The Second Question
Every AI initiative eventually arrives at a similar moment. A leader stands in front of a room and reports the results of the latest AI investment. Those results typically look like time saved, processes automated, and maybe even headcount costs avoided. The numbers are real and usually defensible. The team worked hard. And yet, when the CFO looks at the income and cash flow statements, nothing has materially moved. Revenue is flat. Margins are where they were. No new cash-generating capability has emerged that the organization can point to with confidence.
The wins are real. The value capture is not.
This is not an AI problem. It is a leadership problem. Specifically, it is a reinvestment problem. And it begins with the question almost no one asks.
Everyone Asks the First Question
The first question is easy: did AI save us anything?
The answer is almost always yes. Generative AI improves execution speed. It reduces cognitive load and accelerates the tasks that used to consume entire mornings. Teams move faster, and the efficiency is genuine.
But efficiency is not value. It is potential. And potential without direction dissipates.
If a knowledge worker finishes her work in three hours instead of eight, but nothing else changes, where does that value go? In most organizations, it gets absorbed. Work expands to fill the time available. Meetings multiply. Incremental polish replaces meaningful output. The five hours return to the system without leaving a trace.
>
> **efficiency gap** (*uh-FISH-en-see gap*) *noun*
>
> The distance between what AI created and what the business actually captured.
>
The second question is the one that closes it: what did we do with what we saved?
Most organizations cannot answer it because they never built the structure to answer it. They celebrated the efficiency win and moved on.
The Two Ways Organizations Get This Wrong
When leaders recognize the problem, they typically reach for one of two responses. Both fail.
The first is control. The instinct behind timesheets, utilization dashboards, and productivity tracking in fine increments is understandable; if we can see where the time goes, we can redirect it. The problem is that this approach measures inputs and signals distrust. It produces compliance, not value. Employees fill timesheets the way they fill parking meters, accurately enough to avoid consequences but not honestly enough to change anything. The dashboards look clean and the reinvestment never happens.
The second is culture. Volunteer days, innovation hours, and unstructured time given back to employees in the spirit of goodwill are not without merit. Employee retention has real P&L implications, and any CFO who dismisses that is leaving money on the table. Reduced attrition lowers recruiting costs, shortens onboarding timelines, and preserves institutional knowledge that is genuinely difficult to replace.
But goodwill without a connected financial outcome is a long causal chain. “AI freed capacity, we invested in culture, retention improved, and recruiting costs dropped.” Each link in that chain requires its own evidence, and most organizations are not disciplined enough to close all three links, which means the argument collapses the moment someone in finance asks for proof.
Neither control nor culture is the answer. The answer is intentionality.
Efficiency Is a Beginning, Not an End
The culture shift that actually produces outcomes is not a shift toward more control or more generosity. It is a shift toward treating efficiency as a starting point rather than a destination.
The question that needs to become normal is not where did the time go. It is what did we build with the capacity we recovered.
That shift sounds simple. It is not. It requires leaders to do three things explicitly that almost no one does.
The first is naming the capacity. Not in hours, but in terms of what kind of work was freed and where in the organization it lives. The nature of the capacity determines what it can become.
The second is naming the reinvestment opportunities. Leaders need a specific, prioritized list of initiatives the organization would pursue if it had that capacity. Revenue generating, margin expanding, risk reducing work. That list needs to exist as a maintained, owned document, not a whiteboard exercise that disappears after the offsite.
Get Canon Pattillo’s stories in your inbox
Join Medium for free to get updates from this writer.
The third is drawing the line to the financial statements before the reinvestment begins, not after. Revenue per full-time equivalent. Gross margin improvement from reduced operational expenses. New product revenue. Reduced customer acquisition cost. The reinvestment plan only shows up on the income statement if someone connected it to a line item in advance. After the fact, the story is always murky.
The Leadership Problem No One Is Talking About
Here is where the honest conversation becomes uncomfortable.
Most leaders will understand this. A hero of mine once observed that good leadership is not complex, but it is extremely difficult. The difficulty of leadership is not conceptual. It is execution. It is doing the hard, patient, unglamorous work consistently, in the face of competing priorities and short-term pressure.
The same is true here. The reinvestment plan is not a complex idea. It is an extremely difficult discipline to sustain.
