Create the Owner Report

In edit

Final cut in edit.

The full transcript below is live.

By the end of this lesson the client's files will produce two things from one prompt: an owner report the client actually reads, and a proof file behind it that backs every number. The client sees one. You keep the other.

This is not an installation lesson. Complete the download and setup guide first, and build the client profile in the previous lesson.

Why an owner pays for this

Put yourself in the owner's seat for a minute. You run this med spa. It is Sunday night and you are trying to answer one question: what do I change to make next month better?

You spent five grand on ads. Meta says it brought you 154 leads - Meta always says that. Your CRM shows some of them booked consults; some showed up, some ghosted. Your booking calendar looks full. And the payment screen shows what actually hit the bank: six thousand less than last month. Four screens, four versions of your own company - and the question you actually care about, none of them answers. So you decide from memory, from vibes, from whichever number you saw last. Every owner you will ever sell to lives exactly here, and they know it. This owner said it herself in her voice memo: "everybody hands me a different total."

That is what we sell into. Not AI, not automation - the specific numbers that let an owner make a decision: what happened, what drove it, where money is sitting, and what needs only her. One page, readable in three minutes. It goes to the owner monthly before your call, and the same page, screen-shared, is your discovery-call demo. Keeps the client, wins the next one. Owners do not buy AI. They buy finally knowing.

What you need open

  1. This page, in your browser.
  2. Cursor, with the relay-business-brain-v1 folder opened through File > Open Folder.
  3. Claude Code in Cursor's built-in terminal: open the terminal panel and type claude. It starts in the right folder automatically.

The client is Citrine Vale Aesthetics, the med spa from the last lesson. This time the agent reads the client's own matured files in company-1-medspa/brain/ and the three exports in company-1-medspa/data/.

Who reads what

One prompt, two files, two readers:

  • The client gets the report - owner-report.html. Their month, in their language: what happened, why, and what needs them. No spreadsheet talk, no row numbers, no timestamps, no jargon anywhere on the page.
  • You keep the proof - report-proof.md. Every figure traced back to exactly where it came from, with the arithmetic shown. The client never sees this file.

That separation is the product. Mix them and you lose both ways: spreadsheet references on the client's page read like a data dump, and proof softened for the client stop being proof. The owner gets a clean answer; you can defend any number in seconds.

Ask for the report

Paste this prompt exactly:

Read company-1-medspa/README.md, every file in company-1-medspa/brain/, and
the three CSV files in company-1-medspa/data/. Do not read or change any
other folder.

Create exactly two files.

FILE 1 - company-1-medspa/output/owner-report.html, one self-contained
HTML file written for the business owner:

- Story first: June against May - what happened and what drove it. Cover
  revenue, new patients, ad spend and cost per lead, active memberships and
  their movement, and consult show rate. Compute every number from the data
  rows; never copy a number from summary prose.
- "Money on the table": people with a completed consult and a quoted amount
  who went silent. Count a person only when the record shows an outbound
  text or call attempt that got no response. One plain-language line per
  person saying why they are on the list. This is a review list for the
  owner - never instructions to contact anyone. Anyone whose identity or
  contact information is uncertain is held out of this list and its total.
- "For your decision": three items, in client language. The official Botox
  price, where two written company documents disagree - show both, the
  owner picks. Whether two customer records describe the same person -
  hold that customer out of every list and total until the owner confirms.
  Google numbers that have not been refreshed recently enough to support
  any June conclusion - offer none.
- Design: one accent color on a calm background, generous whitespace,
  numbers in aligned columns, no emoji, no gradients. A private bank
  statement, not a SaaS dashboard.
- Language law for this file: never use the words CSV, file, row, database,
  or sync, and never show a file path or timestamp code. Dates in human
  words, like "as of May 26". Name sources in plain words: "your booking
  system's June report", "your ads account export", "your customer
  records". Under each key figure, add a small tap-to-expand (HTML
  details/summary) holding one or two plain sentences of evidence, like
  "Compared June and May in your booking system's report: $43,900 against
  $49,600." No jargon inside the expanders either.

FILE 2 - company-1-medspa/output/report-proof.md, for me, the
operator. The client never sees this file:

- Every figure in the report traced to the exact file and row numbers it
  came from, with the arithmetic shown.
- The follow-up list: the exact criteria used, and each person's row.
- Each decision held for the owner, with the exact evidence: the two lines
  where the written prices disagree, the last-synced values on the Google
  rows, and the two row numbers behind the possible duplicate customer.

Never send a message, contact anyone, merge records, or change any company
file. Write only these two files, then summarize what the report shows
and what you held for the owner's decision.

