Review season is arriving for a lot of small businesses right now, and a growing number of managers are opening ChatGPT or a built-in HR tool before they open a blank feedback form. The technology hasn't changed what good feedback looks like. It has changed who — or what — writes the first draft.
Performance reviews have always been a weak point for small businesses. Owners and managers wear too many hats to spend hours crafting individualized, specific feedback for every employee, so reviews tend toward vague phrases like "needs to improve communication" that don't actually help anyone. That gap is exactly where AI writing tools have found an opening. Managers now feed in notes, project outcomes, or even Slack threads, and ask an AI assistant to turn them into structured, professional-sounding feedback.
Most major HR platforms aimed at small businesses — payroll and performance-management systems like Gusto, BambooHR, and Lattice — have added AI drafting features over the past two years. These tools generate review language, suggest goals, and in some cases flag when feedback is too generic or too harsh. The pitch is consistency: every employee gets a review written to the same standard, regardless of how rushed the manager is.
The underlying advice on what makes feedback effective hasn't moved. Specific examples beat general statements. Clear expectations beat vague ones. A two-way conversation beats a one-way memo. What's changed is that AI can now produce the structure of good feedback — the specificity, the balanced tone, the action items — even when the manager supplying the raw material hasn't done that thinking themselves.
That's the part worth paying attention to. AI can make feedback sound well-organized without making it accurate. If a manager types in three sentences of loose impressions, the tool can dress that up into a polished paragraph that reads better than it deserves to. Employees increasingly notice this. Generic, AI-smoothed feedback that doesn't reflect their actual work has become its own minor complaint in workplace surveys over the past year.
This fits a broader pattern in small business software: AI is arriving first as a drafting assistant, not a decision-maker, in HR functions. The same thing happened with job postings and resume screening over the last three years — AI wrote the first draft, humans were supposed to check it, and enforcement of that "check it" step was uneven. Regulators have since stepped in on the hiring side, with New York City and Illinois both passing laws requiring disclosure or audits when AI plays a role in employment decisions. Performance reviews are the next logical place scrutiny could land, especially if AI-influenced feedback ends up tied to raises, promotions, or terminations.
For small business owners, the practical upside is real: AI drafting tools cut the time cost of writing thoughtful reviews, which is often the actual reason feedback gets skipped or phoned in. The risk is treating the AI output as finished rather than as a starting point. A review still needs manager-specific detail — a missed deadline, a client compliment, an actual project name — or it reads as boilerplate to the employee receiving it.
There's also a documentation risk worth flagging. If a review generated with AI assistance is ever cited in a wrongful termination or discrimination dispute, the manager should be able to show the underlying facts came from their own observation, not from the AI filling gaps with plausible-sounding language.
Watch for state and local rules on AI use in employment decisions to expand beyond hiring into performance management, and watch whether HR platforms start adding audit trails showing what a manager wrote versus what the AI generated.
The practical takeaway: AI can speed up the writing of a performance review, but it can't supply the specific, observed detail that makes feedback useful — that part still has to come from the manager who actually watched the work happen.