And organizations make it harder by failing to reward it. Companies celebrate the efficiency win and move to the next initiative. The leader who spends eighteen months carefully reinvesting recovered capacity into something that produces durable revenue growth is rarely recognized the way the leader who cut fifteen percent from the labor line is. Until incentives align with reinvestment behavior, the plan will keep dying in the gap between intention and execution. That is not a character failure. It is a systems failure.
There is a second cost that organizations are not tracking. Some leaders have already figured this out. They are the ones quietly redirecting freed capacity into customer relationships, product improvements, and team development. They are not making noise about it. They are building. Because they are not making noise, they are also the easiest to overlook when compensation decisions are made. And so they leave. The organization never knows what it lost, because it never measured what they were building. Any organization that is failing collectively at reinvestment should look carefully before assuming everyone is failing equally. Some of their best leaders have already solved the problem. The question is whether they will recognize them before someone else does.
The CFO and COO Cannot Solve This Separately
This is the part of the conversation that most AI content avoids.
The COO is closest to where the efficiency is created. They understand the operational texture of what was freed, where in the organization the capacity lives, what the team is realistically capable of taking on, and which reinvestment opportunities are credible versus theoretical.
The CFO controls what gets measured, reported, and ultimately rewarded. They have the tools to build the measurement architecture that makes reinvestment visible on the financial statements, and they have influence over the compensation structures that can change leadership behavior over time.
Neither can solve this alone. The COO who reinvests capacity without the CFO’s measurement architecture is doing good work that no one can prove. The organization will eventually ask for evidence, the evidence will not exist, and the program will be cut or deprioritized. The COO will be frustrated. The work will stop. The CFO who builds reinvestment metrics without the COO’s operational intelligence is measuring a machine that is not running. The numbers will not move because the activity is not connected to how work actually flows through the organization.
The reinvestment plan is the artifact they build together, with named capacity, named opportunities, connected financial outcomes, and leadership incentives tied to the long game. This conversation does not happen enough. Vendors sell to the COO on efficiency. Analysts brief the CFO on ROI. The handoff between the two, which is exactly where reinvestment either happens or doesn’t, is left to chance.
It should not be.
Solid Starting Point: A Practical Roadmap
I’ve spoken about the issue for some time. It’s time to try, fail, and try again to create the solution. The leaders in operations, finance, and HR working to close the efficiency gap struggle with the same core issues. We haven’t aligned on a perfect answer. But the path is under construction.
What follows is a starting point, not a finished answer. The sequencing of the first two steps matters and depends on where your organization is standing. If you are planning an AI investment and have the opportunity to be proactive, Step 2 arguably comes before Step 1, you decide where you want to reinvest first, then design the efficiency to deliver that capacity. If you are working retrospectively with gains already realized, you start with Step 1. Either path leads to the same place.
1. Step 1 — Identify the efficiency. If you are planning ahead, name the efficiency you intend to create before the investment is made. Define what kind of work will be freed, where it lives in the organization, and what its strategic value is. If you are working retrospectively, audit what AI has already freed in operational terms, not hours. What kind of work is no longer being done manually, and what was it worth to the business?
2. Step 2 — Build the reinvestment menu. Develop a prioritized, maintained list of initiatives the organization would pursue with that recovered capacity. Each initiative should be revenue generating, margin expanding, or risk reducing. Each should have a named owner and a connection to a financial outcome before any work begins. For proactive organizations, this list is the starting point that shapes how the efficiency is designed in the first place.
3. Step 3 — Assign ownership. Not committees. Named individuals who are accountable for specific reinvestment buckets, with the authority to execute and the expectation that they will report results.
4. Step 4 — Build the measurement bridge. This is the CFO’s job. Connect each reinvestment activity to the line item it is intended to move, and define what progress looks like at six months, twelve months, and eighteen months. Without this step, reinvestment remains invisible to the financial statements regardless of how well the work is being done.
5. Step 5 — Redesign at least one reward. Identify one leadership incentive that is explicitly tied to reinvestment outcomes rather than efficiency creation alone. One deliberate signal that the organization means what it says is enough to change the conversation.
The Honest Close
We do not have all of the answers. The reinvestment plan is not a solved problem. Organizations are still learning how to build it, and the ones doing it well are not celebrated loudly enough.
But the organizations that will win over the next decade are not the ones that automate the most. They are the ones that build the discipline to recapture what automation frees and redirect it toward something that compounds.
That is not an AI story. It is a leadership story. And it starts with the second question.
