Hey, welcome back!
And if you are new here, welcome to the SaviteckX family. Really glad you found us.
We are six weeks in and I am still learning as I go. Honestly, that is kind of the whole point.
This week was a good one. I am halfway through a book that I think is a hidden gem. The kind where you want to highlight every single page. I will tell you more about it soon, I promise.
I also did something fun this week. I put a friend's AI prompting skills to the test. She was confident. Very confident. Claude had other ideas. The results were eye opening. I cannot wait to tell you the full story in a future issue.
But first, let's get into today's AI skill.
Catching up? Here is where we have been:
Issue 01: Prompt Engineering - how to talk to AI so it actually listens
Issue 02: Taste Curation - learning to spot good AI output from bad
Issue 03: The Master Prompt - introducing yourself to AI properly
Issue 04: Output Iteration - shaping drafts until they feel right
Issue 05: System Prompts - turning your best work into reusable templates
Just click any link above if you missed one.
Here is something nobody really tells you about AI.
By default, it wants to make you happy. You show it your business plan and it says great. You share a pricing idea and it says sounds good. You pitch it something and it nods along, adds a bit of detail, and sends you off feeling confident.
That is not helpful. That is a yes-man.
A yes-man does not make you better. It just repeats your own thinking back to you in slightly nicer words. If your plan has a gap in it, a yes-man AI will not catch it. It will cheer you on and wave you off anyway.
The people getting real value from AI have figured out how to flip this.
Give AI a different job
Instead of asking AI to help you, ask it to challenge you.
Here is a prompt you can copy right now:
"Act as a devil's advocate. I am going to share [my plan / my idea / this email]. Your job is not to agree with me. Find every flaw, weak assumption, and risk I have not considered. Be direct."
That one instruction changes everything. AI stops looking for reasons your idea works and starts looking for reasons it might not. That is exactly what you want before you commit to something.
Three ways to use this
1. Stress-test before you launch
Before you put out a new offer or change your prices, ask AI to play sceptical customer.
"Here is my offer at this price. Act like someone who is not convinced. Tell me every reason you would not buy this."
You will get objections you had not thought of. Write them down. Then go fix them.
2. Find the real root of a problem
When something is not working and you cannot figure out why, ask AI to go deeper.
"This has been failing for three months. Do not give me quick fixes. Tell me what might be fundamentally wrong with my approach."
Surface level advice is everywhere. This gets underneath it.
3. Check your blind spots before a big decision
Before any important move, try this one:
"I am about to [decision]. What am I most likely not seeing? What would I regret not thinking about in six months?"
Blind spots are called blind spots because you cannot see them yourself. AI can look where you are not looking.
Three things to take away this week
One: Try the devil's advocate prompt on something this week. Pick one decision you are about to make and run it through.
Two: The more context AI has about you and your goals, the sharper its feedback will be. Always prime it before asking for hard feedback.
Three: Next week we are covering Context Compression. What happens when you give AI too much information? It gets confused. We fix that next time.
Harriet
Founder, SaviteckX
If this landed for you, the kindest thing you can do is forward it to one person who needs it.
No big ask. Just one person. That is how this community grows, one real conversation at a time. Thank you for being here.