How 6 Companies Use AI for Performance Management
How 6 companies use AI for performance management, from Meta and J.P. Morgan to high-growth startups like Rho and Case Status.
TL;DR: Companies are adopting AI for performance management in two distinct ways. Large enterprises like Meta, JPMorgan, and Microsoft are layering AI into existing review processes — and in some cases turning AI usage itself into a performance criterion. High-growth companies like Rho, Case Status, and Nirvana have reinvented their performance review cycles to be AI-native. Both patterns are accelerating fast, but they reflect different bets about what AI is good for.
Below: six companies, what they’re actually doing, and what HR leaders can take from each approach.
Why Companies Are Turning to AI for Performance Management
Two pressures are meeting at once. Managers spend disproportionate time on review prep: gathering work artifacts, recalling six-month-old context, drafting from memory. At the same time, AI capabilities have crossed the threshold where drafting and synthesis are reliable enough to trust in production.
That’s produced three distinct adoption patterns. Some enterprises are using AI to draft and summarize within their existing review processes. Others are going further and treating AI usage itself as a performance criterion. Smaller, AI-forward companies have skipped the legacy step and reinvented the review cycle as AI-native from end to end.
Meta — Tying “AI-Driven Impact” to Performance Ratings
Meta is formally baking AI usage into 2026 performance reviews. In a November 2025 memo, Head of People Janelle Gale told employees that “AI-driven impact” will be a core expectation across all roles, from engineers to marketers, with reviewers assessing both how employees use AI and whether they build tools that improve productivity.
Meta also rolled out an internal “AI Performance Assistant” to help staff write reviews, and many employees have begun using Metamate, Meta’s internal AI bot, to draft self-review content. The shift was reported by HR Dive and HR Grapevine, citing the internal memo. Meta is treating AI adoption as a measurable performance dimension, not just a tool.
JPMorgan Chase — Drafting Reviews with LLM Suite
JPMorgan Chase rolled out LLM Suite, its in-house large language model platform, to help employees draft year-end performance reviews. Within eight months of launch, the platform had onboarded 200,000 users, one of Wall Street’s most extensive AI deployments. The tool generates a draft from employee prompts that the user then edits.
Internal guidance specifies that AI-generated text is a starting point only, the employee remains responsible for the final review, and the system is not used in pay or bonus decisions. Boston Consulting Group reports a 40% reduction in writing time when staff use AI to draft reviews, with quality scores rising 20%. The pattern at JPMorgan: scope AI to drafting and synthesis, keep humans in charge of ratings and comp.
Microsoft — Making AI Usage a Performance Criterion
Microsoft made AI tool usage mandatory in mid-2025. In a memo from Julia Liuson, Corporate VP of Microsoft’s Developer Division, leaders were told that “AI is now a fundamental part of how we work” and that “using AI is no longer optional — it’s core to every role and every level.” Managers were directed to factor employees’ AI usage into their performance evaluations.
The Developer Division oversees GitHub Copilot, and the move is widely read as a forcing function for internal Copilot adoption. The shift was first reported by Business Insider. Microsoft is using performance reviews as a lever to accelerate AI adoption, a pattern likely to be controversial and copied.
Rho — Cutting Review Hours by 83% with an AI Agent
Rho, the modern banking platform for startups, runs its full performance review cycle on Windmill. The People team replaced a legacy approach where managers spent hours gathering work artifacts and feedback across tools. Their cycle now completes in 8 days end-to-end (self-reviews in 2.5 days, 360s in 5) versus an industry average of several weeks.
Results: 83% reduction in total hours spent on reviews and 93% of employees preferred Windmill to the prior process. As Rho’s CFO Michael Szarowicz put it, “It wrote the draft of the review in a very human way… I’m just polishing what was a better quality review than I would’ve written if I’d spent 3 hours on it.” Read the full Rho case study.
Case Status — Solving the “Blank Page” Problem
Case Status, a client experience platform for law firms, used Windmill to solve what the team called the “blank page problem”: managers and employees staring at empty review forms trying to reconstruct months of work from memory. Windmill’s AI surfaced work context from Slack and other tools, generating summaries that became the starting point for every review.
