Connected reading

This is Part 2 of Erin’s case.

If you’re starting here, begin with:

• Part 1 story: The Dashboard Said It Was Her Best Offer. It Was Quietly Draining the Business.

• Part 1 guide: Why Your Best-Looking Number May Be Lying to You

For the practical companion to this article, read:

• How to Build a Recovery Plan for a Small Business Using AI

Then continue to the finale:

• She Had a Recovery Plan. Then Sales Slowed Down.

• How to Use AI to Review a Plan Before Drift Gets Expensive

Friday was for fixing the mess

Erin was dripping wet when she pushed through the coffee shop door Friday morning.

A heavy summer thunderstorm had rolled in just as she got out of the car, and the two-block dash had been enough to soak the back of her blouse, frizz her hair, and soften one corner of the notebook tucked under her arm.

She looked automatically toward her usual table.

Taken.

A loud group of retirees had claimed it and were already halfway into an argument about whose alma mater was going to win the SEC next season.

“Perfect,” she muttered.

The past week had been hell. Not dramatic — just the ordinary, grinding kind that happens when you discover something important and still have no time to deal with it. Four straight days of long hours for clients who were taking more out of the business than they were leaving behind.

She had cleared Friday on purpose. Moved meetings. Delayed calls. Told the team she’d be unavailable unless something was actually on fire.

Today was for fixing the mess.

She ordered a drip coffee with unlimited refills. No oat-milk latte. This was a legal pad, calculator, and hard-truths kind of morning.

The only open table was small, sticky, and dusted with crumbs. Far enough from the retirees to count as a win.

She sat down, opened the laptop, and found the page she had written Monday morning.

At the top, underlined twice:

The problem is not revenue.
The problem is how much money the business keeps.

That had helped.

It had not fixed anything.

She still had nine new retainer clients in the business. A team already stretched thin. Work due Monday for clients who were paying too slowly and asking for too much. A business that had quietly reorganized itself around the wrong engine.

Knowing what was wrong was useful.

It was not the same as knowing what to do next.

That was what this Friday was for.

She opened ChatGPT.

Not with a neat planning prompt. Not with the voice she used when she wanted to sound strategic even to herself.

With everything she had.

She pasted in the January-through-May financials. Revenue by offer. Rough economics by offer. Notes about where clients actually came from. Which work sold fastest. Which work paid fastest. Which work took the most founder attention. Which work the team was actually best at. Which work had been quietly draining the life out of everyone.

Then she typed:

I need help building a practical recovery plan for my business.

Please do not give me generic business advice. I need clear, actionable thinking based on the information I provide.

I want you to help me do six things:

1. identify what is hurting profit or cash flow most right now

2. identify which products, services, or client types are still healthy and worth keeping

3. help me figure out which offers make the most profit after the real cost of delivery

4. suggest what I should stop selling first, what I should keep, and what I should lead with instead

5. suggest what changes could improve the economics of the weak offers — price, scope, revisions, payment terms, onboarding

6. build a practical 30-day and 90-day transition plan I can follow while the business is still operating

Here are the facts:

• recent financials

• revenue by offer

• rough economics by offer

• current offer mix

• cash timing issues

• capacity constraints

Here are the realities the numbers do not show:

• the kind of business I actually want

• what work we are best at

• what has been draining us

• what cannot break

• what I am afraid to change

Here is how sales actually work in this business:

• where clients come from

• which channels bring which offers

• how long each offer takes to close

• which offers bring cash fastest

• how much founder effort each offer takes to sell

Please keep your answer in plain English, organized in this order:

A. What is hurting the business most right now
B. What is still working
C. Which offers appear most profitable after the real cost of delivery
D. What I should stop, keep, or change first
E. A 30-day plan
F. A 90-day plan
G. One weekly metric I should track
H. What assumptions or blind spots I may still be missing

I do not need accounting perfection. I need a plan that is realistic and good enough to test in the real world.

She hit enter and sat back.

Even writing it that clearly had changed something.

The answer came back slower than the clean growth advice she had gotten in January.

That felt right.

This time, she hadn’t asked for certainty.

She had brought a mess.

It started here:

Before you optimize, stabilize.

The business does not have a scale problem first.

It has a bleed problem first.

Erin stared at that line.

Not because it sounded smart. Because it named exactly what the week had felt like.

She didn’t need a vision board for the future. She needed to stop making the current problem worse.

The answer continued:

The offer actively hurting the business most right now is new retainer work sold at current price and current scope. It is not just underperforming — it is amplifying strain across the rest of the business.

Your first priority is not finding the ideal future mix. Your first priority is stopping additional damage while protecting the parts of the business that are already healthy.

Under that, it sorted the business into three categories.

Stop. Keep. Rebuild.

Stop what is hurting the business

Stop

• defaulting new leads into retainers

• using strategy as a sales mechanism

• treating all retainers as if they have the same economics

Keep what is still working

Keep

• legacy retainers

• strategy work

• website refreshes

Rebuild around a healthier business

Fix

• current new retainers

• onboarding

• payment terms

• scope

That was the first hour. No vision. No growth language. Just triage in plain English.

And once she could see the wound clearly, the next question got easier.

“Fine,” she typed. “If I stop what is hurting the business, what should this business actually be built around?”

The answer that came back did not give her one winning offer.

It gave her tradeoffs.

Based on what you’ve shared, the healthier version of this business likely does not lead with new retainer work.

