Paper cut-out robot marking script pages beside blank speech bubbles and an empty portrait frame, representing an AI advisor skill based on mentor feedback

Quick answer

To turn a mentor’s feedback into an AI advisor skill, collect their past comments on your work, ask the AI to pull out the recurring principles, write those into a skill file with examples and a clear review format, then test it on a script your mentor has already reviewed to see if it spots the same issues. Add new feedback every time you get it. Ask your mentor first, keep it private, and treat it as a practice partner, not a replacement for the real person.

Key takeaways

  • Principles, not quotes. The skill should capture why your mentor suggests changes, not just copy their words.
  • Test it against known feedback before you trust it.
  • Ask permission. It is their expertise; be open about what you are doing.
  • Keep updating it with each new review so it improves over time.

This guide follows on from How to Build a One-Person Media Machine With Grok Bot: Louise de Sadeleer’s Video Workflow. In that episode, Louise de Sadeleer explains to Greg Isenberg how she created a “Miranda feedback” skill from the script reviews her short-form course mentor gives her. After writing a script, she runs the skill to predict what Miranda would say. Here is how to build your own with Grok Bot, Claude, ChatGPT or any assistant that supports skills or custom instructions.

📘 Main article: See how Louise uses idea, hook and script advisor skills in How to Build a One-Person Media Machine With Grok Bot: Louise de Sadeleer’s Video Workflow.

What is an AI advisor skill?

An advisor skill is a saved set of instructions that makes your AI review your work from a particular point of view, such as a mentor, an editor or your ideal customer. Instead of a generic “make this better”, it gives feedback in a consistent, familiar way.

Why turn feedback into a skill?

Because good feedback is limited by your mentor’s time. A skill lets you get a first round of the same kind of feedback whenever you write, so the reviews you get from the real person can focus on bigger things. Louise also uses separate advisors for ideas, hooks and script structure.

Step 1: Collect the feedback you already have

Gather everything: comments on drafts, emails, call notes and voice memos (transcribed). Put each piece of feedback with the draft it refers to. Five to ten examples is a good start. Label each one: what you wrote, what they said, and what you changed.

Step 2: Ask permission and keep it private

Tell your mentor what you plan to do and why. Most will be flattered, but some may prefer you did not. Keep the skill for your own use, do not share or sell it, and do not present its output as their actual opinion.

⚠️ Warning: Do not upload private messages from other people, or paid course material, to an AI tool without checking the tool’s privacy settings and the course’s terms.

Step 3: Pull out the principles

Paste your examples into your AI tool and ask:

“Here are examples of feedback my mentor gave on my scripts. Identify the 5 to 10 principles behind their feedback. For each, give the principle in one sentence, what it looks like when I get it wrong, and one example from the feedback. Do not invent principles that are not supported by the examples.”

Read the list carefully and remove anything that does not sound like them.

Step 4: Write the skill

A simple skill file has four parts:

SectionWhat to include
Role“Review my script the way [mentor] would, based on the principles below.”
PrinciplesThe 5 to 10 principles from step 3, each with a short example
Review formatOverall verdict, top 3 issues in priority order, one rewrite suggestion per issue, what is working
Limits“If something is not covered by these principles, say so rather than guessing.”

In Grok Bot or Claude you can ask the assistant to create the skill for you: “Turn these principles into a skill called Mentor feedback.” For tips on keeping skills lean, see how to improve an AI skill without filling it with one-off rules.

Step 5: Test it against real feedback

Take a draft your mentor has already reviewed, but do not show the AI their comments. Run the skill and compare. If it catches most of the same points, it is working. If it misses something important, add or sharpen a principle. If it invents points your mentor would never make, tighten the limits section.

💡 Tip: Test with two or three different drafts. A skill that only works on one example is probably just copying that example.

Step 6: Keep it learning

Every time you get new feedback, add it to your examples and ask: “Does this new feedback fit an existing principle, or is it a new one?” Update the skill only when a point comes up more than once, so it stays focused.

Can you build an advisory board of several mentors?

Yes. Greg suggests an advisory board of several mentors, because different experts give different advice. You could have one skill per person and a final step that asks: “Where do they agree, and where do they disagree?” That diversity of views often leads to better decisions than a single voice.

Frequently asked questions

Can AI really predict what my mentor would say?

It can approximate their recurring advice if you give it enough real examples, but it will not match their judgement on new situations. Treat it as a first-pass reviewer.

How many feedback examples do I need?

Five to ten is enough to start. The skill improves as you add more.

Is it okay to create an AI skill based on someone else?

Ask them first, keep it for your own use, and never present its output as their real opinion.

What other advisor skills are useful for creators?

Louise de Sadeleer uses an idea advisor, a hook advisor and a script advisor alongside her mentor skill.

author avatar
Garry Knight
I'm Garry Knight, the person behind Prodify Digital. I write about email list building, email marketing, SEO, AI search and the tools that connect them. My aim is to make online marketing easier to understand, so creators and small business owners can make informed decisions about building an audience and keeping people engaged. Here you'll find straightforward guides and product reviews that explain what something does, where it fits and which limitations matter. The focus is on clear explanations and useful next steps—not hype, shortcuts or promises of easy earnings.

Leave a Reply