by Andy Trattner
A podcaster named David Senra has read more than four hundred founder biographies. Ten of his episodes are on Elon Musk alone. Last March he opened a two-hundred-page book that another guy spent five years compiling out of Musk's own words, organized by maxim, and reported on his own show that he was finding patterns in it that ten episodes and many biographies had not given him.
That is strange and it is the whole point. A clip captures one move at one location. A book captures the moves across many locations. The pattern that connects the moves only becomes visible when you can hold all of them at once. This is also why one consultant with agentic LLMs can now do something McKinsey cannot.
McKinsey runs on clips. A few weeks of executive interviews. A sample of dashboards. The standard playbook applied to whatever ten or twelve data points the engagement budget allowed them to gather. The output is a deck of recommendations sized to the engagement fee, which is sized to the firm's overhead, which is sized to twenty thousand consultants in offices on three continents.
The model can't read your whole business. The cost structure doesn't allow it. So they read the surface, apply pattern-matching from other engagements, and hand you a snapshot. Six months later the assumption on slide 23 is wrong because the snapshot was already a snapshot when they delivered it.
This is the original sin of consulting. Not dishonesty. Just the math of large firms billing by the day.
A single operator with agentic LLMs has a different math.
The whole archive of a $5M business — every email, every P&L line, the standard operating procedures, the customer notes, the founder's own writing about why the company exists — fits inside one context window with room to spare. One mind plus a few well-aimed agents can hold all of it at once.
Holding all of it is the move. It is what reveals how the business actually decides — the assumptions, the inherited constraints, the trapped value. McKinsey can't read this because the engagement model doesn't price for it. One operator working for free in the first week absolutely can.
Your business has five years of accumulated decisions inside it. Maybe twenty employees, a few thousand customers. You have a head full of context you have never seen from the outside. The size that fits inside one consultant plus modern tooling is exactly the size most of America's actual economy runs at. The model the big firms use was never designed for this size. Now it doesn't have to be.
Here is what it looks like in practice.
Three calls of two hours each. The first one is reading, together, the things you have never put in front of a stranger. The second one is the model talking back: here is what your archive says about how your business actually decides, and here is the gap between that and what you say it does. The third one is the cut list. Three or four moves with high expected value and high reversibility. Pick the ones you want.
For a business with $5M in revenue, the floor I will guarantee is $200k of palpable gain in the first week. Not in the meeting. In the engagement. That number is not marketing. The trapped value is large because nobody has ever read the whole archive at once.
No program. No deck. No retainer. Flexible to whatever the week needs to be. The week is free. If it works, you tell a friend.
The reason this starts in Greenville and not Manhattan is structural, not modest.
Local is where the corpus is the right size. A regional business with twenty employees has a knowable archive. A Fortune 500 does not — its archive is too large for one mind plus today's tools to hold, even with the longest context windows on the market. McKinsey has the engagement budget to throw twenty people at a Fortune 500 and partially read it. One operator with agents has the bandwidth to fully read a Greenville business. Different scales, different feasible methods.
The opportunity isn't that big-firm consulting fails locally. The opportunity is that nothing has ever read most of these businesses whole. Reading the operating layer the first time is where the gain comes from. Each business you read carefully teaches you how the next one might decide. The reading compounds.
A McKinsey partner reads twenty clients in a career, billed at strategy-engagement scale. One operator with agents could read a hundred Greenville businesses in two years, billed at whatever the market for "$200k of palpable gain on your $5M business in a week" turns out to be. The compounding rate is different by an order of magnitude. The math is on the side of the corpus.
The thing I want a business owner reading this to take away is concrete.
I am here. I am in Greenville. I will work for free for the first week. The constraint is that you have to be excited about a project — open about the problems you would actually like a sharp outside reader to look at, and willing to spend the equivalent of three two-hour calls on it. I cannot do this for a thousand businesses at once, which is the point. The constraint is the feature. Depth on one business at a time.
If you want to know what your business says about itself when somebody reads the whole archive carefully — you can find out. The cost of finding out is one week of your time and zero dollars.
The corpus is small enough now. The reader is here.
Andy Trattner — Greenville, SC
P.S. The structural argument compresses a longer one. David Senra interviewing Eric Jorgenson, "How Elon Thinks," on Founders, March 24.