"If leaders don’t understand AI, they can’t lead...": Accenture CEO Julie Sweet wants execs to ‘touch the keyboard’

Corporate leaders talking up AI without understanding how it works is no longer good enough. This is the core argument made by Julie Sweet, chair and CEO of Accenture, who said companies will only see real returns from AI if learning starts at the top of the structure.Speaking to Bloomberg Television at the World Economic Forum in Davos, Sweet said leaders must first build their own understanding of AI before expecting their workforce to adapt. Without that, companies will struggle to change how they operate or what they offer, she believes.“If leaders don’t understand AI, they can’t lead the company through the changes,” Sweet said. “In three years, you should be able to say, ‘My company has different services and has different insights.’ That requires a depth of learning from leaders first, and then you have to bring everybody along the way. So leader-led learning is absolutely critical.”Sweet put AI adoption as a test that will play out over a short horizon. Within three years, she said, companies should be able to point to concrete differences in services and insights that are enabled by AI. If they cannot, that signals a failure of leadership rather than technology. Her perspective shows how Accenture has approached AI internally since late 2022, following the release of widely used generative AI tools. Instead of starting with mass employee training, the company focused first on its senior leadership. According to Sweet, most of the early training efforts were directed at Accenture’s top executives, with the aim of ensuring they understood the tools well enough to make informed decisions.That approach, she said, is not theoretical. Asked by Bloomberg’s Francine Lacqua for a practical example that could apply to workforces or even citizens, Sweet described what one client had learned the hard way.“One of my clients told me that, until they had their 300 leaders touch keyboards and see what AI could do, they couldn’t get moving,” she said. “That’s a very tangible example.”Accenture followed a similar path, she added. After November 2022, the company prioritised hands on learning for its top 50 leaders. The goal was not to turn executives into engineers, but to ensure they understood what AI could and could not do, and how it might change business processes.“That’s what we mean by leader-led learning,” Sweet said. “And once you have that, it unlocks the possibilities of what it really means that it’s going to change everything.”Sweet also extended the argument beyond companies. In regulated industries, she said, the pace of AI adoption will depend in part on whether regulators themselves understand the technology. A lack of understanding, she warned, could translate into restrictions that slow or block deployment.“It’s also critical for regulators and regulated industries,” she said. “If regulators block AI, they won’t be able to scale or succeed.”As a result, Sweet argued that governments and non-profits face the same challenge as companies. Leaders in these sectors must invest in learning first, and then design strategies to train their organisations more broadly.Many organisations are rolling out AI tools to employees while senior decision-makers remain removed from AI's day-to-day use. Sweet says that this gap creates a risk of superficial adoption, where AI is added on rather than integrated into how work is done.For Sweet, leaders must use the tools themselves, understand their limits and implications, and only then guide others through the transition. Without that foundation, she said, AI strategies are likely to stall.So leaders must sit down, use the technology themselves, and learn before asking others to follow.