Performance Reviews Are Changing: How AI Fluency Is Becoming a Key Metric
The performance review just changed. Most managers haven’t caught up. Today, AI fluency is moving from a nice-to-have skill to a formal measure of performance. But many managers still don’t know how to evaluate it fairly. This shift is happening fast, especially at big tech companies like Meta, Google, and Microsoft. As more companies add AI skills to their review criteria, managers need a clear playbook to assess employees accurately and help them grow.
How the Performance Review Is Changing
Performance reviews now measure new behaviors. While teamwork and output still matter, managers are also asked to evaluate how employees use AI to get work done. Meta’s new Checkpoint system is a good example. It tracks over 200 data points for software engineers, including how much code they write with AI tools versus on their own. Top performers can earn a 200% bonus multiplier, and Meta’s top-tier award offers a 300% multiplier.
Other big tech companies are following. In June 2025, a Microsoft executive told managers that AI use was no longer optional. That same summer, Google CEO Sundar Pichai shared a similar message during an all-hands meeting. Now, this expectation is spreading to the rest of corporate America. The challenge? Many managers still don’t have a clear playbook for what AI fluency looks like, how it varies by role, or how to evaluate it fairly.
Why Mid-Market Companies Are Behind
AI performance expectations are moving beyond big tech faster than many companies can support. Large employers often have internal AI academies, custom tools, and change management teams. Most mid-market companies are trying to make the same shift with fewer resources.
The pressure shows up in hiring data. According to the RSM U.S. MMBI Special Report 2026, 52% of middle-market executives expect moderate to significant hiring needs over the next year. To meet demand without adding headcount, many are turning to AI and skills training. But they still lack the training, tools, and support to scale AI effectively.
This gap matters most for companies in the $50 million to $1 billion revenue range. Demand for AI skills has climbed sharply since 2023, but the supply of workers with verified AI competencies hasn’t kept pace. Employees at large companies build AI fluency through structured programs, while many mid-market workers learn on their own.
Where Leadership Is Misaligned
The push to adopt AI is moving faster than many leadership teams can align around it. According to Grant Thornton’s 2026 AI Impact Survey, CIOs and CTOs are five times more likely than COOs to say their workforce is ready to adopt AI. That’s a problem because the executives setting AI strategy aren’t always the ones seeing how those mandates play out in daily workflows.
When strategy and operations work from different assumptions, employees can end up being evaluated against expectations they aren’t equipped to meet. This tension shows up in performance reviews, where managers are asked to evaluate AI fluency without clear standards, and employees are asked to prove impact without consistent support.
What the New Performance Review Requires
Closing the gap between AI mandates and manager readiness takes more than a memo. Here are four shifts that can help managers evaluate AI fluency in a way that reflects employee contribution fairly.
1. Redesign workflows first
Training employees on AI tools without restructuring how work gets done leaves them with skills they can’t fully apply. McKinsey argues that AI adoption should be treated as a change management initiative, not just a training program. Workflows need to be redesigned before employees are evaluated on how well they use the tools.
2. Define role-specific criteria
Generic AI literacy doesn’t translate neatly into a performance rating. Grant Thornton’s survey found that training is often disconnected from actual workflows. A useful starting point is defining what strong AI use looks like in each role, whether that’s a marketing coordinator, an operations analyst, or an account executive.
3. Establish baseline metrics
Tying ratings to AI adoption only works if there’s a baseline to measure against. Without one, managers risk evaluating perception instead of measurable improvement. Leaders need to understand current AI usage and productivity across their teams before holding employees accountable for progress.
4. Align the C-suite
The CIO/COO divide Grant Thornton identified can show up at review time as inconsistent expectations across teams. Closing that gap before review season gives managers clearer standards to evaluate against. It also protects employees from being graded against shifting criteria they can’t anticipate.
What Managers Should Do Now
The 2026 review cycle will test how well companies can translate AI ambition into fair evaluation. Managers who treat AI fluency as a checkbox may widen the gap between what leadership expects and what employees can realistically deliver. Managers who treat it as an ongoing feedback loop have a better chance of closing it.
That means using review conversations to identify where employees lack tools, training, or workflow support—not just where they fall short. It also means pushing back when expectations don’t match the resources available. A performance rating only works as a development tool when the standards behind it reflect what employees have been given to work with.
The performance review is becoming one of the primary levers companies are using to drive AI adoption. Whether that lever helps employees grow or leaves them behind will depend on how managers handle the gap between mandate and capability.
Comment