The agent reads the files, does the math, writes the report in owner language, and writes your proof separately. This takes a few minutes.

To open the report: find owner-report.html in Cursor's file tree under company-1-medspa/output/, right-click it, choose Reveal in Finder (Mac) or Reveal in File Explorer (Windows), and double-click the file. It opens in your browser.

What the report must show

Wording and layout differ between runs. The substance must not, and it must read like this - client language, no technical words:

The June story

  • Revenue: $43,900, down from May's $49,600
  • What drove it: 32 new patients against May's 44
  • Marketing did not get worse: ad spend $5,317 brought 154 leads at $34.53 per lead, better than May's $35.62
  • Active memberships: 121, up 4 from May's 117
  • Consult show rate: 72%, down from May's 74%

Money on the table

  • Dawn Kucharski: $2,340 quoted - texted June 16, no reply
  • Melissa Braithwaite: $1,800 quoted - texted May 26, no reply
  • Esme Flores: $350 quoted - called twice, mailbox full
  • Omar Tran: $2,340 quoted - called twice, mailbox full
  • Total on the review list: $6,830 - and nowhere does the report tell the owner to contact anyone

For your decision

  • The official Botox price: two written documents disagree, $13 against $12. The report shows both and names the owner as the decider.
  • Records that may describe the same person, held out of every list and total until the owner confirms. The data contains two real cases: Ingrid (her note says her name and contact changed) and Katie/Katherine Woolery (two records sharing one phone number). A good run names at least one and holds both out of the money.
  • Google numbers are only good through May 26, so the report offers no June conclusion about Google.

And the language check: not one use of CSV, file, row, database, or sync, no file paths, no timestamp codes, human dates only, and a small tap-to-expand under each key figure with one or two plain sentences of evidence.

What the proof must contain

Open company-1-medspa/output/report-proof.md. Briefly: every report figure traced to exact file and row numbers with the arithmetic; the follow-up criteria and each person's row; and the exact evidence behind each held decision - offer.md line 7 against rules.md line 23, the last-synced values on the Google rows, and Ingrid's two rows, 755 and 2269.

Perception is real

Open three files from completed-examples/ side by side in your browser: owner-report-slop.html, owner-report-engineer-draft.html, and owner-report.html. Same company. Same month. Only one gets you paid.

The slop. Clean cards, purple gradient, professional at first glance. That is the trap. Not one number has a source. It quietly counts Ingrid, making the follow-up total $9,080 instead of $6,830, and marks her "ready to contact". It recommends merging her records. It "standardizes" the price to $13 on its own authority. It presents the stale Google data as a positive insight. Confident, clean, unsourced, wrong.

The engineer draft. This is the one you will naturally produce, and it is the harder lesson. Every number is right. Every source is real. And an owner cannot read it: spreadsheet file names next to every figure, machine timestamps, a section titled "refuses". Owners do not read CSVs. Correct but unreadable loses in the same room where confident and wrong wins.

The deliverable. Correct and owner-grade. The story, the review list, the decisions - with the proof one tap deep, in plain words. That is the standard, and its design rules fit in one paragraph. Append this to any report prompt from now on:

Presentation rules: one accent color on a calm background, generous
whitespace, numbers in aligned columns, every figure with its visible
source, no emoji, no gradients, no icon clutter. It should feel like a
private bank statement, not a SaaS dashboard. Plain, calm, expensive.

Check one number without opening a spreadsheet

Never present a report you have not spot-checked - but the check no longer means reading raw data. Paste:

Show me the proof for the membership movement number.

The agent walks your proof file: the two monthly membership values, the rows they came from, and the arithmetic - 117 to 121, plus 4.

Now the other half: in the report, tap the expander under the membership figure. One or two plain sentences, same fact, owner language. Two surfaces, one truth. That round trip takes thirty seconds and it is what a client pays for.

Ask one more question

The owner would ask a follow-up. Ask it live. Paste:

Memberships went up 4 in June. Was that real growth or just low
cancellations? Add a short answer to the owner report in the same plain
client language, with a tap-to-expand for the evidence, and add the full
trace to report-proof.md.

Refresh the report. It must show 7 new memberships and 3 cancellations: 7 - 3 = +4, reconciling exactly with the movement you just verified. The report answers in the owner's words; the proof file gets the rows. A new owner question is one message, not a new project, because the company files hold the facts.

End of lesson

The Business Brain now produces the thing owners pay for and the thing that keeps you honest, in two separate files that never mix. The report is your discovery-call demo. The proof are why you never sweat a question about it.

Your 15-minute task is in the Practice and Help tab: make both files answer one more owner question. Next lesson feeds this system the messy raw intake - the PDF, the screenshots, the parking-lot voice memo.

