Five live builds on your own client work. One work week.

Day 1 · Discovery

Outcome: a scored AI readiness read on one real client or prospect, with evidence, posted
in the channel before you close the laptop.

Why this one

Every agency in your market is selling AI right now. Almost none of them can answer "is this
client actually ready?" with anything but a feeling. That gap is the whole opportunity. A read
you can defend is the difference between a proposal and a guess, and it is the cheapest thing
you will build all week.

Start here because the other four days need a subject. Today you pick the patient.

The teach

Readiness is not one thing. It is five, and they fail independently:

1. Strategy and ownership. Is there a named person who decides, and a reason to do this

that survives a bad quarter?

1. Technology stack. Can the systems they already run be read and written by something

other than a human clicking?

1. Operations and delivery. Is there a repeatable process to automate, or is every job

a snowflake?

1. Governance and risk. Who checks the output, and what happens when it is wrong?
2. Talent and adoption. Will the people who have to use it actually use it?

Score each 1 to 5. The score is not the deliverable. The evidence sentence under each score
is the deliverable, because that is the sentence you read aloud on the call.

The rule that makes it real: no evidence, no score. If you cannot write the sentence, write
"don't know." A "don't know" is a finding. It is often the most valuable line on the page,
because it is a question you get to ask them, and questions are how you get a second meeting.

The build (20 minutes)

1. Pick one real client or prospect. The one you know best, not the biggest.
2. Set a timer for 15 minutes. Fill the five rows. One sentence of evidence each.
3. Force a spread. Give at least one dimension a 1 or 2, and at least one a 4 or 5. If

everything is a 3 you have not scored anything, you have shrugged in a table.

1. Write one line at the bottom: the single dimension that is costing them the most money

right now.

1. Write the question you would ask them about your lowest score. That is your opener.

The template

CLIENT: ____________________          DATE: __________

Strategy and ownership      [ /5 ]   Evidence: ______________________
Technology stack            [ /5 ]   Evidence: ______________________
Operations and delivery     [ /5 ]   Evidence: ______________________
Governance and risk         [ /5 ]   Evidence: ______________________
Talent and adoption         [ /5 ]   Evidence: ______________________

Biggest money leak right now: ______________________________________
The question I would open with: ____________________________________

Your prompt

Paste this into any AI tool. Fill the brackets first. Then argue with what comes back, because
the argument is where the real evidence surfaces.

You are helping me score one company on AI readiness across five dimensions:
strategy and ownership, technology stack, operations and delivery,
governance and risk, talent and adoption.

Here is everything I know about them:
[dump everything: what they do, size, tools they use, how they deliver,
who decides, what they have already tried, what annoys them]

For each dimension:
1. Give a score 1 to 5.
2. Give ONE sentence of evidence from what I told you. Quote me.
3. If I did not give you enough to score it, say "don't know" and write the
   exact question I should ask to find out. Do not guess to fill the table.

Then name the single dimension costing them the most money now, and say why
in one sentence.

Do not soften anything. If my information is thin, say it is thin.

That last instruction matters more than the rest. Without it you get five confident paragraphs
built from nothing, and you will read them on a call and lose the room.

The one-command version

With Agency Script OS installed, the same job is one line, because the method above is already
a skill and the assessment shape is already canon:

/agency-script-os:as-sales-client-discovery-report acmeplumbing.com challenge-day-1

Company name and URL in, branded PDF out: a scored readiness read where every score quotes the
page element it came from, the bottleneck, the first three moves in order, and one next step.
It also writes down what it did not assess, which is the part that makes a report safe to hand
to a stranger.

Same method as your sheet. It scores from public evidence only, so it is a starting read, not
the version you would charge for. You just stop rebuilding the shape every time.

You do not need it today. Do it by hand today. You will feel exactly which part you never want
to do by hand again, and that feeling is worth more than being told.

Ship it

Post the five-liner in the channel. Client name optional or masked. What we want to see is the
evidence sentences and your one opening question.

The bar: someone else in the cohort should be able to read your five lines and tell you which
dimension you are about to sell into. If they cannot, your evidence is too vague.

Common stalls

• Picking the hardest client. You picked the enterprise one because it feels important.

Pick the one you know best. You are learning the instrument, not performing a concert.

• Everything is a 3. Force one low and one high. A flat sheet means you are avoiding a

judgment, and the judgment is the product.

• Scoring the technology and ignoring the people. Talent and adoption is the dimension that

kills the most projects and gets scored the least honestly.

• Writing evidence you cannot say out loud. If the sentence would be rude on a call, rewrite

it until it is true and sayable. That is a skill.

Tomorrow

Tomorrow we take your lowest-scoring dimension and turn it into a one-page blueprint a client
would pay to see. Bring today's sheet. You will need the evidence sentences.