Two weeks ago I posted a breakdown of the Apple lawsuit against OpenAI. Seven details from the filing, including the 400 former Apple employees now working at OpenAI and the engineer who kept access to internal servers after leaving.
That post reached 457,000 people. The comments split into two camps arguing about who was the villain, and almost nobody picked up the part that actually matters for anyone running a company.
Apple did not lose a database. Apple lost people, and the knowledge those people carried was never written down anywhere Apple could keep.
Your company has the same problem at a smaller scale. It shows up in revenue before it shows up anywhere else.
in
“Apple just sued OpenAI. The details are wild. Here are the 7 craziest facts from the lawsuit: 1. OpenAI has hired over 400 former Apple employees, including design legend Jony Ive.”
What a GTM agency actually fixes
Most founders hire a GTM agency expecting more meetings. That is the pitch on every website, mine included at some point, and it is the least interesting thing the work does.
Here is what actually happens on a good engagement. The founder has been selling for two or three years. They close at a rate nobody else in the company gets close to. They know which objection means the deal is dead and which one means the buyer is interested but scared. They know the three questions that separate a real opportunity from a polite conversation.
None of it is written down. It lives in one head, gets applied by instinct on every call, and disappears the moment that person stops taking calls.
So the first job is extraction. We sit inside the calls, we read the closed-won and closed-lost threads, and we turn instinct into something a second person can run. The meetings come later, and they come from a system rather than from one person's talent.
The knowledge that leaves when people leave
Every company I work with has some version of this list.
The best SDR knows which job title actually replies in each vertical. Nobody wrote it down.
The founder knows the pricing floor and the exact moment to hold it. Nobody wrote it down.
Someone in customer success knows which onboarding question predicts churn nine months out. Nobody wrote it down.
When those people leave, and they always leave, the company does not just lose headcount. It loses the compressed version of every mistake it already paid for. The replacement gets to pay for those mistakes again, on the company's time.
Apple had to rebuild an entire design group. A twelve-person B2B company loses a head of sales and rebuilds pipeline from scratch for two quarters. The mechanism is identical.
How we turn a founder's playbook into a system
The process is boring, which is why most teams skip it. It runs in four steps.
1. Record everything for thirty days. Calls, the reasoning behind the pricing decisions, the messages that got replies. The raw material has to exist before anyone can systemize it.
2. Find the repeatable moves. Out of a hundred calls there are usually six or seven patterns that decide most outcomes. The rest is noise that feels important in the moment.
3. Write them as instructions a new hire can execute in week one. If it needs a paragraph of context to make sense, it is a principle. If it can be run, it is a play.
4. Put the plays where the work happens. In the CRM, in the sequence, in the call prep doc. A playbook nobody opens is the same as no playbook.
The AI layer helps at step two and step four. Agents read months of call transcripts and surface patterns faster than any human reviewer, and they keep the plays updated as the market moves. The judgment about what belongs in the system still comes from the people who closed the deals.
Where a GTM agency does not help
I want to be straight about this, because the category is full of promises that do not survive contact with reality.
A GTM agency cannot fix an offer nobody wants. If you have not closed anything yourself, there is no playbook to extract, and systemizing an unproven pitch just makes you wrong at scale. Go sell it yourself first. Twenty conversations is usually enough to know.
It cannot fix a product with a retention problem. Filling the top of a leaking funnel is expensive theater, and everyone involved can see it in the numbers by month three.
It also does not work when the founder stays out of the room. The knowledge we are extracting is theirs. I have watched engagements stall because the person holding the actual playbook delegated the extraction to someone who never sold the thing.
What the numbers look like
The honest version, from our own engagements at Growth Cab.
Extraction takes three to five weeks before anything measurable happens. Founders hate this part, and the ones who quit during it never see the return.
After that, the pattern is consistent. A second seller reaches roughly seventy percent of the founder's close rate within a quarter, against six to nine months of slow bleeding without a documented playbook. Ramp time is the metric that moves first, and it is the one that compounds.
The economics follow from ramp. A rep who takes nine months to produce costs a full year of salary before contributing. A rep who takes three months costs a quarter of that and starts feeding the pipeline while the market conditions still resemble the ones the playbook was built on.
The uncomfortable part
Documentation feels like overhead when things are working. That is exactly when it is cheapest to do.
Apple is spending years and a lawsuit trying to recover something a set of written processes would have kept. It is a large company with unlimited legal budget and it still cannot get the knowledge back, because the knowledge was never in a form anyone could keep.
Your version costs less and hurts sooner. One person leaves, and two quarters of revenue go with them.
The work is unglamorous. Sit in the calls, write down what works, put it where people can run it. Whether a GTM agency does it or you do it internally matters far less than whether it gets done at all.
Your IP walks on two legs. Write it down before it walks.
