AI for Advisors newsletter
Here’s something most advisors don’t realize: Every major AI model includes a team of personas ready to give you professional, relevant, and actionable critiques about anything you’re creating with AI.
These “critics” could be a compliance reviewer, a skeptical client, or even a senior advisor who’s seen every client challenge imaginable. They’re all in there, ready to find what you missed and help you fix it. Most people never call on them because they simply don’t know to do it.
The current challenge
A typical advisor workflow looks something like this: You open a chat, describe what you need, AI delivers a decent working draft, you clean it up, and send it on its way.
That’s using a fraction of what AI can do. The hard part, the part that genuinely determines whether your work is good, is evaluation of your work. And the AI can do that, too. You just have to ask.
This is what power users figured out: You can turn the AI back on its own output. The same model that generated your draft can critique it—from any perspective you choose. Give it a role, a set of standards, a point of view that isn’t yours, and it will find problems you wouldn’t have thought to look for.
That client email you sent last night probably covered the basics. But did it flag the assumption you made about the client’s tax situation? Did it catch the paragraph that drifted into language no normal person would follow? Did it anticipate the question the client’s spouse is going to ask at dinner tonight—the one that unravels your carefully worded explanation?
One reflection pass, through the right lens, catches all of that before you hit send.
The core idea
This technique is sometimes called “reflection prompting.” It is a step you add to your existing process. Instead of asking AI to generate a final answer, you ask it to:
- Generate—produce the draft
- Critique—evaluate it against specific standards
- Improve—revise based on what the critique found
In the generate step, the AI writes as usual, optimizing for fluency, filling in the blanks, getting to the end.
In the critique step, it reverses direction. It reads what it just wrote and asks whether the claims hold up, whether the logic is tight, whether a reader would follow the argument or get lost halfway through.
In the improve step, it revises. Paragraphs get reorganized. Unsupported claims get qualified or removed. The gap between what you meant and what the draft says gets closed.
Add pressure to the critique
The basic reflection loop works. But you can make it significantly more powerful by sharpening the lens you’re critiquing through. Generic critique produces generic improvements. Specific perspectives produce specific, actionable findings.
The real leverage comes from assigning a specific role to the critique. You’re telling the AI who to be when it reviews—and each perspective surfaces a different category of problem.
Client-side lenses
“Critique this as the client’s spouse—someone who wasn’t in the meeting, is reading this email cold, and needs to understand it well enough to feel comfortable with the recommendation.”
Your email might make perfect sense to the person you met with. Their partner, reading it at the kitchen table with no context, is a completely different audience.
“Critique this as a high-net-worth client who has worked with multiple advisors, has heard every pitch, and has zero patience for vague reassurance. Flag anything that sounds like filler or lacks specificity.”
This one strips out the comfortable generalities advisors default to when they’re not sure what else to say.
Compliance lens
“Critique this as a highly experienced compliance expert reviewing this communication for suitability, balanced presentation of risks and benefits, and adequate disclosure. Flag any language that could be interpreted as a guarantee, a promise of specific outcomes, or a recommendation without sufficient qualification.”
Most advisors already think about compliance. Asking the AI to play that role explicitly turns a vague awareness into a specific checklist.
Professional peer lens
“Critique this as a CPA reviewing the tax implications. Are there assumptions about the client’s tax situation that haven’t been verified? Are there tax consequences mentioned without adequate qualification?”
Cross-disciplinary critique is something most advisors never get unless they have a CPA down the hall. This simulates that conversation.
Content and marketing lens
“Critique this as a prospective client who is comparison-shopping financial advisors and reading this content to decide whether you’re worth a phone call. What impression does this create? Does it differentiate you or could it have come from any advisor in the country?”
If you produce articles, newsletters, or any public-facing content, this is the lens that separates work that builds your reputation from work that fills a slot on your calendar.
Each lens reveals different problems. Used together, they work like a supercharged editorial team—one that’s available 24-7 and doesn’t charge by the hour.
How this article was critiqued
Before publishing, I ran this article through its own critique. Here’s the prompt:
“You are a senior editor with 30-years’ experience improving draft articles and who also teaches other advisors how to use AI. You’ve read hundreds of AI how-to articles. Critique this article for: (1) Does it make a clear, compelling case for why an advisor should change their current workflow? (2) Is there anything that feels padded, repetitive, or like filler? (3) Would a busy advisor finish reading this, or would they bail halfway through? (4) Does the reader walk away knowing exactly what to do and how to do it?”
If you want to see how well reflection prompting works, try running this article through that prompt yourself. Then try it on the next email, memo, or piece of content you create with AI.