The AI Shift Most Advisors Haven’t Made

AI for Advisors newsletter

Most of the conversations I have with advisors about AI start in a familiar place: Writing better emails. Cleaning up client communications. Drafting social posts or articles faster.

That makes sense. For many people, those are the most visible and immediate benefits of tools like ChatGPT.

And to be clear, that use case isn’t wrong. But it quietly reinforces an earlier mental model, where AI is primarily a helpful, efficient assistant, but one limited to producing better sentences on demand.

That framing, more than any lack of effort or interest, is what’s holding most advisors back from fully experiencing the power of AI.

Even advisors who use AI daily often think of it as advanced word completion. It speeds up tasks. It reduces friction in communication. It gets drafts on the page more quickly.

That mental model made sense early on. The first widely used versions of these tools were optimized for fast, single-pass responses. You asked a question. You got an answer. The interaction ended.

But that’s no longer the system you’re actually using.

What quietly changed

In late 2024, ChatGPT creator OpenAI made a meaningful shift with the introduction of its first reasoning models for users, starting with its “o1 reasoning model.” These models were explicitly designed to “think” through multi-step chains of logic before answering, rather than immediately generate a response.

That’s a fundamentally different capability than earlier flagship models like GPT-4o, which were optimized for speed and fluency in a single pass. Fast models are excellent for writing, summarizing, and everyday communication. Reasoning models trade a bit of speed for depth, making them far stronger at complex, multi-step problem solving.

The label matters less than the implication.

AI is no longer confined to producing an answer to a prompt. It can now participate in an extended reasoning process. Once you understand that, the role of AI changes.

From writing tool to thinking partner

The difference shows up clearly in how advisors interact with AI. You might ask, “Help me write an email about portfolio rebalancing,” receive a competent draft, make a few edits, and move on. That interaction ends as soon as the output is delivered.

But something very different happens when the starting point changes. Instead of asking for words, you might prompt something like:

At that point, the interaction becomes exploratory rather than transactional. You aren’t asking AI to produce something. You’re asking it to help them reason through a situation.

The value is no longer the final paragraph. The value is the clarity that emerges before the paragraph is ever written.

That experience of “thinking with AI rather than prompting AI” is the shift many advisors haven’t yet experienced.

The ‘AI Usage Ladder’ reframed

This is where the AI Usage Ladder becomes useful. But not as a hierarchy, and not as a judgment. The ladder is best understood as a set of “modes,” different ways of using AI depending on what the moment requires.

Lower rungs tend to be task driven: Draft this. Improve the wording. Summarize this. These uses are fast, bounded, and output oriented. They show up constantly throughout the day, and they should.

Higher rungs are situation driven. They show up when something is unclear, complex, or consequential such as a client decision, a strategic choice, a team conflict issue or new business challenge. These interactions are longer, more open-ended, and oriented toward reasoning rather than output.

But it’s important to understand that a real power user is not someone who lives at the top of the ladder. A real power user is someone who can move up and down the ladder deliberately.

Ladder mobility and mode switching

Understanding how to use AI at the strategic, decision-making top of the ladder does not mean that’s all you do. In practice, real-world AI use involves constant “mode switching.”

You might use a fast, output-oriented mode to draft an email. Then move up the ladder to think through the underlying message or client concern. Then move back down to articulate it clearly. Later in the day, you may return to the top of the ladder to reason through a business decision, a staffing issue, or a client scenario with no obvious answer.

Not every moment requires deep co-thinking with AI. But most advisors, especially those managing clients, teams, and growing firms, face ongoing challenges that genuinely benefit from spending time at the top of the ladder.

The value comes from moving up and down the ladder, not staying “parked” in one spot.

You don’t need to wait for a perfect strategic question to test this. In fact, the best way to understand what top-of-ladder thinking feels like is to bring a real-world challenge you’re facing right now—big or small—and explore it in conversation with AI.

Try a prompt like this:

The exact wording doesn’t matter. Just pick a situation you’ve been wrestling with—something open-ended, with no obvious solution—and give it five to 10 minutes of thoughtful dialogue. The point isn’t to get an answer. It’s to improve your thinking.

Here are a few real examples other advisors have shared that work well for this kind of exploratory conversation:

This kind of thinking session may feel unfamiliar at first. But once you experience this even once, the capability becomes obvious. And once it’s obvious, it becomes something you can return to repeatedly when the situation calls for it.

Ready to make the leap? Horsesmouth’s AI for Advisors Pro training programs provide the structured, advisor-specific approach that transforms occasional users into confident practitioners. Learn more at www.horsesmouth.com/aipro.