The COVID-19 pandemic has reemphasised the power of AI as companies now use it to predict trends in real time, personalise customer experiences, and even explore coronavirus treatments and vaccines. But the crisis has also exposed the limits of how AI is traditionally used in business. Models needed to be retrained to continue working during a period of dramatic change.
As the pandemic propels businesses into a digital-first world, AI will become a key driver of corporate growth and competitiveness, with the growth in European companies’ annual spend on the technology set to double in the next three years – from 4.44% last year to 8.83% by 2024. And as digital leaders know, AI is not a silver bullet or a one-size-fits-all solution. AI solutions can fail to deliver if the wrong business case is selected, the right data is not identified, the data is prepared incorrectly, or the model is not built for scale. That is why AI results vary so much today.
To help executives drive ROI from AI, ESI ThoughtLab, together with a group of AI leaders, including Cognizant, conducted a worldwide benchmarking study among 1,200 organisations across industries. The benchmark revealed that almost two-thirds of senior executives see AI as highly important to the future of their businesses, rising to three-quarters (73%) in the UK. Recognition of AI’s future importance is nearly unanimous, noted by 98% of AI leaders and 85% of the world’s largest organisations (with revenue over $20 billion). Below are some additional noteworthy results:
Companies globally are continuing to struggle to maximise the financial benefits of AI, with two-fifths (40%) of AI projects losing money or just breaking even. Even among those that succeed, the average ROI is currently just 1.3%, and while AI can offer significant ROI, results do not come overnight. The study found that there are often overly optimistic expectations of fast rewards when it in fact takes time to identify the appropriate business case, acquire and prepare the right data, and then build, test, refine, and deploy working models. Payback periods of six months are the exception, not the rule. The average payback period for AI investments is around 17 months.
“After COVID, firms are not going back to the old, unproductive ways of working,” says Bret Greenstein, SVP, Global Head of AI and Analytics, Cognizant. “Now that they are forecasting daily, they will not return to monthly. Executives realise that real-time access to data and the use of AI and ML for forecasting and planning makes you that much more efficient and resilient. Once the bar has been raised on efficiency, companies will keep going with fewer people to preserve margins and stay competitive.”