AI Can Now Help Explain Why You Deserve a Raise.

You did the work. Now comes the separate assignment of describing it in a way that sounds like you did the work exceptionally well.

The annual self-review asks employees to become the historian, publicist and management consultant of their own calendar. A year of decisions and unfinished projects must fit into a few boxes, preferably with measurable impact and evidence of growth.

According to James Faris’s September 28 reporting for Business Insider, Disney encourages approved tools such as Microsoft Copilot for review preparation: brainstorming goals, translating past feedback into development objectives and refining self-assessments. Employees retain responsibility for their thinking and for agreeing on priorities with their manager.

That is a reasonable use of a writing assistant. It also raises a question that extends well beyond one company: when the writing gets easier for everyone, what does the review actually tell us?

A better sentence is a useful service

There is no particular virtue in spending an evening trying to remember what happened in February. Nor should the ability to write fluent corporate prose determine whether someone gets credit for solving a difficult problem.

An assistant can help turn scattered notes into an outline, expose a vague objective or suggest a clearer description of a result. For someone who finds self-promotion uncomfortable, the blank box can be a substantial obstacle. Removing it could make a review more useful.

Consider a hypothetical employee who resolved a recurring billing error. Their first draft says they helped finance. A better account identifies the error, explains their contribution and records what changed. Software can help organize that explanation. The evidence still has to come from somewhere.

The distinction matters because a confident sentence can travel farther than the facts supporting it. An impressive verb does not establish who made the decision, how much help they had or whether the fix lasted beyond the reporting period.

If AI removes a writing disadvantage, that is progress. If it merely gives everyone the same impressive vocabulary, the evaluator needs a better way to distinguish the underlying work.

The review can become a conversation between summaries

Now imagine the next steps. An employee uses AI to draft goals. Months later, an assistant helps describe progress against them. A manager summarizes those descriptions with another tool. An HR system organizes the resulting information into fields that can be compared across a team.

This is a possible evaluation chain, not a claim that Disney has automated ratings, promotions or pay decisions. The reporting about review preparation does not establish any of those things.

Each individual step can look sensible. Together, they can put distance between an actual event and the person making a consequential judgment about it. Qualifications disappear. A shared result becomes an individual achievement. A difficult assignment gets reduced to an incomplete target.

The danger is especially clear when a manager reads only the summary. An employee who prevented a bad launch may have less visible output than someone who shipped a routine update. A concise comparison could make the second contribution look stronger simply because it fits the form.

If AI helps write the goals, the self-review and eventually the manager feedback, what exactly are we evaluating? Ideally, the decisions and results beneath the prose. That requires someone to look beneath it.

Make the evidence easier to inspect

A useful review process would ask for a small number of specific examples: what was expected, what happened, what the employee contributed and what circumstances changed. Links to work, dates and concrete outcomes give the conversation something sturdier than adjectives.

Goals need the same discipline. A well-written objective can still describe work the team does not need, depend on resources nobody has approved or measure something the employee cannot control. Agreement with a manager must happen before elegant phrasing hardens into an obligation.

Managers also need to retain the awkward parts of the job. They have to ask follow-up questions, reconcile conflicting accounts and explain why one result mattered more than another. Software can prepare the materials for that conversation. It cannot make the manager’s explanation unnecessary.

There is an opportunity here to spend less time performing diligence through prose and more time discussing the work itself. That would mean judging the quality of an example rather than the confidence of its introduction, and treating a missing fact as a reason to ask rather than an invitation to generate.

The self-review may soon be the easiest document anyone writes all year. Someone still has to explain the rating in the meeting.