From customer service analytics to risk modelling, the potential business applications of artificial intelligence are wide ranging. But according to new research by management consultancy McKinsey, only a minority of companies are drawing business value from AI adoption. And Covid-19 is widening the divide between the AI leaders and laggards, it suggests, as the leaders double-down on their investments.
McKinsey’s analysis of a survey covering more than 2,000 companies indicates that AI adoption can drive profitability. Companies that derive 20% or more of their pre-tax profit from AI – the so-called AI leaders – are nearly twice as likely to have reported profit growth of over 10% in 2019, according to The State of AI in 2020 report.
But these AI leaders represent fewer than 10% of the companies surveyed. Fewer than a quarter (22%) derive 5% or more of their pre-tax profit from AI.
This partly reflects limited adoption: 50% of respondent organisations have not used AI for a single project. But it also suggests that many AI adopters have so far struggled to extract significant value from the technology.
This chimes with the results of a 2019 survey by MIT Sloan Management Review and Boston Consulting Group, in which 70% of companies reported little or no impact from their AI initiatives.
AI adopters were more likely to report revenue gains as a result of their implementations rather than cost reductions. Three out of ten respondents that had applied AI in any function said revenue had increased by 5% or more as result; half as many said costs had reduced by the same margin.
Covid-19 is widening the gap between the AI leaders and laggards, the study also shows. The majority of AI leaders have increased their investment in technology in every major business function over the past six months, while only 30% of other businesses have done the same.
This growing divide between companies that have boosted technology investments during the pandemic and those that have not is not limited to AI. A global survey of CIOs by recruitment consultancy Harvey Nash and Big Four company KPMG similarly found that 47% of companies are accelerating their investments in digital transformation.
“We are likely to see a divide grow, with digital leaders better positioned to pivot and scale into new opportunities, leaving behind organisations that resist, or are unable to invest in their innovation journey,” says Harvey Nash CEO Bev White.
McKinsey argues that what distinguishes high performers is that their AI strategies are led from the top. “It is really the business leaders in charge of scaling these techniques [who] need to be convinced to do it,” says Nicolaus Henke, senior partner at McKinsey and chairman of its advanced analytics and AI company QuantumBlack. “It is a fundamentally different way to run the entire thing.
“[Leaders] constantly have many projects going on and constantly make improvements,” says Henke. “That requires a completely different way of using technology and there’s essentially a big software programming team at the core of it, which only the leading companies have; while lagging companies still have a lot of little projects in lots of different corners.”
Lagging companies looking to bridge the gap need to gradually overhaul their strategy, starting with the areas where the largest gains are to be made, adds Henke. “For companies who are on the journey not starting as a tech company, they need to start somewhere and ideally in high-value areas,” he says. “[They need to] learn how it works and then fund additional investments from the improvements they’ve already made.”
Despite the increasing prevalence of AI across business functions, only a minority of companies recognise most of the risks associated with the technology, according to McKinsey’s research. The share of respondents flagging risks as relevant has not increased on last year’s survey despite accelerated adoption, with cybersecurity still the only risk that a majority of organisations consider relevant.
Concerningly, even fewer are actively working to address the threat in every case. Even when it comes to the most addressed risks, cybersecurity and regulatory compliance, only 51% and 38% of respondents respectively stated that they were working to mitigate the threat. Unsurprisingly, in most instances, AI high-performers remain more likely to recognise and act to address risks.
The lack of risk awareness and action is concerning, particularly when it comes to customer data and the inherent biases in many algorithms, says Henke, adding that companies need to take action even without the threat of regulatory scrutiny.
“[We are] in the very early stages of a very fundamental shift with AI – think of it like the steam to electricity transition,” he says. “We tell our clients to behave as though you are regulated; don’t wait for a regulator because, particularly if you’re a good practitioner in AI, it’s likely that you know more about the risks than the regulator.”
Featured photo by Zapp2Photo/Shutterstock.
Amy Borrett was the resident data journalist at Tech Monitor.