I meet with many executive suites of mid-market enterprises who have bought into a broken narrative about AI transformation. They’ve been led to believe that if you buy enough licenses, host enough hackathon, and tell your staff to “experiment,” the future will bubble up from the cubicles. You wait for the magic to happen. You expect an AI revolution to rise from the bottom, fueled by curiosity and a shared Slack channel for prompt engineering.
It will not happen.
For two years, leadership teams have watched these experiments with a patient, indulgent eye. They track adoption rates like high school attendance. They celebrate when ninety percent of the marketing team logs into a new AI tool. They point to a “prompt library” built by the staff as if it were a digital library of Alexandria. But an operator looking at the books sees a different story. The reality is cold. This grassroots movement creates some excitement and energy. But mostly noise. It builds a chaotic client zoo of fragmented tools. And, unfortunately, it ususally leaves the profit and loss statement exactly where it found it.
Writing an email ten percent faster does not change the fate of a company. Summarizing all your meetings does not alter your competitive standing. The MIT NANDA study recently revealed a grim ceiling for this approach: while many investigate these tools, only five percent of organizations reach a state of sustained productivity or profit change within six months. The rest are merely playing with digital toys while the clock runs out.
The Sharper Pencil and the Swamp
The flaw in the bottom-up approach is how it views the work itself. It treats these tools as personal efficiency aids. They are often designed to accelerate existing workflows. In other words, it gives the knowledge worker a sharper pencil.
When you crowdsource a strategy, you ask employees to fix their own small, isolated corners. A copywriter uses it to churn out more adjectives. An analyst uses it to scrub a spreadsheet. These are local victories. They are incremental gains that the existing bureaucracy quickly swallows whole. If a marketer saves two hours a week, those two hours are not returned to the company in the form of higher profit. They are lost to the friction of the next meeting or the next approval chain.
Profit does not hide in the individual task. It hides in the spaces between departments. It sits in the handoffs, the cross-functional knots, and the decision-making walls that slow an organization to a crawl. When you treat these tools as features for the individual rather than an architectural layer for the enterprise, you lose the game before it begins. You cannot fix a broken assembly line by giving every worker a faster hammer. You have to redraw the line.
The Force of Top-Down Intentionality
Moving past the crowdsourcing myth requires a shift in conviction. It demands that leadership shifts from asking “what can our people do with this” and start declaring what the organization will become. You cannot spread your AI budget thin across thousands of employees hoping for a spark. Instead, you must choose the two or three core workflows which if re-imagined them from the ground up, can deliver the most business impact.
This work begins with a shift in what we measure. We must stop counting output artifacts like the number of emails sent, code lines written, or reports generated. Instead, we must solve for the time it takes to make a decision. The goal is to see how fast a piece of market data or a customer insight can be turned into an executive action. If a new system allows a leader to see a shift in the market and pivot the supply chain in hours instead of weeks, that is a result you can see on a balance sheet. That is the kind of work we focus on in our Board & CEO Advisory sessions: moving the needle from activity to impact.
The next move is to pull these tools out of the hands of the “experimenters” and consolidate them into core workflows. You identify the processes that define your advantage. Perhaps it is predictive risk underwriting in a fintech firm or deep compliance validation in a healthcare provider. You build centralized, managed pipelines for these specific functions. You do not let individuals pick their own disconnected tools from a menu of a thousand options. You create a single, high-force engine that drives the entire process.
The Necessity of Structural Change
Finally, you must accept the weight of a structural redesign. If a re-engineered workflow allows a team of five to do the work that once required a department of fifty, the old structure cannot remain. You cannot keep the same hierarchy, the same middle management, and the same reporting lines when the nature of the work has changed. The change in the process must force a change in the organization itself.
Many leaders hesitate here. They fear the disruption. They worry about the identity threat that comes when long-standing roles are rendered obsolete. But keeping the old structure in place is an expensive form of nostalgia. It ensures that the gains from your technology investments are eaten by the overhead of a ghost organization. You must have the authority to redraw the workflow and the courage to clear away the ruins of the old one.
It’s no longer enough to host the best hackathons or built the largest internal libraries of generic tips. The winners will be the ones whose leadership had the clarity to see through the myth of the bottom-up revolution. They will be the ones who took the wheel, ignored the noise of vanity adoption, and rebuilt their core business for speed and profit. They will be the ones who realize that while great ideas and innovations can come out of bottom-up experimentation, true organizational change does not typically bubble up. It is envisioned and driven down from the top.
