I've done a terrible job of getting to know you...

After more than a year of sending this newsletter (admittedly, sometimes sporadically), I realize I know almost nothing about who's reading it or what you actually need. That's on me.

So I took a course from Jason Resnick, who helps service providers build better client relationships through strategic email sequences. Jason's approach is refreshingly direct: Stop guessing what people want and start asking them.

New subscribers will get a four-email welcome sequence designed to understand their biggest challenges. But I will be launching it for all of you first. I hope you'll respond so I can better understand what will actually resonate with you.

Speaking of getting real feedback instead of polite assumptions...

“ONE LAST THING…”

You know Lieutenant Columbo’s signature move, right? Just when the suspect thinks they’re in the clear, he turns back with that rumpled raincoat and says, “Oh, one last thing…”

That’s exactly what I’ve started doing with AI. And it’s transformed how I work.

The Real Problem Isn’t What You Think

Everyone complains that AI is “too nice.” But that’s not the real issue.

The real problem? We’re asking terrible questions.

When you prompt AI with “What do you think of this draft?” or “How can I improve this strategy?” you’re basically asking for a participation trophy. Of course you get generic cheerleading.

But when you ask a sharp, specific question, you get sharp, specific answers.

Before and After: Watch This Work

Here’s what happened when I was trying to define my ghostwriting niche. My detailed prompt came back with vague advice about targeting “senior executives who need sophisticated thought leadership.” Essentially, everyone with a budget.

So I pushed back, got slightly better advice about “mid-sized financial services firms,” then realized I was still getting mediocre output because I was asking mediocre questions.

Then I tried this:

“Step back and critique this work as if you were a tough editor with high standards. What assumptions are weak? Where is the logic flawed? What sections drag? What would make someone stop reading? Be specific about what needs to change.”

Here’s the response I got:

You’re 65, with significant student debt and insufficient retirement savings. You don’t have time to “find your niche.” You need to get ruthlessly focused on what pays the most, fastest.

You have one massive competitive advantage most ghostwriters will never have: You can write content that sounds like it came from a bank C-suite because you actually sat in those rooms for 17 years.

...Your menu of services is precisely the problem. You have 30+ different services across 6+ different target markets. That's not a business strategy; that's a buffet. And buffets make everyone mediocre at everything.

That stung. It was also exactly what I needed to hear.

I was working on a brand positioning document for a 50+ client and long-time friend. I was pretty happy with where we were, but because he was a friend, I wanted to make sure I wasn't pulling my punches. So I plugged in the prompt and here's what I got:

"David is facing a perfect storm of job search obstacles, which explains why he's struggling to get traction. Here's the unvarnished truth:

It went on to list a number of red flags with multiple sub-bullets, including (1) age bias is real and working against him; (2) geographic limitation is crippling; (3) his current positioning is confused and outdated; (4) industry consolidation has eliminating his target roles; (5) his skill set appears dated to modern employers; and so on.

I wrestled with sending it to him, but he emailed back that he appreciated the pushback. He's been stalling on repositioning his brand, and I believe I finally have his attention. We're planning to talk this week about reframing what I had already written.

Why This Actually Works

AI systems are trained on massive datasets of human feedback that reward helpfulness and politeness. When you ask for generic feedback, you get the averaged-out response the system thinks you want to hear.

But when you explicitly request criticism using editorial language, you’re accessing a different part of the training data. You’re pulling from tough editors, brutal reviewers, and honest critics rather than supportive coaches.

Your Two Essential Prompts

For any written work: Use the prompt above.

For strategy or planning: “What’s the biggest flaw in this approach? What am I not considering? Where would this likely fail in the real world?”

That’s it. Two prompts that will get you more useful feedback than a dozen polite requests. Or use the more detailed one I provide earlier.

The Uncomfortable Truth

Your AI assistant isn’t your friend. It’s your tool.

The best tools give you honest assessment, not emotional support. When you start treating AI like a sharp-eyed editor instead of a supportive cheerleader, your work gets demonstrably better.

I’ve used this technique to improve client proposals, strategic plans, and yes, this very newsletter. The feedback isn’t always pleasant, but it’s always useful.

Your Turn

Next time you’re working with AI, try the Columbo approach. Get that initial response, then turn back and say:

“One last thing… now tell me what’s actually wrong with this.”

You might be surprised by what you learn. You’ll definitely be stung by some of it.

But that sting means you’re getting real feedback that can actually make your work better.

And isn’t that what we’re really after?

What’s your experience with AI feedback? Hit reply and tell me whether you’ve found ways to get more honest assessment from your digital tools.

FOUR GREAT POSTS FROM FOUR SMART PEOPLE

1. Gini Dietrich offers advice on doing a smart mid-year check-in on your communications plan.

2. Carla Johnson offers a case study about the long wait for baggage at Houston's George Bush Intercontinental Airport and the passenger complaints. The airport reduced the wait time but not the complaints, leading her to ask "Are You Solving the Wrong Problem?"​

3. Andy Crestodina gives you a fun behind-the-scenes look at the camera, mics, and lights for 21 top marketers.

4. Are you or your company truly differentiated from the competition? David C. Baker ponders the message you’re sending with your positioning choices.

TWEETS THAT ACTUALLY RULE

MATT GRAY @matt_gray_ The 1-3-5 Rule: Each morning, identify: • 1 big mission-critical task • 3 medium-impact tasks • 5 small maintenance tasks This system eliminated my overwhelm and doubled my productive output in weeks. Stop trying to do everything. Start doing what matters. 2:8 PM • Jun 14, 2025 83 Retweets 792 Likes Read 99 replies

​

I welcome your comments or suggestions for future issues. Drop me a note here. If you found this on LinkedIn or had it forwarded to you, you can subscribe by clicking the button below.