A financial advisor gave AI one paragraph. Scoping showed what was missing.
He wanted AI to find $100,000 of value in how his office runs. One paragraph filled in the goal. Scoping showed the three things nobody had said yet.
Financial advisory practice · One paragraph of input · Scoped before any build
1
Paragraph of raw input
$100K
The value he set as the win
3
Gaps that had to close first
5
Steps in the build order
The Ask
"Most of what we do is just how we've always done it"
The owner of a financial advisory practice brought this to a scoping session. Here it is, as he wrote it:
"I want to be able to take a look at the processes and ways we currently manage our financial advisory practice. Most of what we do is just because that's how we've always done it. I'd like to be able to spot inefficiencies, and come up with better ways to manage our office. I think we're trying to raise quality, and that means proposing processes or ways of working that we haven't done before. This working in 90 days: Claude has a high-level view of our systems and processes. I can bring in a specific workflow, point to the SOPs or process docs we currently have and explore ways we could re-engineer them. That doesn't just mean tech, but AI or automation can definitely help. Single result: redesign at least one specific process that saves over $100,000 in value."
It's a good ask. It's also not a plan yet.
What Scoping Did
One paragraph filled the goal. The rest had holes.
AI sorted the paragraph into a scope doc, section by section. Each section was marked filled, partly filled or still open.
Filled
The outcome
Find better ways to run the office than "how we've always done it." Raise quality. Win: redesign one process worth $100,000 or more.
Filled
The 90-day picture
AI holds a high-level view of the practice. He brings one workflow and its process docs, and they redesign it together.
Partly
The inputs
Some process docs exist. Nobody knows which processes are written down and which live in people's heads.
First job: gather the docs into one place, and list which processes have none.
Partly
The output
A redesigned process, plus a living map of how the practice works. The format, and who reads it, weren't stated.
Open
How the work runs today
The actual processes, who runs them and where the time goes. None of it was in the paragraph.
Question to close it: "Walk me through one process start to finish, the way it really happens today."
Open
His standard for "better"
Without it, AI proposes redesigns he wouldn't sign off on.
Question to close it: "When you see a process that works well, what's true about it?"
Open
Compliance and trust
Financial advice is regulated. Record-keeping rules, client confidentiality and how accurate a redesign must be had to be settled before any build.
The Insight
The bottleneck was the paperwork, not the AI
The hard part was getting the process docs in one place and knowing which ones existed at all.
That's what scoping is for. It shows you the real first step before you spend a dollar on the exciting one.
The Plan
Smallest useful version first
20-minute inventory. List every process bucket in the practice. No detail yet.
Pick one. Choose the process most likely to clear the $100,000 bar.
Document today. Write down how it really works now.
Redesign it. Propose the new version, with AI and without.
Ship and measure. Put it in place, measure the saving, then move to the next process.
Without scoping
Start building on a paragraph
Find the compliance wall halfway in
AI guesses at what "better" means
No way to tell if you hit $100,000
With scoping
Goal and win written down
Gaps named, with the question that closes each
Real bottleneck found on day one
A five-step plan, smallest step first
Scoping is the conversation that closes the gaps before anyone builds.
Do This Yourself
Get the scoping skill
This is the skill I use to scope ideas like this one. Install it in Claude, then tell it what you want to build. It maps the real steps, the honest cost and the first move, and writes a one-page brief you can decide on and hand off.