Managers are using AI to write performance reviews – and it shows

Highwire’s survey of 1,034 corporate employees found 78% of managers used AI to help prepare feedback or performance reviews in the past year, while only 16% of non-managers said they were told AI influenced their review. Employee reactions were mixed: 54% said feedback became more specific and actionable, but 34% found it more generic and 32% less useful. Nearly one in four employees rehearse difficult workplace conversations with AI, amid a reported gap in manager support; Synthesia launched its Sessions practice feature, but its effectiveness remains unproven.

Analysis

The investable signal is less “AI replaces managers” than “companies may automate the preparation layer while leaving accountability with managers.” That can improve review consistency, but if generic or poorly grounded feedback erodes employee trust, savings may be offset by retention, grievance, or performance-management costs—risks that are diffuse and unlikely to show up quickly in vendor revenue.

Economics favor platforms that already sit inside enterprise workflows and can prove adoption, governance, and measurable learning outcomes. Standalone coaching and training tools face a harder case: simulation features are relatively easy to copy, so usage may rise without supporting incremental pricing. Buyers could capture most of the productivity benefit. Sensitive performance data and recorded practice sessions also create privacy and governance exposure; a high-profile misuse or policy restriction could slow deployment.

Near term, this survey and product launch are weak catalysts, not a basis for a sector position. Over 1–3 months, look for procurement evidence, paid-seat expansion, and renewal commentary rather than feature announcements. Over 6–18 months, the differentiator is likely validated improvement in employee outcomes and safe integration—not avatar realism. The survey is self-reported and sponsored by a professional-development firm whose incentives may favor human-led training; Synthesia’s product rationale is also a company claim, not efficacy evidence. The contrarian risk is that investors overread usage as monetization: workplace AI use can expand while standalone vendors face bundling and price pressure.

Key Decisions for Investors

• No standalone trade on this item. Treat it as a watch signal for enterprise software and learning-and-development vendors; do not infer revenue impact from reported employee usage.
• For the next 1–3 months, monitor enterprise procurement, paid-seat growth, renewal rates, and customer evidence of improved training or review outcomes. Upgrade the theme only if vendors demonstrate willingness to pay beyond existing software bundles.
• Track governance as a downside catalyst: verify whether deployments use sensitive performance data, what retention and access controls apply, and whether customers impose restrictions after privacy or employee-relations incidents.
• Falsify the adoption thesis if product launches fail to convert into paid expansion or renewal uplift; falsify the trust-risk thesis if independent customer evidence shows reviews becoming more actionable without worsening employee sentiment or disputes.