How to get useful advice from an AI mentor

The quality of AI mentor advice depends on how you engage. Step-by-step: bring specific questions, share context, push back on generic responses, and commit to action.

May 21, 2024Updated Apr 21, 2026

Quick answer

You get useful advice from an AI mentor by doing the work of being a good advisee. That means bringing a specific question, sharing context the mentor cannot know, pushing back when the advice feels too general, asking for examples, and ending each session with one concrete next step. The mentor is only as useful as the material you give it to work with.

What you need before you start

Gurus installed on your Android phone
A specific situation or question in mind (not just a topic area)
Willingness to be honest about the context, including the parts that reflect poorly on you

Step 1: Bring a specific question rather than a broad topic

The fastest way to get generic advice from any mentor, human or AI, is to open with a vague topic.

"I want advice on my career" produces frameworks and generalities. "I have been offered a promotion that would mean managing a team for the first time. I am not sure whether to take it because I am not confident I am good at managing people, but I also know this kind of opportunity does not come often. What should I be thinking about?" produces a targeted, useful conversation.

Before each session, write down the specific question you want to answer. Not the general area, the actual question. If you cannot write it in a single sentence, the question is not specific enough yet.

Step 2: Share relevant context the mentor cannot know

The AI mentor does not know your background, your history with this goal, or the specific people and circumstances involved. It only knows what you tell it.

The more relevant context you share upfront, the more targeted the advice. Context worth sharing includes:

The specific situation in enough detail that someone unfamiliar could understand it
What you have already tried or considered
What the constraints are (time, relationships, resources, risk tolerance)
What outcome you are actually trying to achieve (this is often different from what you asked about)

A useful habit: before each session, spend 2 min writing a short brief. Situation, what you have tried, what you need help thinking through. Paste this into the opening message. You will get better responses than if you give the context piece by piece as the conversation unfolds.

Step 3: Push back when advice feels too generic

AI mentors sometimes default to applicable-to-everyone advice. This is safe but not particularly useful for your specific situation.

When the advice feels generic, say so directly: "That applies to most situations. What specifically would you advise given what I told you about my context?" or "I have already considered that approach. What would you suggest given that it did not work for me?"

Pushback forces the conversation to your actual situation rather than a version of it. A good mentor, human or AI, responds to specific challenge with more specific thinking. If the response is still generic after you have pushed back twice, share more context. The issue is usually insufficient information, not insufficient capability.

Step 4: Ask for examples or follow-up questions

Two prompts that consistently improve the quality of AI mentor advice:

"Can you give me a concrete example of what that would look like in practice?"

Abstract advice is hard to act on. An example makes it tangible. If the mentor says "be more direct in how you set expectations with your team," an example of what that sounds like in a real conversation is more useful than the principle alone.

"What questions should I be asking myself that I have not asked yet?"

This prompt surfaces the blind spots in your framing. The mentor may have noticed something in your description that you did not flag as important. This question invites it to surface that, rather than waiting to be asked about it explicitly.

Both prompts work at any point in the conversation and tend to unlock the most useful turns in a session.

Step 5: End each session with a concrete next step

The difference between a productive mentor session and an interesting conversation is what you do afterward.

Before you close Gurus, explicitly ask: "Based on this conversation, what is the one thing I should do before we talk again?" Then write down the answer.

The commitment should be:

Specific (not "work on my communication" but "have one direct conversation with my manager about the timeline by Thursday")
Time-bound (with a deadline you name)
Within your control (not dependent on someone else doing something first)

Write it somewhere you will see it. Review it before your next session. Report back honestly on whether you did it.

The sessions themselves are not the output. The behaviour between sessions is.

Common problems

Advice is too generic (fix: be more specific in your opening). If three sessions in a row have produced generic advice, the input is too vague. Go back to Step 1. Write the specific question. Add more context, especially the parts that feel embarrassing or complicated. The AI cannot infer what you have not shared. If you find it hard to be specific, that difficulty itself is worth exploring: what is making it hard to name the actual question?

Over-relying on the AI (fix: set a session limit). If you find yourself reaching for the AI mentor before you have spent any time thinking yourself, or having multiple sessions per day, the interaction may be serving as reassurance-seeking rather than genuine development. Set a limit: two or three sessions per week on a specific goal. Between sessions, sit with your questions rather than immediately externalising them. The thinking you do alone is part of the process.

Not acting on advice (fix: write down one commitment). This is the most common failure mode. You have a good session, feel clear and motivated, and then do nothing differently. The fix is structural: always end with a written commitment. If you review your last three sessions and find no commitments made and no change in behaviour, the sessions are pleasant but not functional. Raise this directly with the mentor in the next session: "I notice I keep having good conversations but not changing anything. What do you think is happening?"

FAQ

What if the AI mentor gives advice I know is wrong for my situation? Say so. "That does not fit my situation because..." is a legitimate response. The subsequent conversation often produces more useful thinking than the original advice. If you never disagree with the AI mentor, you are probably receiving too much validation and not enough challenge.

How do I get the AI to be more direct with me? Ask for it explicitly: "I want you to be direct with me and not soften the feedback. What are you actually observing about my situation?" Most AI mentors will increase directness when asked. If you want ongoing directness throughout a session, state this at the beginning: "Please be direct and challenge my assumptions rather than just supporting what I say."

Should I keep a record of my sessions? Yes. Keep brief notes after each session: the main insight, the commitment you made, and the outcome you reported at the start of the next session. Over weeks, this record shows you your own patterns: the rationalisations you keep using, the commitments you keep not keeping, the type of question that produces the most useful responses for you. That pattern data is valuable in itself.