Performance review season is arriving earlier for a growing number of small businesses, and so is a new helper: AI. Managers who once stared at a blank feedback form are now opening ChatGPT, Claude, or a built-in HR tool to draft the first pass.
The practice of conducting structured performance and development reviews isn't new. What's changed is the toolkit. HR platforms aimed at small businesses, including Lattice, 15Five, and BambooHR, have added AI features over the past two years that summarize peer feedback, flag vague language, and generate draft write-ups based on a manager's bullet points. Standalone chatbots get used the same way, informally, by managers who don't have dedicated HR software at all.
The appeal is straightforward. Writing thoughtful, specific feedback for five or ten employees takes hours most small business owners and managers don't have. A tool that turns rough notes into a structured paragraph can cut that time significantly, and several HR software vendors now market this as a headline feature rather than an afterthought.
But the mechanics of a good review haven't changed just because a machine is involved. Reviews still need concrete examples, measurable goals, and a record of how feedback connects to actual performance over the prior period. AI can smooth the language; it cannot supply the specifics a manager hasn't already observed and documented.
This fits a broader pattern in workplace software. Scheduling, hiring screens, customer service replies, and now performance feedback have each gone through the same cycle: a mundane, high-volume writing task gets offered as an AI shortcut, adoption spreads quickly among time-strapped small business owners, and legal or quality concerns surface a year or two later. Recruiting saw this with AI-screened resumes, which drew discrimination complaints and new state disclosure laws. Performance management is arguably next in line, since written reviews are frequently used as evidence in wrongful termination and discrimination claims.
For small businesses, the trade-off is time saved versus documentation risk. A review drafted by AI from thin manager notes can come out generic, and generic reviews are precisely the kind that attorneys for a terminated employee look for, because they suggest the company wasn't tracking performance closely before the decision to let someone go. On the other hand, a manager who uses AI to turn solid, specific notes into clear prose faces no greater legal exposure than one who writes the same thing by hand; the tool is just a drafting aid, not a decision-maker.
Employment lawyers who work with small businesses generally advise treating AI-drafted reviews the way you'd treat a draft from a junior HR assistant: useful for structure and tone, but requiring a manager to verify every factual claim, specific date, and performance example before it goes in an employee's file. A handful of states, including Illinois and Colorado, now have or are drafting disclosure rules for AI use in employment decisions, and performance reviews that feed into raises or terminations could eventually fall under those rules.
Watch for two things over the next several months: whether more states extend AI hiring-disclosure laws to cover performance management, and whether mainstream HR platforms start adding audit trails that show how much of a review was AI-generated versus manager-written. Both would signal that regulators and vendors see this as a bigger issue than a convenient writing shortcut.
For now, the practical move for small business owners is simple. If a manager uses AI to draft a review, require that every specific claim be checked against real documentation, and keep a record of the manager's original notes separate from the AI-polished version in case the review is ever challenged later.