AI on Old Processes: The $1 Billion Mistake Companies Keep Making

Jan de Vries · 2026-09-16

AI is moving into boardrooms faster than most companies can figure out what to do with it. But according to Jim Harris, a specialist in GenAI and disruptive innovation, the bigger risk isn't moving too slowly. It's bolting new technology onto old ways of working and expecting magic to happen. Harris has advised executive teams at American Express, Barclays, IBM, SAP, Walmart, and the UK Cabinet Office. In 2024, TEC Canada named him Speaker of the Year for his work on AI. In this exclusive interview with the AI Speakers Agency, he explains why companies need to redesign processes from scratch, how leaders can ease employees' fear of AI, and why keeping humans in the loop still matters as adoption accelerates. ### Why 'AI Fairy Dust' Doesn't Work Harris shared a story from the Consumer Electronics Show in Las Vegas. He interviewed the head of AI for Ernst & Young in the US. EY has 400,000 employees and has spent a billion dollars on AI. They had a 46-hour risk management process they'd run for clients for years. They thought, "Let's apply some AI to it." With the people who designed the process, they got it down to 32 hours. A 20% gain. Not bad. Then they tried something different. They took people with no experience in the process but who were tech-first, AI-first, and paired them with risk managers who'd never been involved. They got the process down to 15 minutes. Then they said, "We're EY. We need data compliance and verification." So they added steps back in and got it to six hours. Look at that: 46 hours, 32 hours, 15 minutes, then six. Harris says this highlights two things. First, you can't just sprinkle some AI fairy dust on your existing processes and hope for the greatest gain. You need to start with a blank sheet of paper and ask: if we were designing this today from first principles, what would we do? Don't take a 2018 process and try to enhance it with AI. Redesign it. Second, the people who designed the original process are wedded to it, consciously or not. You need fresh eyes. When you go from 46 hours to six, that's more than an 80% reduction. Staggering. And now you can redeploy people to more important activities. ### How to Introduce AI Without Scaring Your Team Harris says the first step is creating psychological safety. Most people in developed nations fear AI. They fear job loss. As adults, we're often uncomfortable with new ways of working. So organizations have to work hard to say three things: - Your job is safe. - We'll provide training. - This will enhance your skill set and employability. That's the foundation. ### The Human Element Still Matters Harris also points to Demis Hassabis, founder of Google DeepMind and Nobel Prize winner in 2024. Hassabis started as a PhD student and went on to lead breakthroughs in AI. But even with all that brainpower, Harris insists humans must stay involved. AI can process data faster than any person, but it can't replace judgment, ethics, or creativity. The companies that win will be the ones that pair AI's speed with human wisdom. > "You can't just sprinkle some AI fairy dust on your existing processes and hope to get the greatest gain." — Jim Harris So before you bolt AI onto that old workflow, ask yourself: if I were building this from scratch today, what would it look like? The answer might surprise you.