A business leader I spoke with recently described his company’s AI strategy in one sentence: “By the end of this quarter, we want to achieve a 10% uplift in productivity using AI.”
That sentence does something clever. It sounds strategic, measurable, ambitious, directional. But it’s constructed in a way that lets everyone off the hook. Ten people in that organization will interpret it ten different ways. No one will be wrong. And when the quarter ends and nothing has materially changed, everyone will have a reasonable explanation for why.
Vague goals aren’t a planning mistake. They’re an organizational defense mechanism.
Productive at What?
The word “productivity” is doing enormous work in most AI conversations, and almost none of it is precise.
When executives talk about productivity, they might mean producing more output with the same resources, producing the same output with fewer resources, completing work faster, improving quality, or expanding organizational capacity. These are not variations of the same objective. They require different interventions, different measurements, and different definitions of success.
A sales leader defines productivity as more customer conversations. An operations leader defines it as more units per labor hour. A customer service manager defines it as more tickets resolved per agent. When a company sets a “10% productivity improvement” goal without specifying which of these it means, it hasn’t set a goal. It’s issued a placeholder.
The real question is never how do we improve productivity by 10%? It’s productive at what, specifically, and for whom?
The Constraint You’re Not Naming
Most AI discussions start with efficiency. That’s the wrong starting point, not because efficiency doesn’t matter, but because efficiency is rarely the actual constraint on performance.
Think about the organizations where AI initiatives have produced real results. In most cases, they didn’t deploy AI and then watch productivity rise. They identified a specific constraint, something that was limiting growth, slowing customers down, or consuming disproportionate capacity, and then determined whether AI could remove it.
The question that unlocks this isn’t how can we automate this task? It’s what is preventing us from doing more of what matters?
Maybe your sales team spends 40% of their time on CRM updates instead of customer conversations. Maybe proposals take three weeks to generate because four people have to touch them. Maybe your support team resolves tickets slowly, not because of staffing, but because agents can’t surface the right information fast enough.
These are constraints. They’re specific, measurable, and solvable. Productivity is a byproduct of removing them, not a goal you can pursue directly.
A Question That Forces Honesty
Here’s a diagnostic worth running before setting any AI target:
If we achieved a 10% productivity gain, what would actually be different?
Not in general terms. Specifically. What would a customer experience differently? What would an employee spend their time on instead? What would a manager be able to see or decide that they can’t today?
If the answer is vague, “we’d be more efficient,” “things would move faster,” the goal isn’t real yet. You haven’t defined the outcome; you’ve described a feeling.
When you force specificity, the answer changes. It becomes: Sales representatives will spend 20% less time on administrative work and 20% more time interacting with customers. Or: Customer onboarding drops from ten days to seven. Or: First-draft proposals that currently take two weeks get generated in two hours.
Now you have something you can design, test, measure, and improve.
Outcomes First, AI Second
Most organizations begin with: We need to use AI. Then they go looking for places to put it. The organizations creating real value do the opposite. They start with a customer problem, a business bottleneck, a growth constraint, or a strategic objective, and then determine whether AI is the right tool to address it. Often it is. Sometimes it isn’t. But the sequence matters.
Starting with AI leads to deployment without impact. Starting with outcomes leads to AI that earns its place.
This is the failure mode behind most “10% productivity” goals. The target was set before anyone asked what problem they were actually solving. AI gets applied somewhere visible. Activity gets mistaken for progress. The quarter ends. The goal gets reset.
Ask yourself: What specific business outcome are we trying to improve?
The harder question
The business leader I mentioned at the start isn’t unusual. Most executives setting AI targets are working under real pressure to show progress, demonstrate ROI, and justify investment. The vague productivity goal is, in part, a response to that pressure.
But pressure doesn’t make imprecision strategic. It makes it expensive.
The companies that will create the most value from AI over the next several years won’t be the ones with the highest deployment counts or the broadest tool adoption. They’ll be the ones who identified the specific constraints limiting their performance, designed interventions around measurable outcomes, and held themselves accountable to results that actually changed something.
Before your next AI planning conversation, bring this question into the room: If we hit this goal, what would be different, specifically?
If no one can answer it, you don’t have a goal. You have a target that lets everyone look busy until the next quarter.
Bottom line: The organizations winning with AI aren’t chasing productivity metrics. They’re removing the constraints that limit performance.
Related Posts
