The AI Plan That Didn’t Survive Reality

Ron Smith isn’t a “growth hacker.”

He’s a founder in geotechnical consulting in a mid-sized city in the Southeast. His work lives in the unsexy middle of real projects: soil conditions, foundation risk, schedule pressure.

The goal isn’t virality. The goal is credible work delivered on time so a project keeps moving.

Most of Ron’s business comes the way professional services still work:

• General contractors who call when they need answers fast

• Civil and structural teams who don’t want surprises

• Architects and developers who remember who made their life easier the last time

Referrals. Repeat relationships. Reputation.

And then one Monday, the steadiness didn’t feel steady.

Revenue wasn’t falling off a cliff. It was just… flat. Softer than it should be. The kind of softness that doesn’t create panic, but does create a low-grade pressure you carry into every meeting.

There was also a practical ceiling nobody liked naming out loud.

Ron’s team could handle only about five concurrent site investigations before turnaround started slipping. And in geotech, turnaround isn’t a nice-to-have. It’s the difference between “they call you again” and “they quietly stop.”

Ron had a number in his head: 25% revenue growth next quarter.

Not because he’d modeled it perfectly. Because it felt like the amount of momentum he needed to feel this year—like the amount that would quiet the thought he didn’t like admitting:

Are we building something… or just staying busy?

So he reached for certainty.

He opened ChatGPT and typed:

“Create a marketing plan to grow revenue by 25% next quarter.”

Minimal context. No baseline numbers. He didn’t even mention the five-investigation ceiling—or what slips when they exceed it. No detail on how buyers actually choose a geotech firm. No mention that this business doesn’t run on funnels.

ChatGPT produced a plan. It was organized. Confident. Multi-channel. It even suggested adjacent services Ron didn’t fully understand or feel equipped to sell.

And Ron felt the relief that comes from a plan that sounds authoritative.

His inner assumption clicked on:

It’s AI… it’s like an MIT PhD in a box. It has to be cutting edge and great, right?

He didn’t notice the trade he was making. He just noticed how good it felt to have something.

By Tuesday, he’d cleaned the output into a “playbook.” Enough structure to look decisive. Enough polish to sound like leadership.

Wednesday: The Meeting

Wednesday morning, Ron called his team in—the core group who would actually have to execute whatever he was about to bless.

Conference table. Laptops open. Coffee. That familiar, quiet question in the room: Where are we going with this?

Ron pulled up the plan.

“Alright,” he said, trying to sound certain. “Here’s the playbook. Let’s execute this.”

For a few minutes, it seemed fine. The plan had all the right words. The kind of language that makes people nod because it feels like a serious strategy.

Then someone asked the question that always exposes the difference between a plan and your plan.

“Who exactly are we targeting?”

Ron opened his mouth.

Nothing clean came out.

He glanced back at the doc like the answer might be hiding in the paragraphs. He said something vague about “construction partners” and “stakeholders,” but even as he said it, he could hear how slippery it sounded.

Another question followed—calm, reasonable, deadly:

“Why this channel first?”

Ron started answering with borrowed language.

“Because it’ll increase top-of-funnel volume and diversify acquisition—”

One of his team members blinked.

“When you say ‘channel’… what do you mean?”

That one landed.

Not rude. Not dramatic. Just clarifying.

A channel is just the path work comes in—referrals, repeat clients, partners, inbound calls, content, ads. The way someone ends up choosing you.

And Ron felt it: he was speaking in someone else’s vocabulary. Borrowed language, borrowed certainty. Like he’d put on a suit that didn’t fit and hoped nobody would notice.

More questions came:

“What are we actually offering?”

“Who owns this next week?”

Ron could feel the shift. Not hostility. Something worse: quiet uncertainty.

Because his team wasn’t asking philosophical questions. They were asking execution questions. And Ron couldn’t answer cleanly, because he didn’t actually understand the plan he’d brought them.

He knew it sounded smart.

He didn’t know what it meant in their world.

That’s when it hit his body first—a stomach-drop moment. Heat behind the eyes. The urge to bluff his way through. The instinct to move the meeting along before anyone noticed.

Then the internal self-talk landed, sharp and honest:

What the hell am I doing?

What the AI plan basically said

“Run SEO + paid search + LinkedIn content + webinars + nurture sequences…”

Great if you’re building a funnel. A distraction if your growth engine is repeat work and referrals.

