Workers use AI for efficiency, but organizations want transformation

What AI enables without strategic direction

Hannah Maxson, marketing manager at Centered Health, is a great example of just how many workers use AI when faced with an expanded scope and limited guidance. Maxson found herself working solo after her team was downsized. She went from handling campaigns and copywriting to adding compliance, operations, and internal communications to her workload.

"My team was downsized drastically," Maxson says. "These aren't entirely new types of projects, but they'd typically be shared across multiple people with different skills and expertise. AI has helped me navigate that learning curve much faster."

Recently, Maxson led a large Salesforce-related compliance and data-mapping project, for which she used ChatGPT to build a four-phase rollout plan. "That alone removed a lot of pressure because I no longer felt like I had to solve everything at once," Maxson says. The AI helped her understand what questions to ask and where to navigate in the system setup, with clear steps on what to turn on.

Even though she used AI to execute, the judgment came from her. "The decision-making still came from me, especially around compliance, industry requirements, and what makes sense for our organization's data," she says. "AI could guide the process, but it couldn't make the final call on what was actually necessary for our business."

She uses AI across a broadening workload, including video production coordination, email tone-checking in a sensitive industry, and daily task planning based on her energy levels. "My workload hasn't necessarily decreased; if anything, expectations have increased, but how I spend my time has changed," Maxson says. "AI has taken over first drafts, outlining, and a lot of emails, which means I spend less time staring at a blank page."

She's clear about AI's impact on her creative work. "AI hasn't made me more creative on its own, but it has made me more confident and capable of executing independently," Maxson says. "By reducing friction and second-guessing, I have more space to think strategically and show up better in the parts of the work that actually require a human perspective."

This is rational adaptation. Given unclear mandates about what AI should enable, workers optimize for what's measurable: speed, task completion, and breadth of coverage. Maxson can now handle compliance projects that would have required outside consultants. She coordinates video production that used to take multiple people. She manages a scope of work that didn't exist for a single person before AI. That's real capability expansion driven by efficiency. Without clarity about what 'better work' means beyond 'more work,' efficiency is what workers optimize for."

Jessica Swain, social media manager at Found, describes her AI use similarly. Her company built a personalized GPT that knows her posts, tone of voice, and career history. "I'll say I need to write a post, and it will spit out a post for me in my tone of voice,” Swain says.

The productivity vs. performance gap

"We're seeing the pendulum swing from 'go drive our AI strategy' to 'why aren't people using our AI well enough,'" says Josh Cardoz, chief creative and learning officer at Sponge, a people enablement agency that works with some of the world’s biggest and most complex brands. "The key metric is productivity-pushing as a silver bullet."

Organizations measure what's easy to track, such as adoption rates, monthly license usage, and time saved on tasks. But Cardoz argues these metrics miss whether AI is actually improving the quality of work. The problem starts with sequencing, he says. Companies lead with the tool rather than the goal or the human-centric skill that needs developing.

He points to a Fortune 100 client where AI is the mandate, but the company is doubling down on storytelling as the core capability first. Rather than starting with how to use AI for storytelling, the program teaches what storytelling is, where individuals fit in, and what workflows look like. AI agents and coaching tools get integrated later as accelerators.

The difference is leading with capability rather than heading straight in with technology. Define what great performance looks like, then introduce AI to amplify it. When organizations flip that sequence and lead with the tool, they get friction and resistance because people don't understand what they're supposed to be achieving.

The result is tension between productivity and performance. Productivity is doing more, faster, whereas performance is doing better work that’s more creative, more strategic, and more aligned with business outcomes. Recent research from Harvard Business Review reveals why the same tools produce different outcomes: AI boosts creativity primarily for employees with strong metacognition, which encompasses the ability to plan, monitor, and refine their thinking. These individuals strategically use AI to expand knowledge and free cognitive capacity. Others use identical tools for efficiency without creative gains.The same technology produces different outcomes depending on whether organizations have defined what creative or strategic AI use looks like in their context.

Measuring application over adoption

When organizations roll out AI training, most track adoption and completion rates. Cardoz recommends measuring something different: lead behaviors that indicate employees are actually applying new capabilities, not just using tools.

"What are the lead behaviors with intended performance?" Cardoz says. In the storytelling example, that means looking for evidence that employees are leveraging frameworks in presentations. "That is an indicator of moving in the direction of what would be desirable behaviors for the organization."

Then AI tools—a coaching agent for preparing pitches, a synthetic focus group before a presentation—become accelerators rather than just adoption drivers, he argues. A tech company Sponge works with takes this approach when training salespeople to sell AI products: build the right mindset first, then integrate AI tools to support it.

Completion rates and engagement scores will always matter, Cardoz says, but the more valuable metric is whether behavior is moving in the right direction; whether people are applying frameworks, making better decisions, or approaching work differently. Those behaviors signal performance improvement, beyond just doing more.

The challenge is that most organizations haven't defined what those behaviors look like. They've rolled out tools, created training on how to use them, and measured adoption. But they haven't clarified what "using AI well" means for a sales manager versus a compliance specialist versus a creative team.

So workers like Maxson and Swain make rational choices. They use AI for what's clear and measurable: getting through tasks faster, handling a broader scope, and executing independently. The efficiency gains are real, but the strategic transformation isn't.

What comes next

The technology is ready, and workers are adapting. The question is whether organizations will define what "better work" means before optimizing for "faster."

Until they do, employees will keep making rational choices like getting more done, handling broader scope, and executing independently. People will keep using AI to navigate complex projects solo and manage expanded responsibilities, still delivering real value to their organizations. But that value is fundamentally about capability expansion through efficiency, not creative transformation or strategic innovation. The gap between what AI enables and what organizations could be getting won't close until companies clarify what performance looks like beyond productivity, create the conditions for it to happen, and measure whether they're achieving it.

Workers are delivering real efficiency gains, but the strategic transformation organizations expected hasn't materialized. The difference comes down to what companies define, what they measure, and what they're willing to invest in beyond tool adoption.