Performance Reviews and Feedback

The Review Everyone Dreads

It’s performance review season. Managers are staring at blank forms, trying to remember what happened over the past six months. They know they should give specific feedback, but the details are fuzzy. So they write generic statements: “Great team player.” “Meets expectations.” “Could improve communication.”

Meanwhile, employees are anxious. They want to know where they stand, but they’re bracing for vague feedback that doesn’t tell them anything useful. The whole process feels like a formality rather than a genuine development conversation.

This is broken. And AI can help fix it – not by writing reviews for you, but by helping you turn messy notes and observations into structured, specific, actionable feedback.

What You’ll Learn

By the end of this lesson, you’ll write performance reviews that are specific, balanced, and genuinely helpful. You’ll use AI to structure feedback, identify patterns, and ensure consistency across your reviews – while keeping the substance deeply personal and accurate.

From Onboarding to Performance

In Lesson 5, you built 30-60-90 day plans with clear milestones. Those milestones become the foundation for performance evaluation. When you’ve defined what success looks like upfront, reviewing performance becomes a matter of comparing actual results to expected outcomes. The best performance reviews start on day one.

Why Most Reviews Fail

Performance reviews fail for predictable reasons:

Vague language. “Good communicator” tells an employee nothing. Good at what? With whom? In what context? How could they be even better?

Recency bias. Managers remember the last 4-6 weeks clearly and the rest of the year vaguely. The employee who had a rough January but crushed October looks great. The one who had a stellar spring but a quiet fall doesn’t.

Inconsistency. Manager A rates everyone highly because they want to be liked. Manager B rates everyone average because they think that’s fair. Manager C uses the full scale thoughtfully. Cross-team comparisons become meaningless.

Avoided conversations. Constructive feedback gets softened to the point of meaninglessness. “You might want to consider possibly looking into ways to enhance your time management skills” translates to: “I’m not brave enough to tell you that you’re consistently missing deadlines and it’s affecting the team.”

Quick Check

Pull up a performance review you’ve written or received recently. Count the number of specific, observable examples mentioned versus general characterizations. If the ratio tilts toward generalities, this lesson will help.

The Performance Review Framework

Every review comment should follow this structure:

Observation: What the employee specifically did (behavior or outcome) Impact: What effect it had on the team, project, or organization Assessment: How this compares to expectations for their role level Direction: What they should continue, start, or stop doing

Example – Strong performance:

Observation: Led the Q3 customer migration project, coordinating across engineering, customer success, and product teams. Delivered two weeks ahead of schedule with zero customer-reported issues.

Impact: The early delivery allowed the sales team to reference the migration in Q4 pipeline conversations, directly contributing to three new enterprise deals.

Assessment: Exceeds expectations for a mid-level project manager. Demonstrated senior-level stakeholder management and risk anticipation.

Direction: Continue owning cross-functional initiatives. Consider mentoring a junior PM on your next project to develop leadership skills.

Example – Needs improvement:

Observation: Missed sprint deadlines on 4 of 6 sprints in Q3. When deadlines were at risk, stakeholders were not notified until the day of the deadline.

Impact: Downstream teams had to scramble to adjust their timelines, and two client deliverables were delayed by a week each.

Assessment: Below expectations for timeline management and stakeholder communication at this level.

Direction: Flag at-risk deadlines at least 3 days in advance. Let’s establish a weekly check-in to review sprint progress and identify blockers early.

Notice: no personality judgments. No vague character assessments. Just behavior, impact, assessment, and direction.

Using AI to Structure Reviews

AI is useful for turning raw notes into structured reviews. The key is starting with your real observations, not asking AI to generate them.

Step 1: Gather your raw notes.

Throughout the review period, collect:

• Specific accomplishments and deliverables
• Feedback from peers and cross-functional partners
• Challenges or struggles you observed
• Goals from the last review and whether they were met
• Notable behaviors (positive or negative)

Step 2: Feed the notes to AI for structuring.

