The Gap Nobody Is Pricing In
Ademola Odewade has a warning for finance leaders that has nothing to do with whether artificial intelligence works. It works. His concern is what happens in the space between what AI can now do and what most finance functions have actually built to govern it. The chartered accountant, founder of K-Dems Consulting, and contracted expert evaluator for a frontier AI evaluation marketplace argues that this gap, not the technology itself, is where the real risk sits over the next two years.
“AI capability is now well ahead of AI governance in most finance functions,” Odewade has said. “The tools are cheap and available. The frameworks for using them safely are not.” It is a distinction that sounds obvious once stated but is, in his experience, routinely ignored by organizations under pressure to show a board that they have an AI strategy.
Two Bad Options
In Odewade’s view, most finance leaders currently find themselves stuck between two flawed responses to AI. The first is paralysis. Having watched the hype cycle unfold and been pitched by a rotating cast of vendors, many leaders conclude that the risk of a control failure, a compliance breach, or a bad number reaching the board outweighs any efficiency gain on offer. So they wait, while their cost base stays exactly where it is and competitors who move do not.
The second response is recklessness. A language model gets plugged into a process nobody has properly mapped, a human review step that existed for a specific reason gets quietly removed, and six months later the organization discovers it has automated an error rather than eliminated one. In finance, Odewade points out, that is rarely a minor inconvenience. It can mean a restatement, a failed audit, or a funding round that falls apart under diligence.
Both failures, he argues, share the same root cause: treating AI adoption as a technology decision rather than a leadership one.
Evidence From Inside the Labs
What separates Odewade from many others making similar observations is where his evidence comes from. Since July 2026, he has worked continuously as a contracted expert evaluator for a frontier AI evaluation marketplace, grading the output of advanced AI systems on finance and business reasoning and writing the rationales that explain precisely why a wrong answer is wrong. He has been promoted across three successive engagements in three months and was certified in AI agent evaluation in September 2026.
That work puts him in a rare position. He is automating professional work with AI agents inside his own consulting business while simultaneously being paid by AI labs to identify exactly where those same systems get professional judgement wrong. Few people writing or speaking publicly about AI governance in finance can claim that kind of direct, ongoing exposure to both sides of the equation.
Built on Big Four Discipline
Odewade’s conviction that governance has to come before automation traces back to his early career. He spent eight years in professional practice, first at PwC and later at KPMG, immersed in the kind of formal controls, audit rigor, and governance structures that large firms build almost by default. When he left that environment to found K-Dems Consulting in London in 2025, his goal was to bring that same discipline to small and mid-sized businesses that cannot afford to build it internally, often operating with a finance team of one or two people and no dedicated audit function at all.
That is precisely the population most exposed to the governance gap he now writes about. Large enterprises have compliance teams and risk committees to slow down reckless AI adoption. Founder-led and mid-market businesses frequently do not, which means the decision about where AI belongs in the finance function often falls to a single overstretched leader with no internal check on the call.
A Framework for Closing the Gap
Odewade’s response to this problem is ACCOUNT, a framework laid out in his first book, Human-Led, AI-Powered Finance: The CFO and Founder’s Playbook for Smarter Decisions, Stronger Controls and Scalable Growth. Rather than treating AI adoption as an all-or-nothing choice, ACCOUNT gives finance leaders a structured way to decide what work can safely move to a machine, what must stay in human hands, and where controls need to sit so that an error gets caught before it reaches a board, an auditor, or a regulator.
His second book, The AI-Powered Future of Work, publishing October 2026, extends the same governance-first thinking to individual professionals through a companion framework called EXPERT, addressing how people, not just organizations, can build verifiable judgement into their own careers as AI reshapes what employers value.
Why the Timing Matters
Odewade is direct about why he considers the next two years decisive. AI tools are only becoming cheaper and more accessible, which means the organizations that fail to build governance now are not simply delaying a decision, they are widening the exposure they will eventually have to address. His hope is that finance leaders start closing that gap deliberately, rather than discovering it the way most organizations discover control failures: after something has already gone wrong.
He also hosts two podcasts, The King Dems Podcast and Diary of a CFO, both ranked in the global top 2 percent on ListenNotes, where these questions of governance, leadership, and AI adoption come up regularly with senior guests including Professor Dave Ulrich of the University of Michigan’s Ross School of Business. Across more than 200 combined episodes, the two shows have reached listeners in over 130 countries.
For finance leaders under pressure to show a board an AI plan, Odewade’s advice is less about speed and more about sequence: build the governance first, and let automation follow it, not the other way around.
Follow Ademola Odewade on LinkedIn and X for more on his work at the intersection of AI and finance leadership.
