Machine learning and AI are the latest technologies transforming business — carrying on from digitization and big data. It is difficult, however, to pull apart the hype from implementation and distinguish the realities from the marketing jargon.

Pete Williams is a “data evangelist” who works with companies to get them started down the road of data-driven decision-making. For Pete, companies fail to understand that AI and ML are use cases for the changes made in data collection over the last decade — not something completely new. Also the founder and CEO of The Decision Playbook, Pete believes the whole process hinges on the volume of interactions now taking place on machines and devices that leave a digital trail.

The changes that have occurred are in gradations of sophistication in the processing capability to handle these high volumes of data. However, the true transformation, according to Pete, is understanding that the potential of data is not in algorithms or technology — it is in how we structure our decision making around the technology that already exists.’s Robin Block sat down with Pete to discuss the challenges businesses face in implementing machine learning and the future of a data-driven society.

Robin: What are the developments you see in the industry — who’s doing things right?

Pete: Almost all companies are now storing their data. But, using that data in successful models to train algorithms isn’t something I have yet seen many come to terms with. Most of the developments I have seen come from people using open source languages rather than tools on their vast datasets. Quite often, what is actually being done in corporate applications of A.I. are large pattern matching operations.

I call myself a ‘data evangelist’ because I spend my time helping businesses and stakeholders understand that there is a different way to make a decision than what they have done in the past. It is hard for people who have grown up in an organisation and learned how that culture works, to take a different perspective and grasp the implications of new technology on old roles and processes. You could say that my job is to help companies get to base camp on the mountain of data discovery safely. If you’ve hitched a ride on the vendor helicopter and been dropped just below the summit, without the necessary training and conditioning to finish the climb, you’re as likely to fall back down the slope as make it to the peak.

What mistakes do companies make in engaging with data?  

Mistakes often come from treating data as a technology problem. Using data to aid commercial growth and decision making is not a process that should fall under the auspice of the CIO. The technology function of an organisation is already busy making sure that technology accurately captures and relays data. The use of data as a decisioning tool is a commercial function — not a technology function.

The next problem is that people put all of their hopes in hiring great data scientists — another variant of the helicopter up the slope! You need talented people who can manipulate data, but that won’t get you anything unless you build an organisation and structure that appreciates and values data.

The basecamp requirement is to create what I call the ‘data literate ecosystem’ — an organisation that puts data architecture at its heart, surrounds those components with a culture that generates a flow of strategic inputs and is capable of acting on that data, once interrogated, to drive commercial outputs and decisions. The most sophisticated data centred operations can be let down by the humans wrapped around the outside. Getting this right is often the hardest and most intractable part of creating a truly data-driven business. It is easy enough to set up a data lake with some visualisation tools — everyone wants to sell you one of those. But, you are just wasting money if you don’t tackle the human problem.

Flip the question. Succeeding with data is ultimately a people, not a technology opportunity.

What do you see in the future for A.I.?

When I first started programming, producing a weekly data set was good work. That has moved to daily, to within-day, to almost real-time. Moving your decision-making process into real-time requires you to see the most relevant and granular information imaginable — the challenge becomes coming to grips with the speed of life. It can be quite scary for people making decisions when it feels like they are always on the spot right now. This is where A.I. can help, by making some decisions for us. Much of the challenge moving forward is therefore not technological development but proper implementation and cultural adaptation to trust the algorithm.

The realisation that can enable truly momentous transformation is that data’s power is in deep connectivity across all business functions. Data illuminates the already existing reality that there are no isolated verticals — every single decision is a horizontal decision. Organisations are simply organised into vertical pillars for our own convenience and limited cognitive capabilities. Data is a pathway to operating in a less abstracted manner more attuned to the reality of the processes already going on within an organisation.

There are already a lot of people interacting with A.I. without realising it. On one hand, the most exciting possibility for the A.I. revolution is the potential for a transition that people don’t really even notice — services that were done by humans will be taken over by machines, everything will continue as before and everyone will benefit from increased productivity. The dark side, however, is that we are entering a new world in which automation isn’t just coming after manufacturing jobs — ‘professional jobs’ are on the road to automation as well. We have never had to survive a transition of redundancy and change like this before. Many of the jobs that people aspire to get coming out of higher education aren’t going to exist in the future. In the long-run, this may completely transform the concept of society — an exciting or worrying observation depending on which side of the bed you woke up on this morning.

Like this article? Subscribe to our weekly newsletter to never miss out!

Previous post

The Role of Big Data and Mobile Apps in Healthcare

Next post

How GDPR Will Affect Data Science