15Five AI-Assisted Reviews

Most reviews run on memory.
Great ones run on evidence.

Reviews built the traditional way often run into problems.

Recency Bias

Reviews written from memory lean on the last few weeks and the moments that stuck. Strong work from early in the year drops out of the picture.

Deadline Cram

Pulling together months of work and turning it into a good review under a tight deadline makes quality slip.

Generic Feedback

Even strong managers can struggle to put good feedback into words. When feedback comes out vague, the employee leaves without knowing how to improve.

Filed & Forgotten

The review gets filed, and nothing happens next. When there's no plan or follow-through, the same gaps show up again next cycle.

Most reviews run on memory.
Great ones run on evidence.

Reviews built the traditional way often run into problems.

Recency bias

Reviews written from memory lean on the last few weeks and the moments that stuck. Strong work from early in the year drops out of the picture.

Deadline cram

Pulling together months of work and turning it into a good review under a tight deadline makes quality slip.

Generic feedback

Even strong managers can struggle to put good feedback into words. When feedback comes out vague, the employee leaves without knowing how to improve.

Filed and forgotten

The review gets filed, and nothing happens next. When there's no plan or follow-through, the same gaps show up again next cycle.

AI handles the drafting and the prep.
The judgment stays human.

AI-Assisted Reviews creates the first draft of a performance review from a year of someone's real work in 15Five and their connected work tools, while the review writer decides what evidence it draws on, edits the draft, chooses every rating, and submits when they're ready.

Feedback Grounded in Evidence, Not Memory

AI-Assisted Reviews creates the first draft of a performance review from what someone actually did during the review period. Every draft is built from goals, feedback, 1:1s, past reviews, and work from connected tools like Slack, Microsoft Teams, Jira, GitHub, and Google Drive. Every claim traces back to its source.

Your Standard for Good Feedback, in Every Review

The company context you set and the prompt instructions you write shape every draft, across every team. Your definition of good feedback shows up in the reviews themselves, cycle after cycle.

A Review Cycle that Compounds

Once both reviews are in, the manager gets a conversation guide for the meeting. From there, an individual development plan can be created in Growth Studio, and it becomes context for the next cycle. Each review makes the next one sharper, so the cycle turns into a continuous loop instead of an annual reset.

How It Works

1. Start with answers drafted from evidence

Every open-ended question comes back with a draft built from real work, gathered as it happened. Connecting a tool backfills up to a year of history. The evidence behind each answer is visible and labeled with its source.

2. Shape the draft your way

Adjust the context a draft is built from and generate it again, edit by hand, or revise through the chat by asking for a shorter answer, a sharper focus on one project, or a version restructured into bullets. Rating-scale questions, judgement, and accountability stay with the writer.

3. Turn the review into a plan

Once the reviews are submitted, the manager gets a conversation guide. An individual development plan carries over into the next cycle.

Same Permissions. Same Protection.

AI-Assisted Reviews runs on 15Five's existing data governance and permission framework, and your data stays secure. Drafts draw only on information the reviewer already had access to, and nothing more.

15Five holds SOC 2 Type II, SOC 3, and TX-RAMP Level 2 authorization, and meets CCPA, GDPR, and PIPEDA. Data is encrypted, access stays restricted by role, and every subprocessor that touches your data is named and open for review before you connect anything.