The cycle ran 84% faster than the prior process built on Google Docs and forms. 93% of employees preferred Windmill. Average time per review: 6 minutes for peers, 7 for upward, 9 for managers. The full Case Status story shows AI’s biggest unlock isn’t review quality but eliminating the prep work that makes reviews painful.
Nirvana — Running a Full Cycle Without an HR Team
Nirvana, a healthcare technology company, ran its mid-year performance review cycle without a dedicated HR function. Windmill handled the coordination, drafting, and reminders that would normally require full-time HR staff. The team hit 100% completion and 80% employee preference for the new process.
For mid-sized companies that don’t have the HR headcount to run a traditional cycle, AI agents close a real gap. Nirvana’s approach proves the bottleneck at this size isn’t strategy, it’s the operational work no one has time to do. See the Nirvana case study.
How Each Company Uses AI
| Company | Use Case | Tool / Platform | Key Outcome |
|---|---|---|---|
| Meta | AI-driven impact as review criterion | AI Performance Assistant + Metamate | Formal rollout in 2026 |
| JPMorgan Chase | Drafting year-end reviews | LLM Suite (in-house) | 200K users in 8 months; 40% faster writing |
| Microsoft | AI tool usage as performance metric | GitHub Copilot adoption | ”AI is no longer optional” memo |
| Rho | Full review cycle on AI agent | Windmill | 83% fewer hours; 93% preference |
| Case Status | AI-drafted reviews from work data | Windmill | 84% faster cycle |
| Nirvana | Cycle coordination without HR team | Windmill | 100% completion; 80% preference |
What HR Leaders Should Take Away
The split between enterprise and AI-native approaches matters more than the headlines suggest. Enterprise rollouts layer AI into existing review machinery. AI-native rollouts replace the machinery itself, which is where the real time savings come from.
Two practical reads:
- Drafting is the easy win. Every enterprise rollout above starts here. If your managers still write reviews from a blank page, that’s the lowest-hanging fruit, and BCG’s own data suggests 40% writing-time reductions are realistic.
- The bigger gains come from coordination. The Windmill customers above hit 80%+ time reductions because the AI handles scheduling, reminders, peer-feedback collection, and synthesis. Drafting is one piece of a much larger workflow.
For more, see our guide on how to use AI for performance reviews and the top AI tools for performance reviews in 2026.
Frequently Asked Questions
Which companies are using AI for performance management?
Major enterprises including Meta, JPMorgan Chase, and Microsoft have rolled out AI tools for performance reviews. Meta is tying AI-driven impact to formal ratings starting in 2026. JPMorgan launched LLM Suite to help employees draft year-end reviews. Microsoft is factoring AI tool usage into employee evaluations. High-growth companies like Rho, Case Status, and Nirvana are running full review cycles on AI agents.
How is Meta using AI in performance reviews?
Meta is tying employee performance reviews to 'AI-driven impact' starting in 2026. Head of People Janelle Gale told staff in a November 2025 memo that AI usage will be a core expectation across all roles. The company also rolled out an AI Performance Assistant, and many employees are using Metamate, Meta's internal AI bot, to draft reviews.
Does JPMorgan use AI for employee evaluations?
Yes. JPMorgan Chase rolled out its in-house LLM Suite in 2025 to help employees draft year-end performance reviews. The platform had onboarded 200,000 users within eight months. JPMorgan's policy makes clear that AI-generated text is a starting point only, the employee remains responsible for the final review, and the system is not used in pay or bonus decisions.
What are AI-native performance management platforms?
AI-native performance management platforms are built around AI from the start, rather than bolting AI features onto legacy review software. Tools like Windmill collect work context year-round through integrations with Slack, GitHub, Jira, and other systems, then use that data to draft reviews and run feedback cycles automatically. Rho, Case Status, and Nirvana run their full performance review cycles this way.
Should mid-sized companies use AI for performance management?
Mid-sized companies often see the largest gains from AI in performance management because they typically lack the dedicated HR resources of large enterprises. Examples like Nirvana — which ran a full mid-year cycle with no dedicated HR team — show that AI can replace much of the coordination and drafting work that previously required full-time staff.