Could AI be the fastest way to make a broken process more expensive? In this episode of Consulting The Future, I’m joined by Bill Briggs, CTO at Deloitte, to unpack why so many organizations can produce impressive AI pilots, yet still struggle to show real business change. Bill has spent nearly three decades advising leaders across the private and public sectors, and he has a clear point of view on what separates pilot activity from outcomes you can measure, defend in the boardroom, and scale across the enterprise. We talk about what “good ROI” actually looks like when you strip away dashboards designed for show. Bill explains why early AI efforts often fall into incrementalism, where teams layer new tools on top of old workflows, then wonder why productivity does not move. He also shares a line that will land with anyone watching cloud bills rise, you can end up “weaponizing inefficiency” and turning yesterday’s background automation into today’s token burn. A big thread in our conversation is the “innovation flywheel,” and Bill translates that into plain English. He argues that innovation cannot live in a silo with nicer office space and a few foosball tables, it has to come from the people closest to the work who know where the friction and “knuckleheadery” hides. He shares Deloitte research that shows a dramatic trust gap, with enthusiasm for AI high in the C-suite, but dropping sharply as you move toward frontline employees. The takeaway is simple, if your people do not feel included, they will resist it, and if they feel threatened, they will quietly avoid it. We also get into what it takes to move fast without losing control as GenAI and agentic systems spread. Bill points to a pattern he sees repeatedly, organizations invest heavily in technology, but underinvest in the parts that make it safe and sustainable, training, guardrails, process controls, and the engineering discipline that bakes security and governance into how solutions are built. That leads into a practical conversation about rethinking processes from scratch, and why “waiting for the next model release” can become a form of paralysis that competitors will happily exploit. Finally, Bill offers a refreshingly honest look at technical debt, including his three-part framing of malfeasance, misfeasance, and non-feasance, and how leaders can take a more surgical approach to modernization instead of repeating tired slogans like “cloud good, mainframe bad.” We close with a human moment too, Beach Boys, pinball machines, and a reminder that the best tech conversations still sound like real conversations. So where do you think most organizations are right now, building real outcomes, or putting a shiny layer on top of yesterday’s problems, and what would you change first, and please share your thoughts?