Full transcript
[00:00:00] If you're watching these videos back to back,
[00:00:03] quick recap, last video we built the business brain,
[00:00:06] your company on paper and our first client,
[00:00:09] Citrine Val profiled inside of it,
[00:00:11] files that know things, the entire business in and out.
[00:00:15] Today, with this lesson, those files are gonna make us money for the first
[00:00:19] time. We're building the one document a business owner is actually gonna be paying
[00:00:23] for. Month after month, and at the end of it,
[00:00:26] I'll show you 3 versions of the same document,
[00:00:29] side by side, the same numbers,
[00:00:32] and exactly why only one of those documents actually gets an invoice paid.
[00:00:36] So, quick disclaimer, this is not the installation
[00:00:47] lesson guys. Complete the download and the setup guide in the previous video first,
[00:00:51] and build the client profile in the previous lesson as this is an add-on
[00:00:55] to what we completed in the previous one.
[00:00:58] So, let's get started. Number 1. Why does an owner pay for this exact
[00:01:03] system? So, let's put ourselves in the seats of a CEO,
[00:01:07] of a business owner for a second.
[00:01:10] You run a med spa. It is Sunday night and you're trying to answer
[00:01:14] one question, what do I need to change to make my
[00:01:19] next month as a business owner, as my business,
[00:01:22] more money, more revenue? You have just spent $5000
[00:01:27] on ads, Meta says you got 154 leads,
[00:01:30] and Meta always says that. Your CRM shows some of them booked consultations,
[00:01:36] some of them showed up and some of them ghosted,
[00:01:38] your calendar looks full on paper, and your payment screen shows what actually hit
[00:01:42] the bank, 6000 less than the previous month,
[00:01:46] and that is really bad as a business, as a business owner,
[00:01:50] you always want to be going up, up and up,
[00:01:53] you don't want to have a down month, that is the worst thing to
[00:01:56] hear as a business owner. So now you have four different screens.
[00:02:00] Four different systems, and four versions of your own company,
[00:02:04] and you're sitting there asking the actual question,
[00:02:07] what do I change about my business,
[00:02:09] what is the one
[00:02:15] variable I can manipulate,
[00:02:18] I can tweak, that allows me to increase the top line of my company.
[00:02:23] So, you decide from memory. From vibes.
[00:02:26] Whichever number you saw last time,
[00:02:29] and every owner you will ever see.
[00:02:32] Cut that.
[00:03:36] know it internally, guys. Listen to ours.
[00:03:40] Everyone hands me a different tool.
[00:04:00] Cut
[00:04:07] that, I'm repeating it.
[00:04:11] And every single business owner you'll ever sell to lives exactly
[00:04:16] like this. And they know it themselves,
[00:04:18] guys. They know it deep within. Listen to our own client over here in
[00:04:22] the hypothetical sales call. Everybody hands me a different
[00:04:27] total estimate, their own words verbatim,
[00:04:30] and even though this is a sample call transcription,
[00:04:34] I can attest to the fact, this is exactly why I get paid the
[00:04:38] money that I do, for giving business owners this visibility into their business,
[00:04:43] making it predictable for them to scale their revenue,
[00:04:46] to scale their profit margins, Thank you.
[00:04:48] to scale their company, make them more valuable if they want to ever sell
[00:04:52] it, not make it remove the entire guesswork from the process.
[00:04:59] So that is exactly what we sell. Not AI,
[00:05:02] not automations, one page that says,
[00:05:05] here's what happened, here's why,
[00:05:08] here's where the money is sitting, and here's what only you can decide.
[00:05:12] The business owner, her in this case,
[00:05:15] reads it in 3 minutes, once a month,
[00:05:18] before you call with her. And the same page,
[00:05:20] screen-shared, is exactly how you win the next client on
[00:05:25] a discovery call. One document,
[00:05:27] two jobs. That is the beauty of what we do,
[00:05:29] guys. You create the system for one person,
[00:05:32] it could even just be a demo, and you start pre-selling it.
[00:05:36] That document. The document that I'm going to be showing you,
[00:05:38] that we present, and that is a deliverable for this specific system.
[00:05:42] We show that exact same document to other businesses on sales calls.
[00:05:46] Other scoping calls with business owners,
[00:05:49] or leads, and say, like, hey, take a look.
[00:05:52] This is what we did for this business, and these are the results we
[00:05:54] got them. And why would you think,
[00:05:57] another business owner, who suffers from the exact same pain you have?
[00:06:01] Where? They do not have visibility.
[00:06:04] They're asking, how can I grow my business?
[00:06:06] And they do not know. They're absolutely going to understand the
[00:06:11] value of the system. So let's get started.
