For years, I’ve been consulting mostly quietly, in the background. For startups, SMEs, and high-net-worth individuals. I genuinely enjoyed it. The cases, the challenges, the kind of thinking-on-your-feet problem-solving that you can’t really replicate anywhere else.

Then, somewhere around 2022, the work dried up. Not because I stepped away. Because the demand almost completely disappeared. I wasn’t mad at all, because I had other things that require my immediate attention.

What Killed the Demand

The easy answer is AI. And it’s not wrong.

Over the last three to four years, a clear mindset shift has happened in certain markets. Especially among startup founders and indie hackers. The logic went something like:

“Why pay for billable hours when Claude or ChatGPT costs $20 a month and can answer the same questions?”

And honestly, who can blame them? When you’re running on a near-zero runway, every dollar matters. I ran a coaching program for a while.

Priced intentionally low, for founders in exactly that position. Even at $99 to $249 per month, 90% of the audience couldn’t justify it. What feels cheap on one side can feel crushing on the other.

Fast forward to 2026, and those same people are now calling $20 for an AI subscription too expensive. The expectation floor just keeps dropping.

The Old Model Doesn’t Work Anymore

The traditional consulting model has a few structural problems that AI didn’t create but made impossible to ignore.

First, the speed gap. AI can now compress a week of desk research into a few hours and produce detailed, structured reports on demand. That used to be a significant chunk of what consultants were charging for. That value proposition is gone.

Second, rigid frameworks. A lot of traditional consultants are still operating on playbooks built for a different decade. The market has changed faster than most consulting methodologies have.

If the deliverable is a slide deck with a SWOT analysis and a generic 90-day roadmap, there’s a real question of what the human is adding.

Third, the shift toward outcome-based pricing. Client expectations have moved hard toward performance fees and results-only models.

The hourly billing structure that sustained consulting for decades feels increasingly out of place when clients want to pay for what actually changes, not for the time spent changing it.

And then there’s the most direct signal from the industry itself: McKinsey replaced over 5,000 consultants with more than 12,000 AI agents in 2025. That’s not a footnote. That’s the direction of travel written in bold.

It’s Not Over If You’re Willing to Rebuild

Here’s where I push back slightly on the doomsday framing: traditional consulting is dying, but consulting itself isn’t.

What dies is the generic model: Broad-scope engagements, generalist advice, slow turnaround, opaque pricing.

What survives, and what I think will actually become more valuable, is the hybrid model. AI is doing the research, synthesis, and first-draft thinking.

Human consultants doing the interpretation, context-reading, relationship work, and quality oversight that AI is still genuinely bad at.

The path forward for consultants and firms is to narrow and go deep. Not startup consulting. Not founder advisory. Something hyper-specific.

A particular industry vertical, a particular type of problem, a particular stage of a company, where domain expertise and pattern recognition built over years actually justifies the fee. The more specific, the harder to replicate.

The consultants who survive this will be the ones who stop treating AI as a research assistant and start treating it as a co-delivery model. That’s a very different relationship with the tools.

I Miss It, Honestly

I do. And that surprised me a little when I sat with it.

Right now, I spend most of my time building micro-SaaS products, Ryoka’s workflow, and content creation for Story Brew and my podcast journal. All of it solo. Claude is my main collaborator. Which works.

But some days I go 12 hours without speaking a single word to another person. That loneliness isn’t dramatic; it’s just there, sitting quietly in the background.

Consulting gave me something I didn’t realize I’d miss this much: real conversations. Proper back-and-forth with real people working through real problems.

The occasional Zoom call that doesn’t feel like a meeting because it’s an interesting problem you’re both trying to solve together.

That’s not something a language model gives you. And I’ve started to feel that absence more lately.

What I’m Going to Do About It

I’m going to try it again. Small, one or two engagements per month, structured as experiments more than as a business. I want to dry-run different delivery models, test what still resonates with clients, and figure out what a 2026 version of my consulting approach looks like.

This isn’t about the revenue. It’s partly about the interaction, and partly because I’m curious. I’ve changed a lot in how I work and think over the last few years. I want to see if the consulting version of me has too.

If it works, it works. If it doesn’t, I’ll have learned something. That’s a reasonable trade.