Hi {{first_name | Reader}},

Three weeks ago I had a client call at 10am.

I was in the middle of something else. The meeting notification popped up. I had twelve minutes.

Then an email landed in my inbox.

The sender was my own system. The subject line: *Pre-Meeting Brief: [client name], Monday 10:00.

I hadn't asked for it. I hadn't scheduled it. I hadn't set a reminder.

I opened it.

Company snapshot. One paragraph. Who they are, what they're building, the industry context. Pulled from the entity file I'd built on them over six months.

Financial position. Where cash stands. What revenue looks like against their last forecast. Two numbers that mattered most, with a one-line interpretation.

Recent signals. The last three emails between us. The open action items from our previous meeting. One risk flagged automatically: a deadline I hadn't looked at in two weeks.

Prep points. Three questions worth asking. One area of probable friction based on the email thread.

Twelve minutes. That's how long I had. I read the brief in four.

I walked into that call knowing exactly where cash stood, what was at risk, what had been said in the last three emails, and what I needed to push on. Every email had been read. Every number had been checked. Every open item had been flagged.

The system didn't replace the preparation. It guaranteed it.

That email wasn't a summary. It was a brief.

A brief is a specific format. One page. Built by a team for someone who doesn't have time to assemble the picture themselves.

The most famous one in the world lands on the US President's desk every morning. An entire intelligence community filters, verifies, and curates information so that one person can read a single page and act.

The format works because the person reading it needs the picture already assembled, already verified, already current.

That's the CFO's problem too. You're walking into a client call. You need to know what changed since the last one. You don't have time to re-read six email threads, pull up the latest trial balance, and check your own notes from three weeks ago.

As of today, the system builds that brief for you. Every time, before every call, without being asked.
Want to know how ready your finance function is for this? The Scalable Finance Wheel maps it across six dimensions. Good baseline before adding an AI layer on top.

Last week I showed you the foundation: VS Code, Claude Code, CLAUDE.md, the folder structure.

That setup is the skeleton. This brief is what the skeleton produces when it has something to work with.

Here's what the system read before writing it.

The entity file. Six months of notes on this client, living in a single structured folder. Their industry, their business model, their goals, the history of our work together. I built this over six months. Every meeting, I add a note. Every deliverable, it goes in the folder. The AI reads the whole thing every time.

The financial data. Live numbers from their accounting software. Trial balance, cash position, AR aging. Pulled live, formatted into context the AI can reason about.

The signal layer. Recent emails. The last meeting's action items. Outstanding tasks. The AI read these not to summarize them, but to look for patterns. What's overdue, what's escalating, what I might have missed.

Having three data sources doesn't mean the brief is reliable. The entity file can exist and still reflect a conversation from six months ago. The financials can be there but pulled from a trial balance that's two months old.

So I wrote a scoring formula.

Seven dimensions, each weighted.

Each one scored 0 to 100 based on what's present and how fresh it is. When was the last financial pull? How recent is the last email? Is the entity file reflecting what you know today, or what you knew three months ago?

Weighted average.

Most people don't have a quality standard for their meeting prep. They trust their memory, open a few tabs, skim the last email, and hope they're caught up.

The brief that arrived at 9:47 scored green across all seven dimensions. That's why I trusted it. That's also why the system sent it at all.

Last week was the foundation.

The next three issues are the layers.

Issue 7: The context layer. Teaching the AI who your clients are. Their goals, their history, how your relationship with each one shapes your advice.

Issue 8: The signal layer. Live inputs: emails, meetings, action items. How the AI reads activity and surfaces what you would have missed.

Issue 9: The synthesis. One document that replaces ten reports. Generated before your call, without being asked.

By the end of this arc, you'll have a system that writes its own pre-meeting briefs. Built once. Runs on its own.

Two things I'm saving for later that are worth flagging now.

There's a post-meeting brief too. Generated after every call. What was discussed, what you committed to, what the client said that's worth tracking. Same system, different direction. Backward-looking instead of forward. It also tells you something about how you ran the meeting. More on that in a future issue.

And getting a new client onto this system is its own process. Onboarding starts with everything that already exists. Every past report, every meeting log, every financial document. Analyzed, structured, loaded. The AI doesn't start from nothing. It starts from everything. That's also coming.

The Scalable Finance Wheel covers the six dimensions that determine whether your numbers are working for you or just sitting there. Useful baseline before adding an AI layer on top.

One thing I'm curious about: what would you want in that brief? If a document arrived in your inbox twelve minutes before a client call, what would make it worth reading?

Hit reply. I read these.

-- Samer

P.S. We're about six weeks out from a live build session. I'll be putting together the whole stack from scratch in 45 minutes. More on that soon.