Hi {{first_name | Reader}},
I was looking at a client's January P&L last week. Revenue was down 27% from the prior period.
My first instinct was to flag it. Bad news. Prepare talking points. Soften the landing for the board.
Then the system I've been building did something I didn't expect. It didn't just show me the number. It showed me what the number was supposed to be.
A 60% annual growth target. Set by the CEO. In a meeting on January 15th. Based on three assumptions: a new product line launching, two new markets opening by Q3, and a seasonal ramp curve that historically peaks in summer.
The 27% drop wasn't just bad. It was bad against a specific plan, built on specific assumptions, stated by a specific person, in a specific meeting. And two of those three assumptions were already flagged as at-risk.
That's a completely different conversation than "revenue is down."
Last time I showed you what happens after you hang up a call. How the system coaches you, updates your brief, creates tasks, all without you touching anything. That was one meeting, one direction: looking backward at what just happened.
This week I want to show you what happens when you point those same meetings forward. When what was said in January becomes the lens for judging what shows up in February's P&L.
Financial numbers only become intelligence when they're evaluated against strategy signals extracted from real meetings, tagged to chart of accounts nodes, and compared against KPI targets set by specific people on specific dates. This architecture connects your ledger to your strategy automatically, and every recommendation comes with a quantified P&L or cashflow impact.
What most finance teams actually do
You see the 27%. You open the GL. You open last month's board deck. You try to remember what was discussed. You check your notes, if you took any. You piece together a narrative from memory and scattered documents.
The strategy lives in your head. The assumptions live in a meeting that nobody transcribed. The targets live in a budget spreadsheet that hasn't been touched since it was approved.
So when you present the number, you're presenting it naked. Without the context that gives it meaning. The CEO hears "down 27%" and reacts to the number. Not to the gap between reality and the plan they set.
How one number flows through the system
Here's what I built last week for a client. The full architecture, end to end.
Layer 1: The chart of accounts.
Every financial line has a node. Gross Revenue, Net Revenue by business unit, Modified Gross Profit, Modified EBITDA, Cash & Banks, Current Ratio. 23 KPIs total across six categories. Each one has a definition, a target, and a source.
Layer 2: The strategy document that writes itself.
CFO meetings happen. Fireflies captures the transcript. But instead of a summary that sits in a folder, a tool I wrote extracts six types of strategic signals: goals, decisions, assumptions, constraints, commitments, and context shifts.
Each signal gets tagged to the specific chart of accounts node it relates to.
"We're targeting 60% revenue growth" tags to Gross Revenue. "Three new locations by Q3" tags to Revenue Growth Rate. "Onboarding delays pushing launch to Q4" tags to both, as a context shift.
The strategy document updates after every meeting. Nobody writes it. Nobody maintains it. It builds itself from conversations.
Layer 3: Every number meets its judge.
When January actuals arrive, the system pulls up the strategy context for each node. Revenue isn't just down 27%. Revenue is down 27% against a 60% growth target, set by the CEO on January 15th. Based on three location ramps assumed the same day. With one of those ramps flagged as delayed in the February meeting.
The number isn't evaluated in isolation. It's evaluated against everything anyone ever said about it, in any meeting, automatically indexed and tagged.
And then the system tells you what to do about it, and what it's worth. Not "revenue is red." Instead: the 60% target assumed three location ramps. One is delayed 90 days. Rebase the annual forecast to reflect two ramps, not three. That closes a $400K gap in the annual projection and protects Modified EBITDA by 3 points. Revisit the hotel vertical assumption in the next meeting.
Every recommendation comes with a number. Not "fix this." Fix this, and here's the cashflow or profitability improvement if you do. That's the difference between a report and a CFO.
The loop
I built this on a Sunday. Monday I walked into the CFO meeting with it.
The difference was immediate. Instead of presenting numbers and hoping I remembered the context, every line item came with its own history. The targets, the assumptions behind them, the moments those assumptions started cracking.
The CEO didn't hear "revenue is down." He heard "the plan assumed three locations by Q3. One is delayed. Here's the impact on the annual target, and here are the two assumptions we need to revisit."
That's the architecture. Chart of accounts to KPI to strategy signal to evaluation to actionable insight. A closed loop where meetings inform strategy, strategy informs analysis, analysis produces specific recommendations, and those recommendations inform the next meeting.
I'll share what happens when this runs against three months of real financials. The patterns it catches that I missed.
---
A few of you have asked me how to learn this. How to build this kind of system for your own practice or your own company.
If you want me to teach it, reply with TEACH ME and tell me what kind of clients you serve. I'm figuring out the best format for this, and your reply will shape what I build.
---
I'm doing a live session soon. The full system, on screen, with anonymized client data. Architecture overview, then workflows walked through end to end.
Know who the knocker-upper was? Before alarm clocks existed, there was a person whose job was to walk through town with a long stick, tapping on windows to wake people up for work. That job disappeared overnight when alarm clocks became cheap.
Spreadsheets are the long stick. Claude Code is the alarm clock.
Don't be a knocker-upper.
Frequently Asked Questions
What is a closed-loop financial intelligence system?
A closed-loop system connects your chart of accounts to KPI definitions, mines CFO meeting transcripts for strategic signals (goals, decisions, assumptions, constraints, commitments, context shifts), tags those signals to the same financial nodes, and evaluates every new number against its full strategy context. The loop closes when the evaluation informs the next meeting agenda.
How do you extract strategy signals from CFO meetings?
We use Fireflies to capture meeting transcripts, then run them through an extraction tool that identifies six signal types: goals, decisions, assumptions, constraints, commitments, and context shifts. Each signal gets tagged to the specific chart of accounts node it affects, creating a living strategy document that updates after every meeting.
What are the six types of financial strategy signals?
Goals (quantitative targets), decisions (explicit strategic choices), assumptions (beliefs underlying the plan), constraints (hard limits on operations), commitments (promises with deadlines), and context shifts (changes in business environment that affect assumptions). Each type has a lifecycle status: stated, active, at-risk, invalidated, or replaced.
How does this differ from a dashboard or BI tool?
Dashboards show you what happened. This system shows you what happened, what was supposed to happen, which assumptions broke, and what to do about it with a quantified P&L or cashflow impact. The strategy context comes from real meeting conversations, not manually maintained targets.
Can I build this for my own finance practice?
Yes. The architecture uses a chart of accounts structure, KPI definitions, meeting transcription (Fireflies), and an LLM-based signal extraction tool. Reply to the newsletter with "TEACH ME" if you want to learn how to build it for your own clients.