Artificial intelligence might be the buzzword of 2018, but few actually know what to do with the technology. That said, Sky seems to be surging ahead of the pack.

At the Telco Data Analytics and AI conference in London, an interesting statistic was put to the audience; 60% of the AI R&D spend in the telco industry is being directed towards network optimization. This is certainly a valid quest, though the problem with inward R&D investment is that it won’t prevent the slow wander towards utilitisation. To create value, telcos need to be investing in projects which actually create value, drive diversification and capitalise on new revenues. This is exactly what Sky seems to be doing.

“We have a data liquidity problem,” said Rob McLaughlin, Head of Digital Decisioning and Analytics for Sky UK. “Getting data is not an issue, we get it without trying, it’s about getting value from it.”

It seems the Sky UK team has a lot of ‘nice to have problems’, which demonstrate the effective steps forward the business is making in the intelligence-orientated world. While many telcos are struggling with the basic concepts, Sky is really setting the pace.

Aside from the overwhelming amount of data, McLaughlin suggested there was almost too much enthusiasm. Here, the wider team isn’t resisting, but asking for solutions which are perhaps not ready. McLaughlin said there is a lot of low-hanging fruit, the simple problems which AI can address, right now, but the more complex applications will take time.

As McLaughlin told the audience, this drive can be a bit frustrating, but at least the management team is embracing new concepts and technologies. This is arguably a perfect scenario however. Change is led from the top of an organization, and McLaughlin seems to be describing a culture which is enthusiastic to embrace change and create value. This is not normal for the telco industry.

Another interesting point made by McLaughlin was a claim there was no POC.

“We launched these projects at scale from day one,” said McLaughlin. “We didn’t want to do a POC as it was a bit of an insult to our intelligence. Why do they need to test whether data is good for the business?”

This demonstrates the much-hyped fail fast business model which has been employed so effectively by the internet giants. These companies don’t need to prove there is value in personalising services, they just need to make it work. The only way to get the algorithms to work is to get them out in the real world, trained by data, honed by machine learning and real-time experiences. This culture of creating results, not trying to prove perfection, will certainly drive value for Sky.

McLaughlin’s team are implementing AI in four different ways at Sky. Firstly, using customer information to cross sell services and products. Secondly, increasing engagement with products and services customers have already bought. Third, anticipating customer needs and problems, a project which is saving Sky millions in customer services and improving the overall NPS score. Finally, AI is being used in media optimisation to improve the advertising platform.

While these projects are still in the early days, the results are clear according to McLaughlin. NPS has been improving, cost saving are being realised and proactive selling of product through personalisation is increasing. With the cross-selling side, the results are quite remarkable. The success of sales of Sky Sport products are up 57% due to two simple changes. Firstly, putting the product in front of the customer at the right time, Saturday afternoon not Friday night for example, and Secondly, selling the product in the right way. If you know you are engaging a football fan, tell them about the football benefits not Formula One.

All of these initiatives are built on identity. For McLaughlin this is the most important aspect of any data analytics and AI programme, and receives more attention than anything else. If you cannot identify your customer, it is impossible to personalise services effectively. It seems simple, but it is an aspect which is often overlooked.

“If we have the opportunity to speak to someone, don’t tell them something, treat them as the person the data says they are,” said McLaughlin.

Sky might not have a reputation as an particularly innovative organization, nothing out of the ordinary at least, but this approach to data analytics and artificial intelligence is certainly worth noting. The culture is accepting and proactive, there is an attitude which is geared toward doing, not planning, and the objectives are clearly outlined. McLaughlin might have his frustrations, but if you want an example of an organization which is proving the value of intelligence, you won’t have to look much farther.