>gabes/the letter
the letter
01 · 00:00:00
// the weekly letter · issue 05 · jul 27 to 31

five signals.
one week.
none of this is
a tech problem._

a wire that left on a familiar face. a client list in someone's personal chatgpt. fourteen subscriptions doing four jobs. five failures, and not one gets fixed by better software. each cost money because a rule was never written and nobody's name was on it.
// today, live on the page

the newest signal, pulled the moment you land.

loading the latest signal.
01warningmonday

your bookkeeper is going
to get a video call from
you. your face, your voice,
your way of asking. and it
will ask them to move money.

one company sent $25 million out the door that way. the thing that stopped ferrari was a single question the fake could not answer.

in early 2024 a finance employee at arup, the engineering firm behind the sydney opera house, joined a video call with the cfo and several familiar colleagues, took the instructions, and made fifteen transfers totalling about $25 million into five hong kong accounts. every other person on that call was generated, built from video and audio the company had already published.

months later a ferrari executive got a call in the ceo's voice, southern italian accent included, pushing an urgent confidential deal. the whole thing collapsed the moment he asked the title of the book the ceo had recommended him days earlier. that is the entire defense, and it cost nothing.

the advice going around is to train your team to spot deepfakes, which asks a person to win a perception contest against software that improves every quarter, on a bad monday, under pressure from someone who sounds like their boss. the smaller the company the worse the exposure, because there is nobody standing between a convincing request and the bank login. write one rule and put it where the money is: any payment, any change to bank details, any transfer outside the normal pattern gets confirmed by calling back a number already in your own records, never the number that called you and never a reply in the same thread. then tell whoever holds that login, in writing, that they are allowed to make you wait ten minutes.

dowrite one rule where the money is, and give the person holding the login written permission to make you wait ten minutes.
don'ttrain people to spot the fake. that is a perception contest against software that gets better every quarter, judged on a bad monday.
the urgency is the attack. remove it and the whole thing stops working.
receipt arup, the engineering firm behind the sydney opera house: a finance employee made 15 transfers totalling about $25 million to five hong kong accounts after a video call where every other participant was generated, early 2024 · a ferrari executive ended the same attack by asking the title of a book the ceo had recommended days earlier
shipped monday ·xin
02truthtuesday

your best employee is
pasting the client list into
a chatgpt account you do
not own and cannot audit.

77% of employees do it. the login is personal, the history belongs to them, and you find out during a security questionnaire.

roughly three quarters of employees now paste company material into ai tools, and about the same share of workplace chatgpt accounts are personal logins with no admin, no retention setting, and no record you can produce.

what goes in is not exotic: pricing, a client list, an unsigned contract, a spreadsheet with names in it, pasted by the people who move fastest because nobody gave them a sanctioned way to move that fast. the bill arrives later and in someone else's words. a client asking where their data sits. a renewal stalled on a vendor questionnaire. an insurer comparing what your policy claims against logs you cannot pull.

the advice going around is to write an ai policy and block the tools, which does not stop the paste. it moves it to a personal phone where you have no visibility at all. do the version that holds: put every employee on a managed company account for one or two approved tools with training and retention switched off, and write a single page naming what may be pasted and what may not, in examples rather than categories. that is an afternoon of admin, and it closes the largest hole you currently have.

doput everyone on a managed company account for one or two approved tools, and write one page of examples: this may be pasted, this may not.
don'tban the tools. the paste moves to a personal phone and you lose the last of your visibility.
the rule only holds if the sanctioned path is faster than the personal one.
receipt about 77% of employees report pasting company material into ai tools · roughly three quarters of workplace chatgpt use runs on personal accounts with no admin controls and no retention settings
shipped tuesday ·xin
03proofwednesday

you did not buy an ai
strategy. you bought
fourteen subscriptions,
and four of them do
the same job.

78% of it leaders got an ai charge they did not forecast last year. the overlap is not a discount you missed. it is a job nobody owns.

the count is the tell. most companies now carry somewhere between eight and twenty ai or software lines and cannot say which one owns which job. 78 percent of it leaders were hit with an unexpected ai or consumption charge last year, because usage pricing means one power user or one looping workflow spends real money quietly and reports it a month later.

forrester has platform spend growing around 40 percent through 2026 against 5 percent for point solutions, and the companies that consolidated their stacks report cost reductions of 20 to 35 percent. that is not a negotiation win. it is the price of the duplication they were already carrying.

the advice going around is to run a software audit and push harder at renewal, which trims the invoice and leaves the sprawl exactly where it sits, ready to grow back by q4. do the version that holds: list every ai and software line, write beside each one the single job it owns and the person accountable for that job, and cancel anything whose job is already owned by something else. then put a spend alert on every usage-priced account so the meter tells you before the invoice does.

