Make a custom GPT that gives artifact feedback from a manager's explicit rubric and examples, while keeping the manager responsible for context, coaching, performance decisions, and any feedback that affects a person rather than the work itself.
Source episode · 03:00
How I AI: Hilary Gridley's Playbook for Scaling Yourself with Custom GPTswith Claire Vo
Hilary collects before and after slide examples, asks ChatGPT to articulate the differences and become far more specific, turns the resulting criteria into instructions for a custom slide evaluator, scores each criterion from one to five, and beta tests the GPT with a teammate before wider sharing.
A reusable feedback assistant that scores and explains artifact criteria, cites specific passages or elements, proposes actionable revisions, states uncertainty, and routes contextual or people related judgment back to the manager.
OpenAI conversational AI
Step by step
Follow the sequence once, then adapt the prompts, checks, and handoffs to your own setup.
5 steps
Collect permissioned examples of the same artifact type, including weak drafts and the manager's edits. Remove author identity and sensitive content, and label the context and why each change improved or harmed the work.
Ask AI to compare the examples and propose observable criteria. Challenge vague phrases and separate general quality, team convention, audience need, and personal preference.
Compare these paired before and after artifacts. Identify observable differences, the likely purpose of each edit, exceptions, and candidate criteria. Separate general communication quality, team conventions, audience specific needs, and subjective preference. Do not infer anything about the authors.
Turn the promising criteria into a specific rubric with definitions, positive and negative examples, evidence requirements, scoring anchors, and a not enough context outcome.
Make this rubric 100 times more specific. For each criterion define what it measures, why it matters, observable evidence, one through five anchors, examples, exceptions, and when to return NOT_ENOUGH_CONTEXT. Keep all judgment about the artifact, not the person.
Generate custom GPT instructions that ask for the artifact purpose and audience, score the rubric with citations, prioritize the most important revisions, and explain rather than simply rewrite.
Write instructions for a custom GPT that evaluates [artifact] using this rubric: [rubric]. It must request purpose and audience, cite exact evidence, score with anchors, explain why each issue matters, propose specific revisions, state uncertainty, and never assess employee performance, intent, or potential.
Test the GPT on held out examples with one teammate, compare its feedback with the manager's, and collect disagreements. Revise the rubric and examples before broader sharing, then review use and failure reports regularly.
Turn an idea into a PRD, user stories, and a plan.
After the steps
How to recover when the loop fails and where human judgment helps.
Recover
Keep building
Log your workouts effortlessly without ever opening a spreadsheet. This workflow uses the Claude mobile app to parse a simple voice note and automatically update your workout tracking spreadsheet in the background.
Create a custom email management interface within Claude to read, triage, and draft replies without opening the Gmail tab. This workflow helps you reclaim focus and compound your learnings by training your AI on every interaction.
Build a fully automated system in Claude to generate personalized, interactive HTML proposals and client onboarding materials, saving hours on administrative work and creating a high-touch experience.
Join 100,000+ product managers who use ChatPRD to write better docs, align teams faster, and build products users love.