I'm writing a performance review for a [Job Title] who
reports to me. Here are my raw notes and observations
from the past 6 months:

[Paste your bullet points, notes, and observations]

Help me organize this into a structured performance review
using this format for each key point:
- Observation: What they specifically did
- Impact: The effect on team/organization
- Assessment: How this compares to expectations for their level
- Direction: What to continue, start, or stop

Also:
- Identify 2-3 key strengths with specific evidence
- Identify 1-2 development areas with specific evidence
- Suggest 2-3 goals for the next review period
- Flag any language that might be gendered or biased

Tone: Direct, respectful, specific. No corporate jargon
or vague generalities.

Step 3: Review for accuracy and bias.

AI structured your thoughts – now verify:

• Does every statement reflect what actually happened?
• Is the language specific enough that someone else could verify each claim?
• Is constructive feedback framed around behavior and impact, not personality?
• Would you be comfortable if this review were audited for fairness?

Quick Check

Look at the constructive feedback section of a recent review you’ve written. Does it describe specific behavior and measurable impact, or does it use personality-based language like “lacks initiative” or “poor attitude”? The former is actionable; the latter is subjective and legally risky.

Checking for Bias in Reviews

Research on performance reviews reveals persistent patterns:

• Women receive more feedback about communication style and personality, while men receive more feedback about technical skills and strategic thinking
• Women are more likely to be described as “collaborative” and “supportive,” while men are described as “visionary” and “ambitious” – even when their contributions are similar
• Underrepresented employees often receive shorter, less specific reviews

Bias check prompt:

Review this performance review text for potential bias:

[Paste review text]

Check for:
- Gendered language patterns (different words used for
  similar contributions based on demographics)
- Specificity gap (are some employees getting more
  detailed feedback than others?)
- Personality vs. behavior language (feedback should
  describe actions, not character)
- Double standards (behaviors praised in some employees
  but criticized in others)

Suggest specific rewrites for any flagged language.

Self-Review Prompts

Many organizations include employee self-assessments. Help your employees write better self-reviews by providing guided prompts:

Write a self-assessment for the past 6 months.

For each section, provide specific examples:

1. Key accomplishments (2-3)
   - What you did
   - Why it mattered
   - What you learned

2. Challenges you navigated (1-2)
   - What was difficult
   - How you handled it
   - What you'd do differently

3. Growth areas (1-2)
   - Where you want to develop
   - Specific actions you'll take
   - Support you need from your manager

4. Goals for next period (2-3)
   - Specific, measurable goals
   - How you'll track progress
   - Resources or support needed

Delivering Difficult Feedback

Writing the review is half the battle. Delivering it is the other half. Here are frameworks for the hardest conversations:

For underperformance:

1. State the specific behavior or pattern you’ve observed
2. Explain the impact on the team and organization
3. Listen to their perspective (there may be context you’re missing)
4. Collaboratively create a specific improvement plan with timelines
5. Be clear about consequences if improvement doesn’t happen

For a strong performer you’re concerned about losing:

1. Lead with specific recognition of their contributions
2. Share your assessment of their trajectory and potential
3. Ask what they need to stay engaged and growing
4. Discuss concrete development opportunities
5. Follow through on what you discussed

Exercise: Rewrite a Vague Review

Transform these vague review statements into specific, actionable feedback using the Observation-Impact-Assessment-Direction framework:

1. “Sarah has a great attitude and is a real team player.”
2. “Mike needs to improve his communication skills.”
3. “Lisa exceeds expectations in all areas.”

Key Takeaways

• Specific, observable examples beat vague characterizations every time
• Use the Observation-Impact-Assessment-Direction framework for every review point
• Start with your real notes and observations, then use AI to structure them – never ask AI to make up the content
• Check every review for gendered language, personality-based judgments, and specificity gaps
• Constructive feedback describes behavior and impact, not character
• Performance reviews work best when goals and expectations are set clearly from the start

Next lesson: beyond individual reviews, you need clear policies, consistent communications, and efficient workflows. Let’s build those.