Tech commentators have been repeating the mantra “data is the new oil” for years. But with new challenges arising across industries, it’s becoming clear that this aphorism doesn’t give the full story. 45% of business leaders in 2019 considered generating and using insight from data effectively as their biggest challenge for the year.¹

Data is no longer the new oil; knowing how to use the data is. Data itself is useless without the insight it gives an organisation. 

At Nimbus Ninety’s “Accelerate your Journey to AI” evening, held in partnership with IBM, members gathered to deep-dive into implementing AI into data and cloud strategy. AI holds so much promise for liberating the world’s information, but executing a fully-fledged AI strategy requires a critically different approach to data and cloud. As IBM’s Director of Cloud Software, Daniel Wilks, asserted, “data underpins digital transformation; AI is crucial to driving this forward.”

CASE STUDY #1: WIMBLEDON & AI INNOVATION

Bill Jinks, IT Director at Wimbledon, described the “new era of competition” that technology-driven opportunity brings. With online streaming of sports on the increase and sports data analysis becoming faster and more detailed, the projected spend in sports IT services in 2025 is $3bn. Meanwhile, there has been a 53% increase from 2018 to 2019 in the number of minutes of people streaming sport. 

For Wimbledon, responding to this shift towards digital consumption of sport is about rapid scale. For this, it needs to be exploiting cloud capability and driving AI innovation within that. Wimbledon’s hybrid cloud strategy is made up of a combination of geographically dispersed public clouds and on-premise data management private clouds - alongside the courtside networks onsite that capture the action and analyse it.

Part of Wimbledon’s customer proposition is to bring the highlights of any match instantly to a fan, whether they are using iOS, Android or something else. Here, Wimbledon uses IBM Watson to draw out AI-generated highlights reels, using sound (a ball being hit; the crowd cheering) and audiovisual analysis to identify what is a “highlight”.

Even here, videographers face AI bias. If it was left to its own devices, Watson would just produce reels featuring Federer, Nadal and Djokovic: their crowds are bigger and cheer louder, so AI audiovisual analysis is skewed. Feeding highlights from other courts is needed to correct the bias, and teach it to nuance its pickings.

CASE STUDY #2: MET OFFICE & WEATHER DATA FOR BUSINESS

With a long history of accumulating huge amounts of data and forecasting from it, the Met Office has changed hugely in the decades it has been operating. From its naval and military origins in 1854, its product (insight drawn from weather data to predict the forecast) has not changed but with the shift to digital and innovation in AI, the process and priorities certainly have.

Understanding the value of the data is of utmost importance to the Met Office, both in terms of observational data (fact) and forecast data (fiction, as it were). Historically, the Met Office produced thousands of bespoke data packages for other organisations. These were essentially products, which required product management, product ownership and a product lifecycle. As tools for data storage and data analysis evolve, this strategy becomes inefficient. Instead, the organisation now wants to move away from managing products and move towards managing data sets.

This is where APIs have great strategic value to data processes. If APIs are properly leveraged, customers can be given the option to benefit from the data insights, without dealing with the data itself. This streamlines the process; it exposes the insight without exposing the data. 

An example of this is tyre pressure, which decreases by 1 PSI for every degree Celsius dropped in temperature. For the Met Office, this is a commercial forecasting opportunity. It could install alerts in vehicles that tell you a prediction for your tyre pressure, and that you need to pump it up - without once exposing the climate data itself. 

ETHICS, PRODUCTIVITY AND CUSTOMER EXPERIENCE

A huge ongoing issue, which perpetrates all conversations around AI capability, is the ethics around AI. No matter what businesses can achieve with AI innovation and data, retaining customers’ trust is central to the conversation if businesses are to succeed at all. 

As mentioned above, AI bias needs monitoring and correction. With the example of Wimbledon’s highlights reel, the ethical consequence is minor; but with technologies such as facial recognition or location data being used to forecast criminal activity, the ethical consequence can be substantial and damaging.

According to EY, leading practices for establishing a trusted AI ecosystem include:

• An AI ethics board
• Established AI ethical design standards
• An AI inventory and impact assessment
• Validation tools
• Awareness training
• Independent audits

It’s well established that AI can only ever be as good as the data that is being put into it. Other issues like trust in AI, executive sponsorship, changing skills demands and needing people transformation alongside the digital, complicate the conversation around AI. With its huge potential to develop business capability, AI should be invested in and implemented to business needs but renewing data and cloud strategy must come first. 

It’s certainly not simple, but the opportunity is huge. 

This event held was in partnership with IBM, an IT solutions and cloud provider.

REFERENCES:

1. Nimbus Ninety. Digital Trends Report 2019.