It wasn’t evil. It wasn’t even wrong in the abstract. It just wasn’t their business.

The Trade-Off

This is the trade Ron made—without meaning to.

He wanted speed and certainty. He paid for it with ownership, clarity, and credibility.

Because when you let something else set your direction, you lose the ability to lead the people behind you.

Relief isn’t clarity. And teams can tell the difference.

The plan didn’t fail because AI is dumb. It failed because automation bias makes your brain treat confidence as correctness—especially when the inputs are thin.

That’s AI Strategy Theater: a plan that reads well and dies on contact with execution.

Ron’s Reset

Ron didn’t swear off AI after that meeting. He didn’t declare it useless. He didn’t pretend it went fine.

He did something quieter and harder: he rebuilt his thinking from scratch.

Ron’s reset looked like this:

• Truth: where work really comes from (referrals + repeat clients)

• Best people: the kinds of buyers who hire them fast and stay happy

• Offer: what they do in plain English, and why it matters

• Hard line: what they can’t break (delivery timelines and reputation)

• One move: one simple way to grow that fits a referral business (reactivate past referrers)

• One number: a weekly signal they can control (conversations booked)

Before he brought anything back into the room, he forced himself to answer the same five questions his team had asked.

Then he wrote a line at the top of the page, like a metronome:

One lever. One owner. One metric.

It wasn’t a marketing plan yet. It was Ron getting his brain back. It was him becoming the author again.

What Ron Actually Wrote (Ugly Draft)

Goal + time horizon: 25% more revenue next quarter.
Baseline: Most new work comes from referrals and repeat relationships — not strangers finding us online.
Best-fit buyers: GCs, engineers, and developers who are under schedule pressure and need clean answers fast.
Offer (plain English): We tell you what the ground will do before you build — so you don’t get expensive surprises later.
Constraint: Delivery timelines and reputation. More work isn’t worth it if we start missing calls and slipping deadlines.
One move: Re-activate the people who’ve sent us work before and make it easy to send the next one.
One weekly number: 5 qualified conversations with past referrers (calls or meetings) every week.

The Second Meeting

The following week, Ron brought the team back together.

He didn’t open with the polished playbook. He opened with the truth.

“Last week,” he said, “I brought you a plan that sounded smart. That was the problem.”

He walked them through the reset in plain language: what’s true, who they’re best for, what they sell, what they can’t break. Then he named the one lever they were pulling this month: reactivation and partner outreach to the people already most likely to send work.

He assigned ownership for next week. Specific names. Specific actions.

They also made one simple agreement: if they didn’t see movement in two weeks, they would change course—not argue, not blame, not add more channels. Just adjust.

By the end of week two, three past referrers had replied, and two calls were already on the calendar. It wasn’t victory. But it was signal.

And when the questions came, Ron didn’t flinch. He didn’t reach for the document like it was a life raft.

He answered like a founder who was actually present inside the strategy.

The room felt different.

Not because everyone was magically confident.

Because the fog cleared enough to move.

The team could feel trust return—not because the plan was perfect, but because Ron was leading again.

If You’re In This Fog

If you’ve done something like this, you’re not alone. It’s a common mistake under pressure.

Here are three footholds that don’t require a reinvention:

• Write your plan in plain English first—even messy. If you can’t say it simply, you don’t own it yet.

• Pick one move, not six. One lever you can actually run next week beats five ideas you’ll “get to.”

• Answer the five questions before you present anything. If you can’t answer them alone at your desk, you won’t answer them under fluorescent lights.

The five questions your team will ask:

• Who exactly are we targeting?

• What are we actually offering?

• Why are we doing this first (and not the other ideas)?

• Who owns this next week?

• What metric tells us it’s working—and when do we change course?

And if you use AI, keep it in its place:

• AI can generate options.

• But it can’t generate ownership.

• AI is a partner, not a pilot.

Ron ended up with a one-page brief he could actually lead from—and a handful of prompts that forced AI to clarify and stress-test instead of taking the wheel.

He didn’t need a louder plan. He needed a clearer one.

If you can’t explain it, it’s not your plan.

Download the Human-First Marketing Plan Brief + copy/paste prompts Ron used to rebuild the plan. No funnels. No fancy tech stack. Just a plan you can defend in a real conversation.