Lesson 5 · 16 min
High technical skill with weak communication delivers less than average technical skill with good communication. The client speaks in adjectives ("a lot," "slow," "confusing"); the people who decide read numbers. This lesson shows you how to turn the vague into a number, talk about budget without awkwardness, and deliver everything on one page.
Start the lesson
Goal: turn a vague request into a requirement with a number, talk about budget without awkwardness, and write the one-page report.
Each step turns what the client said into something the board approves.
The success question helps at every step: "Six months from now the project worked; what exactly changed in the business?"
What the client says and what you give back.
Ask for the quantity and the peak.
How many per day? And what's the peak of the week?
Ask for the unit of time.
Slow by how many hours: per day, per week, per month?
An error with no cost doesn't go in the proposal.
How many errors per hundred handoffs, and what does each one cost?
Sum it up in one sentence and ask if that's right.
When the summary is wrong, the client corrects it with details they would never give away for free.
So the real problem is the approval queue, and the form just exposes that queue. Is that right?
Four moves and the right close.
Say who you are and tie your goal to theirs. It's not a zero-sum game.
My goal is for this project to lower your cost per ticket. If it does, you'll call me again.
Show up with 2 or 3 hypotheses and ask for corrections. It's easier to edit a finished text than to face a blank page.
From my research, these are the three main points. Correct me if I'm wrong.
"What's your budget?" sounds like "can you afford it?".
Don't apologize before talking about price, and don't push the topic off to an email later.
Is there any budget already set aside for AI initiatives? It helps me work backward from your expectations.
Never end with "I'll send you a proposal." Book the next meeting before this one is over.
A slow proposal is read as lack of priority.
The report is the piece your contact takes to the board without you in the room.
<COMPANY> · <area> · <date> To: <title of the decision-maker> PROBLEM (2 lines) <the pain in the client's own words and why it costs money> NUMBER Today: <metric> = <value>, measured by <how>. Estimated cost: <value per month> (assumption: <which>). RECOMMENDED OPTION <what to do, in 3 lines, with deadline and owner> Cheaper alternative: <which one and what is lost> RISK <what could go wrong and how we'll know early> Not included: <...> DECISION REQUESTED <yes/no for the pilot> by <date>.
BUDGET - Is there any budget already set aside for AI initiatives at the company? - It helps me work backward from your expectations. CONSTRAINTS (they weigh more on price than scope) - Which systems does this process go through, from start to finish? - Who grants access to them, and how long does that usually take? - Is there any data here that can't leave the company? SUCCESS - Picture six months from now. The project was a total success. What exactly changed in the business?
Research this company and give me back a meeting briefing. I don't want flashy ideas: I want the most practical ones to implement. Company: <public name and sector> Area I'll be talking to: <area or title> Approximate size: <number of employees> Return in this order: 1. Business model and where the revenue comes from, in 5 lines. 2. How the area probably operates today: steps and tools. 3. The 5 most practical scenarios, ordered by ease of implementation. 4. For each scenario, the business justification in one sentence. 5. The 3 questions I'll ask to confirm or knock down each hypothesis. Explicitly mark what is public fact and what is your own assumption.
Use the case from your proposal in lesson 3.
List 3 vague phrases your client (or the clinic from lesson 2) said, and write the question that turns each one into a number.
Done when: the 3 phrases each have a question next to them, and each question asks for quantity, time or cost.
Fill in the executive report template. Read it out loud: you should be able to decide in two minutes.
Done when: the report fits on one page and has the problem, a number with how it's measured, a recommended option, the risk and the decision requested with a date.
4 questions. The explanation appears after your attempt.
Question 1 of 4
The meeting ended well and the client showed interest. What's the best close?
Question 2 of 4
Which question about money triggers the least defensiveness?
Question 3 of 4
The client says "the process is slow." What do you say back?
Question 4 of 4
What must not be missing from the one-page report?
Pick an answer. The explanation shows up after you try.