business from multiple platforms, increasing exponentially, that it has become overwhelming.”

For most businesses, data has been swirling into pools and lakes of information, impossible to analyse and use to business benefit without the right systems in place. So much of that data is gathered without context, without considering where it is coming from or how it is relevant to the business, that it remains untapped and too complex to leverage for the average organisation.

Pattern Recognition

Humans are incapable of finding patterns in these oceans of data as there is too much for them to assimilate and understand. This is where ML and AI step in.

The neural networks created by these learning and evolving technologies are capable of sifting through massive amounts of data to glean both insight and context, in seconds.

Supervised vs Unsupervised AI, ML

“There are two types of learning that accompany the introduction of AI and ML to the business,” says Herbst. “The one is supervised learning where we feed the systems the information they need until they understand a basic concept or idea, the other is unsupervised learning where the machines are left to dig through the data and find patterns through experience and understanding.

“From a business perspective, this means that, to fully eke out any value from data, there needs to be a very clear understanding of the question.”

The challenge isn’t how to sift through the data or perhaps what data to sift through, but to establish exactly what questions the business needs answered by the data.

This will immediately refine every other step that follows and ultimately determine whether or not the business gains any value from the data. Conversely, the wrong questions will only deliver the wrong answers.

The right questions should be driven by business need and strategy, and then this should be further supported by the right business culture.

Invest in a learning culture

“The organisation needs to invest into a learning culture that recognises the value of data and has a need to understand it,” says Herbst.

“This can become increasingly complex, expanding outwards from the application of business intelligence tools to describe the data better and into diagnostic analytics and predictive analytics that can predict patterns and assess probabilities to allow for more agile business decision-making.”

A business that puts machine learning at its core will allow for deep learning and the ability to pick up on patterns and look to the past to see how this could influence decision-making in the future.

This can only work if the business constantly feeds data into the system and recognises the value of this partnership to drive value and growth.

Automatic triggers

“Once the business shifts into the realm of predictive analytics it can take advantage of algorithms that are capable of predicting what will happen, and prescriptive analytics that can trigger actions based on past experiences without the need for human intervention,” adds Herbst.

“The technology can pick up the patterns from the data and trigger actions that minimise the need for human interaction while maximising business potential and reach.”

3 Foundational pillars: Interrogate, Invest, Innovation

There are three pillars that should form the foundations of any partnership with AI and ML.

The first is to interrogate the data, to ask what it is telling you and whether or not the patterns that emerge are relevant for the business. This is where it is critical to establish what questions need to be asked to ensure the longevity of the business. These can be anything from ‘Are we in the right business?’ to ‘Is there a pattern in the data that says we should change direction or focus?’.

This pillar can amplify the potential insights provided by data into customer engagement, marketing, feedback and other critical touchpoints across the organisation.

The second pillar is to invest into the right platform for the business. This can only be determined once the business has interrogated the data and established how it is relevant and what patterns need to be assessed.

“Everyone talks about machine learning and artificial intelligence as casually as if they can step in and translate the data into relevant insights in a matter of minutes,” says Herbst. “This is not how it works.

“You cannot invest into a solution until you know what patterns you need to find and what questions you need answering. Always stay focused on the problem that you are trying to solve and what data you have, or need, to solve that problem.”

Finally, the pillar that should stand alongside interrogation and investment is innovation.

True, it is a word overused by business and media, but the principles at the core of this word have never been more relevant or important than they are today.

Standing out in busy, noisy and demanding markets is challenging and complex. This makes it absolutely essential that the business invest into its people and their ability to think outside the box and to innovate.

“So much of our ability to question life and situations and solutions is educated out of us by outdated education systems,” says Herbst.
“What we want is to ignite the skills of critical thinking and problem-solving in the youth and in those already in the workplace. It is never too late to re-learn how to ask questions, to challenge the status quo, and to approach every situation with fresh and intelligent eyes.”

Remarkable Potential

Machine Learning and artificial intelligence have remarkable potential and are, in themselves, remarkable technologies.

And yet, they have to be built on solid foundations and engage in constant, relentless learning in order to deliver on this potential. They also have to be leveraged with intelligence – recognising that no machine or algorithm can interpret data correctly without the right parameters in place.

“If your questions are not intelligent, then the machines will not respond with intelligence that can be used,” concludes Herbst.