Despite the huge hype surrounding machine learning and AI, many organisations have not yet put these concepts into practice. Information Age sat down with Dave Webber, Director of concept management at Callcredit, one of the UK’s biggest credit reference agencies, to talk about how they are an exception to this trend and how they’ve overcome barriers leading to AI and machine learning success.

According to Gartner’s 2018 CIO Agenda Survey, meaningful artificial intelligence deployments are just now beginning to take place. The study shows that only 4% of CIOs have implemented AI, while a further 46% have developed plans to do so.

Despite the modest adoption rates, in 2018 many businesses are already starting to capitalise on this tech. Callcredit is one of these organisations.

Speaking at Microsoft’s event, Leading Transformation with AI, last month, Webber took me through their AI journey and how their agility helped them succeed.

“As Director of Concept Management I run an innovation function which fundamentally develops growth opportunities for the business.”

>See also: How can CIOs help corporate digital transformation?

“We undertake research and development to find out what has potential.”

“AI has effectively given us the tools to take things to a new level. We can do things better and quicker than before, whether it’s in the area of credit risk, or in the area of fraud risk.”

Callcredit is now using Microsoft’s Azure Machine Learning to identify criminals who pretend to be other people when they try to access credit reports and borrow money.

The service has been successful in stopping fraudulent access to Callcredit’s credit reporting and scoring service, named Noddle, and protecting consumers from subsequently having bad loans taken out in their name.

Overcoming roadblocks to adopt AI

As Gartner’s study shows, many organisations have been slow on the uptake of AI. They cite security and privacy concerns, the complexity of integrating AI, finding the skills and determining how to best to measure use cases as the main reason holding them back.

According to Webber: “We didn’t see many challenges around security given how we are used to dealing with consumer data and have been doing so for years. Security has always been top of mind for us and any concerns about privacy were around doing things in the cloud not so much AI.”

>See also: Data, trust and ethics – 2018’s key growth drivers

“The biggest challenges were around transparency. We needed to have a model that you could explain to a regulator. We needed also to explain the concept to our customers so they could explain to their consumer how decisions would be made about them.”

Importance of agility in digital transformation

According to Webber, at the heart of their ability to overcome roadblocks in adopting AI and machine learning was their innovation function within the organisation. This function gives us the remit to experiment with new technology and new data sources.

According to a study by PwC, 67% of business executives see the potential of AI to automate processes and increase efficiency, despite this many organisation are struggling to really understand clearly how it could work specifically with their operations.

>See also: ‘Company culture biggest factor holding back digital execution’

Webber said: “Going back in time, we had the luxury to undertake a two-year study into how machine learning and AI could be used in our organisation to make things better. We mapped a number of different problem domains which we faced, by doing this we came to understand how these concepts could work and which models of deployment would work best.”

Related: How to put machine learning models into production

Filling the AI skills gap

One of the major problems most organisations face adopting AI is finding qualified staff to help. According to analysis from the job site Indeed, there are at least twice as many jobs in artificial intelligence as there are suitable applicants. The report states that the number of roles in AI has risen by 485% in the UK since 2014, but the digital skills gap continues to hold back this innovation.

Webber said: “Finding the right kind of people is hard, we need to work with the brightest data analysts who have a real passion for technology, who can put things into context and make that leap in processing data for AI and business benefits.”

>See also: How can the UK tackle the AI skills gap? 

“We work quite closely with Leeds University and the Consumer Data Research Centre (CDRC) there. We take on interns and sponsor graduates.”

“Beyond that, we use traditional means to recruit the right people to our organisation and then train them.”

Making AI transparent to the whole organisation

Despite the skills gap in AI, advances in this sort of technology have historically been associated with a reduction in staff headcount. While reducing labour costs is attractive to business executives, it is likely to create resistance from those whose jobs appear to be at risk. Webber argued that companies need to think carefully about this.

Webber added: “There were a lot of myths in the early stages, most of it before my time, but we had to go through an education plan starting at the top, setting an expectation of what this technology could bring to the organisation.”

“It’s not just about educating people about the tech and getting them up and running; it’s about what happens when you scale it out, who else needs to get involved. It needs to stretch across the entire organisation.”

“We run a community session whereby we get all interested parties, mainly data analysts, together in a room and bring in experts from around the UK as well as clients to talk through these concepts.”

>See also: Robots could replace 250,000 public sector workers by 2030

“So we are in a phase of continual education. We look to engage all key stakeholders, getting them involved, getting their feedback and input into the project.

“The thing with this system is that it helps people realise the benefits, there is then a natural appetite for workers to then embrace new tech, as they understand it will make their lives easier.”

“For AI and machine learning success, it is vital that you do not treat it as an isolated project, treat it as a programme of work, treat it as a real opportunity for business change. Otherwise, the project will come and go, and that’s the end of it. If you treat it as a real opportunity for business change, you must accept that the challenges will be difficult along the way and you should accept that some projects will fail. If you go ahead with the right mentality, then you’ll have an easier ride.”