As The Scaling Executive Coach and host of The Scaling Executive Podcast, I sat down with Chris Crowe, founder and CEO of CMBYND, a global consulting firm built on an AI-enabled advisory model. Crowe spent two decades inside major financial institutions, including RBC Capital Markets, Citi, and ING, before building CMBYND from zero to roughly $100 million in revenue.
When AI starts producing the analysis your team used to produce by hand, Crowe holds that the CEO’s job shifts from managing output to managing thinking. “The vast majority of people take what is produced from AI as gospel,” Crowe told me, and CMBYND’s own internal data shows that pattern across roughly 75% of the people who use it. Left unchecked, that habit quietly shifts a company’s culture from judgment to acceptance. Crowe’s fix is structural: retrain the organization to question AI output the way it once questioned a junior analyst’s first draft, and rebuild the org chart itself around that discipline.
This episode is for CEOs scaling a knowledge-work business, consulting, financial services, or professional services, who are adopting AI faster than their team’s ability to question it.
Key Takeaways
- CMBYND found that roughly 75% of employees accept AI-generated output without question, a pattern Crowe says slows an organization’s thinking and can quietly reshape its culture.
- Crowe rebuilt CMBYND’s staffing model from a traditional consulting pyramid, heavy with junior analysts at the base, into a diamond shape that narrows entry-level headcount and routes AI-assisted research directly to client-facing knowledge workers.
- CMBYND retrains staff to prompt AI the way they would direct a new hire: stating who they are, what they need, and why, instead of typing a bare search term.
- A Canadian court held Air Canada to a commitment its chatbot made to a customer, a ruling Crowe treats as proof that legal accountability for AI-generated communication sits with the company, not the AI system.
- Crowe’s operating rule: AI will confirm you are right and tell you your question is a great one, so the discipline is to interrogate the source, the context, and what the AI left out before accepting its answer.
How CEOs Restructure Consulting Firms From a Pyramid to a Diamond for the AI Era
Traditional consulting firms, including CMBYND before Crowe rebuilt it, are staffed like a pyramid: a small group of knowledge workers at the top and a wide base of junior staff underneath doing the analysis that supports them. Crowe found that clients valued the top of that pyramid and resented paying to train the bottom of it.
AI removed the need for that wide junior base. Crowe’s team began feeding its own research, papers, and thought pieces directly into AI tools, letting knowledge workers get finished analysis without routing it through a layer of junior staff first. “We could actually build a diamond instead of a pyramid,” Crowe said, describing a structure with a narrow base and a wider middle of experienced staff supported directly by AI-assisted research.
The shift was not easy. Crowe’s team had to be retrained before the new structure worked, which is why CMBYND built an internal lab dedicated to teaching staff how to use AI as a research partner rather than a search engine.
How CEOs Train Employees to Prompt AI Like a New Hire, Not a Search Engine
CMBYND’s lab exists because Crowe found that staff defaulted to typing AI prompts the way they typed Google search terms, which produced shallow, generic answers. The fix was to teach staff to treat AI the way they would treat a new employee on day one.
“Ask them questions, tell them who they are, tell them what you’re looking for,” Crowe said, describing the instruction his team gives staff before they open an AI tool. Crowe watched his team’s prompting shift from single search terms to structured requests that stated context, role, and goal, the same information a manager would give a new hire before assigning a task.
That retraining is what made the diamond structure work. Without it, the firm’s shift to AI-assisted research would have produced faster answers without better ones.
How CEOs Build Accountability When AI Makes Mistakes With Customers
CMBYND’s clients are mostly large financial institutions built on lean, shareholder-driven staffing models, which means the teams Crowe works with rarely have time to sit back and think. As those institutions move into agentic AI, the risk shifts from staffing to accountability.
Crowe points to the Air Canada case, where a court held the airline to a commitment its chatbot made to a customer, as the moment accountability for AI output became a legal question rather than a reputational one. “Accountability now is something that’s sitting on all the organizations,” Crowe said.
That is the governance work CMBYND does for its bank clients: building the oversight mechanisms that let a company use agentic AI in client-facing work without losing control of what gets promised on its behalf.
The Framework: Scaling an Organization’s Thinking, Not Just Its Output
| Principle | What it means in practice | Named evidence from this interview |
| Diamond over pyramid | Narrow the entry-level layer of the organization and route AI-assisted research straight to knowledge workers, instead of staffing a wide base of juniors to produce it | At CMBYND, Crowe rebuilt the firm’s staffing model into a diamond shape, letting the firm deliver research and thought pieces to its client-facing team faster than the pyramid model allowed |
| Prompt AI like a new hire | Give AI the context you would give a new employee: who you are, what you need, and why, instead of typing a bare search term | CMBYND built an internal lab to retrain staff away from Google-style search habits toward structured prompting, a shift Crowe calls central to the firm’s AI adoption |
| Doing to thinking | Once AI produces an answer, the CEO’s job is to keep the organization’s thinking active: challenge the source, the context, and what was left out before accepting the output | Crowe cites CMBYND’s internal data showing roughly 75% of people agreed with AI-generated conclusions without question, which pushed the firm to retrain staff to interrogate AI output instead of accepting it |
| Own the AI’s mistakes | Legal accountability for AI-generated commitments to customers sits with the company, not the AI system, so governance must scale alongside AI use | Crowe points to the Air Canada case, where a court held the airline to a commitment its chatbot made to a customer, as the model for how CEOs must now treat AI-driven communication |
Quotes from This Episode
- “The vast majority of people take what is produced from AI as gospel.” — Chris Crowe, CEO, CMBYND
- “It will always tell you that you’re right.” — Chris Crowe, CEO, CMBYND
- “We could actually build a diamond instead of a pyramid.” — Chris Crowe, CEO, CMBYND
- “You have to let people fail.” — Chris Crowe, CEO, CMBYND
- “Accountability now is something that’s sitting on all the organizations.” — Chris Crowe, CEO, CMBYND
Frequently Asked Questions
How do I get my team to stop treating AI answers as fact?
Chris Crowe, CEO of CMBYND, found that roughly 75% of his team accepted AI-generated output without question until the firm built a dedicated training program teaching staff to interrogate AI answers the way they would question a junior analyst’s first draft. The discipline is to ask what source the AI used, what context it was missing, and what it left out before treating the answer as final.
Should I restructure my team now that AI can do the work junior staff used to do?
Crowe restructured CMBYND from a traditional pyramid, with a wide base of junior staff producing analysis, into a diamond shape with a narrower entry-level layer and AI-assisted research feeding directly to client-facing knowledge workers. This works only if the organization also retrains staff on how to prompt AI effectively, since the structural change alone will not produce better output.
Who is legally responsible when a company’s AI chatbot makes a mistake with a customer?
Crowe points to the Air Canada case, where a court held the airline accountable for a commitment its chatbot made to a customer, as evidence that legal responsibility for AI-generated communication sits with the company, not the AI system. CEOs moving into agentic AI for client-facing work must build governance and oversight frameworks before, not after, an AI system starts making commitments on the company’s behalf.
CEOs Work with Glenn Gow to Scale Their Companies and Careers
Glenn Gow is The Scaling Executive Coach. He coaches ambitious executives into the CEO seat and CEOs into successful exits. With 25 years as a CEO and 5 years in venture capital, Glenn helps leaders scale their companies by scaling themselves first. If this conversation was useful, you can apply for executive coaching with Glenn Gow or apply to be a guest on The Scaling Executive Podcast.
