Connected reading

This guide is the practical companion to Part 3 of Erin’s case.

• Read the story: She Had a Recovery Plan. Then Sales Slowed Down.

If you missed the earlier parts, start here:

• 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

• Part 2 story: Knowing the Right Number Didn’t Fix the Business

• Part 2 guide: How to Build a Recovery Plan for a Small Business Using AI

The plan was better. The business still felt slow.

There is a frustrating stage that comes after the first good plan.

You made the plan.
You stopped the worst of the damage.
The team has more room than it did.
The old mistake is no longer getting bigger every week.

And yet something still feels off.

The direction seems sound. The work is happening. The effort is real. But the business is not moving the way you expected it to — and you cannot quite put your finger on why.

That is the moment most owners handle badly.

Some panic and throw out the strategy too soon.

Others hold on too long because admitting drift feels like admitting failure.

Neither helps.

Because the problem usually is not that the strategy is wrong.

The problem is that the plan has started drifting away from reality — quietly, gradually, in ways that do not look like failure until they have had enough time to get expensive.

That is where this guide begins.

What drift usually looks like

Drift is rarely dramatic at first. Usually it looks like almost.

You are doing the work. The activity is there. But the result is not showing up at the pace you expected.

Sometimes that drift shows up in sales:

• conversations are happening, but not enough turn into signed work

• proposals go out, but decisions stall

• the pipeline looks active, but cash is still late

Sometimes it shows up in delivery:

• the work is selling, but taking more time than expected

• revisions keep spreading

• quality starts slipping because capacity is tighter than it looked

Sometimes it shows up in margin:

• revenue holds steady, but profit gets thinner

• one offer looks fine in the report while quietly costing more to deliver than anyone expected

Sometimes it shows up in the team:

• the numbers look stable, but people are getting worn down

• more and more of the work needs senior attention just to keep it steady

Sometimes it shows up in timing:

• the strategy is right, but the buying cycle is slower than you assumed

• cash arrives later than the plan expected

• the market is not moving at the speed your plan was built around

The common thread is simple: the plan has not fully failed. It has started to drift. And drift is easy to miss because it often looks like progress that is not turning into results quickly enough.

The two kinds of numbers you need now

Once the plan is running, one of the most useful distinctions you can make is between lead indicators and lagging indicators.

Lead indicators

Lead indicators tell you whether the right activity is happening — the inputs you can influence directly.

Examples:

• serious sales conversations completed

• proposals sent

• projects started

• follow-ups made

• deliveries finished on time

• client check-ins completed

These numbers tell you whether the machine is moving.

Lagging indicators

Lagging indicators tell you whether the activity is producing the result you wanted — the outputs that follow.

Examples:

• signed contracts

• cash collected

• profit left after delivery

• projects completed profitably

• client retention

• team strain easing instead of worsening

These numbers tell you whether the machine is moving in the right direction.

You need both.

Because this is exactly how drift hides: lead indicators can look healthy while lagging indicators quietly disappoint. Activity can be strong while results fall short.

When the lead indicators look fine but the lagging indicators are falling short of what the plan expected, that is the moment to bring AI back into the conversation.

What AI is for at this stage

The first way most people use AI is obvious.

They ask it for ideas, ask it for a plan, ask it to help organize the mess.

That can be useful.

But it is not the full opportunity.

The bigger use is bringing AI back into the work while it is still in motion — not just to make the plan, but to review it against what is actually happening.

Used once, AI can help you make a plan.

Used regularly, AI can help you keep the plan honest.

That is a different kind of relationship.

You are not asking for a brand-new strategy every time something feels off. You are asking AI to help you compare the plan to reality, see where the gap is opening, notice patterns you are too close to see, and think through one or two smart adjustments before you overreact.

A plan reviewed regularly stays closer to reality. A plan reviewed once a quarter has already had three months to drift.

What to bring AI when the plan starts slipping

The quality of AI’s help at this stage depends almost entirely on what you bring it. Generic inputs produce generic outputs. Real inputs produce useful thinking.

Bring four things.

1. The original plan

What were you trying to do? The offer you planned to lead with, what you planned to stop or reduce, what you expected to happen in 30 or 90 days. Do not rely on memory — pull up the actual plan.

2. What actually happened

Not what you meant to do. Not what it felt like happened. What actually happened — calls completed, proposals sent, work signed, cash collected, profit, delays, reschedules, missed targets.

3. Where the work seems to be getting stuck

This is where lead and lagging indicators become useful together. Where is activity healthy but results weak? What stage is slower than expected? What part of the plan feels heavier, slower, or messier than it was supposed to?

4. What people are actually saying

This is where hidden truth often lives — emails, notes, objections, feedback, follow-up threads, client comments, team observations, project debriefs.

If you have recorded and transcribed conversations, use them only with permission. That material can be especially valuable because AI can help spot repeated hesitation patterns across more conversations than you can hold in your head at once. Not because AI gets to decide what they mean — because it can help you see what keeps showing up.

The more specific and honest the material you bring, the more useful the review becomes.

A simple review process

Five steps, in order.

Step 1: Compare the plan to what actually happened.

Start with the gap.

Ask:

What did you expect by now, and what actually happened instead?

You are not building a case against yourself. You are looking for where reality started to separate from the plan.

Step 2: Find where lead indicators are healthy but lagging indicators are weak.

That is where drift is most likely hiding.

Examples:

• lots of conversations, too few signed contracts

• plenty of work sold, too little profit left after delivery

• revenue holding, cash timing worsening

• projects increasing, quality starting to slip

The question is not whether the whole plan is bad — it is where this plan is no longer turning into the result you expected.

Step 3: Read the human evidence.

