The fastest-growing segment in consulting is AI strategy. Every firm — from the Big Four to the two-person boutique — now has an "AI practice." They'll audit your readiness, map your use cases, build you a roadmap, and hand you a 120-page deck.
The dirty secret? Most of the people writing those decks have never managed a budget. Never led a team through a quarterly earnings call. Never been accountable when a product launch fails. Never had to explain to a board why a million-dollar initiative didn't deliver.
They've studied AI. They haven't deployed it under pressure.
The consultant's sandbox vs. the operator's reality
There's a fundamental difference between building an AI proof of concept in a controlled environment and deploying AI inside a real business with real constraints.
In the sandbox, everything is clean. The data is curated. The stakeholders are aligned. The timeline is generous. The success metric is "it works."
In operations, nothing is clean. The data lives in seven different systems, three of which are legacy. Half the team is skeptical. The budget was cut last quarter. And the success metric isn't "it works" — it's "it moves the number."
I know this because I've lived on both sides. As CEO of Yaneken — Chile's second-largest multi-brand retail group — I didn't hire someone to tell me where AI could help. I built it. We deployed AI-powered customer service across brands. We automated inventory processes. We used machine learning for demand forecasting across 158 stores.
None of it looked like a consulting slide. All of it required understanding the business first and the technology second.
The three things consultants miss
1. Change management is the actual product.
The hardest part of AI deployment isn't the model. It's the people. Store managers who've done things the same way for fifteen years. Operations teams who see automation as a threat. Middle management who doesn't understand what a language model does and doesn't want to ask.
Consultants build the technical solution and call it done. Operators know the technical solution is 20% of the work. The other 80% is getting humans to adopt it, trust it, and use it correctly.
2. ROI isn't optional.
In consulting, the deliverable is the recommendation. Whether it works is someone else's problem.
In operations, every dollar spent on AI is a dollar not spent on hiring, inventory, marketing, or rent. You don't get to run experiments forever. You need to show a return — in months, not years. That pressure changes how you think about which AI to deploy, where, and when.
I've killed AI projects that were technically impressive but commercially pointless. That's a judgment call you only develop by carrying the P&L.
3. Integration beats innovation.
Consultants love novel solutions. The latest model. The most sophisticated architecture. The bleeding-edge framework.
Operators love solutions that integrate with what already exists. The warehouse management system from 2019. The POS that runs on Windows. The ERP that nobody wants to replace because the migration would take eighteen months.
Real AI deployment is ugly. It's duct tape and API calls and workarounds. It's making GPT-4 talk to a system that was built before GPT existed. Consultants don't do ugly. Operators live in it.
The question you should ask
Before you hire an AI consultant — or an AI advisory firm — ask one question:
Have you ever had a P&L?
Not "have you advised someone with a P&L." Not "have you worked at a company with a P&L." Have you, personally, been responsible for revenue, costs, and margin?
If the answer is no, they're guessing. Educated guessing, maybe. Thoughtful guessing, sure. But guessing.
If the answer is yes, they know what it means to deploy something that has to work on Monday morning. They know the difference between a strategy that looks good in a deck and a strategy that survives contact with reality.
Why this matters now
We're at an inflection point. AI is moving from "interesting experiment" to "operational necessity." The companies that get this right in the next two years will pull ahead. The companies that don't will spend the decade catching up.
The stakes are too high for guesswork. You need advisors who've been in the seat — who've managed the budget, led the team, and lived with the consequences.
That's not a consulting engagement. That's an operator who now advises.
That's what we built Menlo & Oak to be.