dowrite the job and the owner beside every software line, then cancel anything whose job is already owned by something else.
don'tstop at the renewal negotiation. a cheaper invoice for the same sprawl grows straight back by q4.
the saving is real. the point is the map.
receipt 78% of it leaders reported an unforecast ai or consumption charge in the last year · forrester: platform spend growing around 40% through 2026 against 5% for point solutions · 20 to 35% cost reduction reported after stack consolidation
shipped wednesday ·xin
// the week's argument, running

list what runs your business.
put a name beside it.
read what is left.

every signal this week ended in the same two lines of homework: name the person, write the rule. put in the things that already run in your company, write who owns the decision, and tap whether the rule is written somewhere a new hire could find it. nothing is saved or sent.
these rows are the week's five, seeded blank on purpose: edit every field, add your own. it updates as you type.
the thing that runswho owns ittap: is the rule written?
04shiftthursday

the entry-level job at
your company now requires
judgement you have
never written down.

ai-exposed entry roles are seven times more likely to demand senior skills. you are still hiring for a rung that stopped existing.

pwc read over a billion job ads across six continents and found the bottom of the ladder moved: ai-exposed entry-level roles are now seven times more likely to ask for judgement, leadership, and decision-making, skills that used to arrive in year five.

the cost lands on you twice, because the cheap junior who used to absorb the rote work is now expected to show up with taste, and the rote work they were supposed to learn on has already been automated out from under them.

the advice going around is to pay up for ai skills, and the market is not wrong about the price: the ai wage premium hit 62 percent this year, up from 57. paying more does not close the gap, because what you are actually buying is a person to exercise judgement your company has never made explicit, and no salary band compensates for that. do the version that holds: take the three decisions your newest people get wrong most often, write the rule, the exception, and one worked example for each, and put them somewhere a person on day four can find without asking.

dowrite the rule, the exception, and one worked example for the three decisions juniors get wrong most, and put it where day four can find it.
don'tpay a premium for judgement you have never described. no salary band compensates for an unwritten standard.
you cannot hire your way out of judgement your company never wrote down.
receipt pwc global ai jobs barometer, over 1 billion job ads across six continents: ai-exposed entry-level roles 7x more likely to require judgement, leadership, and decision-making · ai skills wage premium at 62%, up from 57%
shipped thursday ·xin
05absurdfriday

the ftc just banned an ai
sales company from ever
selling to businesses again.

they promised small businesses an ai closer. the settlement says the earnings pitch was the actual product.

in march the ftc barred air ai and its owners from marketing business opportunities, after charging that the company sold small businesses on ai sales agents using income claims it could not support.

the comedy is structural, not personal: a company selling software to replace your sales team was itself running a sales pitch no software could have made good on. the cost does not land on the people who wired them money, it lands on the whole category, because buyers now expect the sequence to be demo, number, subscription, with the working system optional.

gartner expects more than 40 percent of agentic ai projects to be scrapped by 2027, and almost none of that is the model failing. it is that nobody could run it in production once the vendor stopped presenting. the advice going around is to vet vendors harder and ask for more case studies, which is asking a sales process to police itself. do the version that holds: before you sign anything, write down the workflow it touches, the single number it should move, and the person inside your company who owns it on the monday after go-live.

dowrite three things before you sign: the workflow it touches, the one number it must move, and the person who owns it on the monday after go-live.
don'task for more case studies. that is asking a sales process to police itself.
a vendor who cannot answer those three in writing is selling you the demo.
receipt ftc action against air ai and its owners, march: barred from marketing business opportunities over unsupported income claims made for ai sales agents · gartner expects over 40% of agentic ai projects to be scrapped by 2027
shipped friday ·xin
// take it with you

five pages nobody wrote.
here is how to write them.

every signal this week ended in one page of writing. these five prompts each produce one of those pages. paste into any assistant (chatgpt, claude, gemini) and use today. each card says what it does and what you get back.
no email. no signup. no follow-up sequence waiting for you. take it and go.
01
warning
writes the payment rule that sits where the money is, in one page.
you get the callback rule, the transfers it covers, and the written permission to make someone wait.
02
truth
writes what may be pasted into an ai tool and what may not, in examples.
you get the paste list, the never list, and the sanctioned path that has to be faster.
03
proof
puts one job and one owner beside every software line you pay for.
you get the duplicate jobs, the cancel list, and the accounts that need a spend alert.
04
shift
turns the three calls your juniors get wrong into a page they can read on day four.
you get the rule, the exception, and one worked example for each of the three.
05
absurd
writes the three answers a vendor has to give you before you sign.
you get the workflow, the single number, and the named owner for the monday after go-live.
one file. opens in any notes app. drop it next to the work. no gate, no catch, nothing waiting for you after.
saved. that is the point of us: the value shows up before the invoice does.
// the signal daily, the letter weekly

five failures.
five pages.
none of them written.

one signal a day, monday to friday, public. the letter every friday, the same hour the fifth signal lands: the week compiled, connected, and pointed at what to do. no quizzes, no funnels, no webinar.
good. fridays, then. bring us the part that keeps breaking.
>gabes · operating systems for companies done improvising · issue 05 · past issues → · past signals →
> scroll. watch what gets claimed.