It likely looks more like this:

• strategy intensives lead the relationship

• website refreshes follow as bounded implementation

• retainers become selective follow-on work, not the default front door

• legacy retainers stay protected

• high-drag new retainers are reviewed, repriced, narrowed, or reduced over time

Then it put a shape to it.

A healthier target mix

That was not the plan to start on Monday.

It was the picture of what a healthier business would actually be built around.

And it sounded like the business she had wanted before recurring revenue and clean-looking charts had flattened the whole thing. The work she and her team were best at had never been endless monthly servicing. It had been diagnosis. Direction. Clarity. Turning noise into a plan.

“That sounds right,” she typed. “But what can I actually change? I still have the business I have.”

Then it helped her work through the changes that could improve the economics.

Restore strategy as a real paid offer. Stop recommending retainers first. Reprice future retainers. Narrow scope. Reduce revision allowances. Tighten payment terms. Bring website refreshes back as the natural follow-on to strategy. Stop selling the retainer profiles that always turn into drag.

She wrote a new header across the top of her notebook:

Stop / Keep / Rebuild

That was better than “growth plan.”

More honest.

The next few hours were the real work — not asking AI for the perfect future state, but working out what had to happen first, second, and third to get from the business she had to the business she actually wanted.

She and GPT worked through it in layers.

What had to happen now. What had to happen in 30 days. What had to happen over 90.

By lunchtime, the outline had sharpened into something she could actually run.

Immediate

• stop pitching retainers as the default

• restore strategy as a paid standalone offer

• review the nine new retainers and rank them by drag, payment reliability, and fit

• protect legacy retainers and delivery quality

30 days

• lead all new conversations with strategy

• tighten how future retainers are quoted

• review collections discipline

• set clear scope and revision boundaries

90 days

• rebuild the pipeline around strategy intensives and website refreshes

• use retainers as selective follow-on work

• decide which retainer profiles are worth keeping long-term

• build the business around work that is profitable, sellable, and actually worth being known for

The weekly metric mattered.

She had learned enough this week to know that a bad metric could ruin six months.

So she asked which single number mattered most while she made the shift.

A few options came back, but one stayed with her:

new work sold by healthy offer type

Not total revenue.

Not number of calls.

New work that actually matched the business she was trying to build.

Because it would keep her from celebrating every sale that made the underlying problem worse.

Then she asked:

“What tells me this plan needs adjusting?”

The answer came back with just enough discipline to be useful:

If strategy work does not begin converting into direct revenue or bounded follow-on work within 30 days, revisit:

• pricing

• offer framing

• sales channel quality

• whether website refreshes should be pushed harder as the faster bridge offer

If you continue losing money while selling the healthier mix, the problem is no longer the mix alone — it is pricing, delivery cost, or overhead.

That gave the plan a spine.

Not certainty.

A structure.

Something that could hold up when Monday started asking things from her again.

“Show me what the transition looks like month by month,” she typed.

This time the answer came back in numbers.

Erin’s Transition Map

And beneath that, one more adjustment she should have made months earlier:

New payment structure

Strategy intensives

• 50% on booking

• 50% on completion

Website refreshes

• 25% on booking

• 25% at milestone

• 50% on completion

She wasn’t just changing what she sold.

She was changing when the money arrived.

That mattered almost as much as the mix itself.

Around 2:00, she stopped typing and just read.

The plan wasn’t elegant. It wasn’t the kind of thing you turned into a keynote slide.

It was better than that. It respected reality. It didn’t pretend she could rebuild the business in one clean motion.

It started where actual recovery starts:

• stop making it worse

• keep what is still healthy

• restore the work that makes good money

• move the mix over time

• measure the right thing

• adjust when reality pushes back

That was the part she had been missing when she was thinking like an analyst instead of an owner.

In a spreadsheet, you solve for the ideal mix.

In a real business, you stop the bleeding first.

Then you earn your way back to the ideal.

She smiled at that, barely.

It felt like the first true thing she had said to herself all week.

Erin stepped out of the coffee shop more than ten hours after she had walked in that morning.

The storm had moved on hours ago. In its place, the heavy wet heat of a summer afternoon. Her shirt was dry. The buzz of her coffee had stopped working sometime around lunch. Her eyes hurt. Her notebook was full.

But she had what she came for.

Not relief.

Not proof.

A plan.

She checked the time — 4:37 — slid the phone into her bag, and started toward the car.

Monday was coming.

This time, she was ready for it.

Part 1 taught her that the dashboard had been answering the wrong question.

Part 2 taught her something just as important: knowing the right number doesn’t fix the business. It gives you the first honest chance to rebuild it.

The difference between Erin’s January prompt and her Friday one wasn’t the tool. It was what she brought to it. In January she handed AI the clean version — revenue data, client counts, a tidy question. In June she handed it everything: the facts, the realities the numbers did not show, how sales actually worked, the drag, the fear, the work the team was actually good at, the business she actually wanted.

AI can help you build a recovery plan. But it can only work with what you give it.

The financial reports tell AI what happened.

The extra context tells AI what matters.

The tradeoffs are still yours.

Continue Erin’s case

• Go back to Part 1 story: The Dashboard Said It Was Her Best Offer. It Was Quietly Draining the Business.

• Review the Part 1 guide: Why Your Best-Looking Number May Be Lying to You

• Read the companion guide to this article: How to Build a Recovery Plan for a Small Business Using AI

• Continue to Part 3: She Had a Recovery Plan. Then Sales Slowed Down.

• Read the final guide: How to Use AI to Review a Plan Before Drift Gets Expensive