Now go back to the notes, emails, transcripts, feedback, and debriefs. Read them looking for repeated patterns — not just what people said once, but what keeps showing up.

This is where AI can help you notice:

• hesitation

• recurring objections

• decision friction

• scope mismatch

• timing issues

• delivery pain

• pricing resistance

• team strain that has not shown up cleanly in the numbers yet

The reports show you what happened. This material often explains why.

Step 4: Ask what still holds and what needs to change.

Do not throw out the whole strategy too quickly.

Ask which parts still look directionally right, which are proving slower or heavier than expected, and what needs to be repackaged, narrowed, priced differently, or reviewed more often.

This is the difference between adjustment and panic. Panic starts over. Judgment changes what reality has shown you.

Step 5: Choose one adjustment for the next 30 days.

The goal is not to reinvent the business every month — it is to choose one or two useful changes that make the current strategy more workable in real life.

One honest adjustment, executed well, beats five technically correct changes nobody has time to implement.

Plan Review Prompt

Here is a prompt worth using when the plan is running but something feels off.

I need help reviewing my plan against reality.

I do not think the overall direction is necessarily wrong, but part of it seems to be getting stuck.

Here was the plan: [PASTE]

Here is what actually happened: [PASTE]

Here are my notes, emails, feedback, or observations from the people involved: [PASTE]

Help me do five things:

1. Tell me where the plan seems to be drifting.

2. Help me separate what actually matters from the normal friction of running a business.

3. Identify the real bottleneck right now.

4. Suggest 2–3 adjustments that fit the business I am actually trying to build.

5. Help me choose one change to make in the next 30 days without throwing out the whole strategy.

Keep this practical and in plain English.

That prompt is useful because it does not ask AI for a grand new plan. It asks AI to review the current one honestly. That is the right job at this stage.

One useful follow-up prompt

If you have richer notes, transcripts, or feedback, this is where AI can become even more helpful.

I want help reading between the lines in this material.

Here are notes, emails, transcripts, feedback, or observations from customers, prospects, team members, or delivery debriefs: [PASTE]

Help me do four things:

1. Identify repeated hesitation patterns or friction points.

2. Tell me what people may be signaling without saying directly.

3. Tell me whether the issue looks more like timing, offer fit, decision friction, delivery strain, pricing, scope, or something else.

4. Suggest how I might adjust the next step based on what they are actually telling me.

Keep this plain English and practical.

Most people have never used AI this way — not to generate output, not to replace judgment, but to help read the truth that is already sitting in the work.

That may be one of the most underused applications in the whole toolkit.

A simple weekly review

If you are actively trying to change the direction of a business, your weekly review does not need to be big. It just needs to be real.

Five things, once a week:

1. What serious work actually happened? Conversations, deliveries, client interactions, proposals.

2. What turned into a real result? Signed work, cash collected, projects completed, margin improved.

3. Where does work seem to be getting stuck? Which stage is slower, heavier, or messier than expected?

4. Does the work being sold and delivered still match the business you are trying to build? Or are you drifting back into what felt familiar?

5. What truth from notes, emails, or debriefs is worth paying attention to before it gets buried? What keeps showing up?

That is enough to catch drift early. The goal is not a perfect dashboard. It is staying close enough to reality that the gap between plan and actual does not widen for three months before anyone notices.

Put this to work

Pull up the actual plan — not the version in your head, the real one. Pull up what actually happened — do not summarize from memory. Pull up the human evidence — notes, emails, follow-ups, feedback, debriefs, transcripts if you have permission to use them.

Then use the Plan Review Prompt above. If you have enough human evidence, use the Read Between the Lines Prompt after it.

By the end, you want five things on the page:

• where the plan is getting stuck

• what still looks right

• what needs to change

• one adjustment for the next 30 days

• one short weekly review list

That is enough to keep the business moving.

Closing

A plan is not something you make once and admire. It is something you keep honest.

That is what this part of the work is really about — not proving whether the plan was perfect, but seeing where it is drifting while you still have time to do something about it.

Used once, AI can help you make a plan.

Used regularly — brought into the work while it is still in progress, given the real picture instead of the polished one — it becomes something more useful than a planning tool.

It becomes a second set of eyes on the work while it is still close enough to change.

That is what compounds.

Not the plan itself.

The habit of keeping it honest.

Download the Plan Reality Check. It helps you compare the plan against what actually happened, find where the gap is opening, and choose one adjustment for the next 30 days.

A few questions worth answering

What if the plan is directionally right but still moving too slowly?
A plan can be strategically correct and still move too slowly to relieve pressure. That does not mean you abandon it. It means you review where the friction is and adjust the packaging, timing, entry point, or delivery until the strategy becomes more workable in real life.

What if lead indicators look healthy but lagging indicators are still weak?
That gap is where the most useful diagnostic work happens. Something is converting activity into results less efficiently than expected. AI can help you examine whether the bottleneck is in the offer, the sales process, delivery, timing, pricing, scope, or buyer fit.

What if I do not have call recordings or detailed notes?
Start with whatever you have — emails, follow-up threads, short notes, team observations. The habit of capturing truth matters more than the format. As you build the practice, the quality of the inputs improves naturally. If you do record or transcribe calls, get permission first.

What if AI gives me too many possible adjustments?
Choose one. The goal at this stage is not to redesign the whole strategy. It is to make the next useful move without abandoning the direction. One honest adjustment, executed well, is worth more than five theoretically correct ones that never get implemented

Continue Erin’s case

• Read the Part 3 story: She Had a Recovery Plan. Then Sales Slowed Down.

• Go back to Part 2 story: Knowing the Right Number Didn’t Fix the Business

• Review the Part 2 guide: How to Build a Recovery Plan for a Small Business Using AI

• 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