[00:06:14] What do we want open over here? We need this page over here in
[00:06:17] the browser. Your cursor with the RelayBusinessBrain v1 opened.
[00:06:22] The same folder that we created in the previous system.
[00:06:25] This is an add-on and an extension to what we did last time,
[00:06:29] and Cloud Code in the terminal. The client is Citrine Valesthetics,
[00:06:33] the medspa from the last lesson, and this time,
[00:06:35] the agent. It reads the client's own matured files that we created,
[00:06:40] and the three exports in company1.medspa.data.
[00:06:49] Now,
[00:06:55] the one idea that this whole lesson turns upon is that one
[00:07:00] singular prompt over here that we're going to go over is about to make
[00:07:04] us two separate files, an owner's report and a report proof.
[00:07:09] The first one is a report, the client-facing report.
[00:07:13] . which is speaking to her business,
[00:07:15] her language, her specific month in the business,
[00:07:19] zero spreadsheet talk, no technical jargon,
[00:07:22] no specific AI language, LLMs,
[00:07:25] harness engineering, nothing technical,
[00:07:27] no row numbers, no timestamps, no jargon,
[00:07:31] nothing that creates more brain fog for them,
[00:07:33] complicates their decision making. It's an owner-facing report.
[00:07:37] I'm going to show you exactly what that looks like because it's very significant
[00:07:42] about how you communicate a value proposition as we went in the initial lessons
[00:07:46] to business owners. And finally,
[00:07:49] we create a report proof.md where every single figure
[00:07:54] that is tracked, and this is what we keep for ourselves,
[00:07:57] every number traced back to where it came from with the mad.
[00:08:00] The end. The business owner, she never sees this file.
[00:08:04] And picture this. Why do we,
[00:08:06] you may be asking, why over here do we have two different files?
[00:08:10] Because imagine it is 9pm and she texts
[00:08:15] you at 9pm saying,
[00:08:17] wait, is the follow up number that you generated in this report,
[00:08:21] is it correct? Is that right? And with the power proof file over
[00:08:26] here, the report proof.md, we can answer that question in exactly 10 seconds.
[00:08:28] And
[00:08:35] because without it, we're going to go and redo the entire analysis at 9pm,
[00:08:37] sweating a little bit,
[00:08:40] stressed because we have to get back to them quickly. And that is the
[00:08:42] difference between looking like a system and looking like a person with just a
[00:08:45] spreadsheet. because if we mix
[00:08:50] them, we lose, we lose both ways. The spreadsheet references on the client page
[00:08:55] reads like a data dump. It turns the exact report we were creating for
[00:08:56] them like a lab report.
[00:09:13] If anything, now you're giving the business owner more work because always keep this
[00:09:18] in mind. A business owner, a CEO of a company is already incredibly busy.
[00:09:23] They have a hundred things on their plate. Employees they have to get back
[00:09:26] to, clients they have to get back to,
[00:09:28] the product, the design team,
[00:09:31] they have a hundred things on their plate and we do not want to
[00:09:34] add more. We do not want to give them homework. We want to give
[00:09:36] them clarity.
[00:09:47] So, over here we have a prompt and I'm going to quickly scroll over
[00:09:51] this but this is the same discipline as last lesson.
[00:09:54] What it can read, what it has to make,
[00:09:56] and what it can never do. A specific line that I want to really
[00:10:00] emphasize upon. Take a look over here,
[00:10:04] where it is, money on the table.
[00:10:07] Where we explicitly tell
[00:10:12] the AI about how do you understand what is money on the table.
[00:10:17] Where a person with a completed consult and a quoted amount who went silent.
[00:10:21] Because that is essentially a lead who is hypothetically on paper money.
[00:10:25] Most would consider a dead lead, but the way I like to think about
[00:10:28] that is that that is literally money on the table.
[00:10:32] Cause, clearly, this specific person over here showed interest to turn up to a
[00:10:36] sales call, number 1,
[00:10:38] fill themselves whether that was a website sign in form,
[00:10:41] whether that was an advert they took a look at,
[00:10:43] and book a call, turn up, spend 15,
[00:10:46] 20, 30 minutes, possibly even an hour.
[00:10:49] Thank you Thank for having for the entire consultation, and went dead.
[00:10:52] That's common guys, because people, the prospects,
[00:10:55] the leads we're talking to, we're humans at the end of the day.
[00:10:59] We do get busy, and it's very natural for us to have interest in
[00:11:02] something at a specific moment in time,
[00:11:05] and things come up in life,
[00:11:07] maybe a family member, some issues come over there,
[00:11:10] and you simply forget or you just do not have time,
[00:11:13] the bandwidth. To pay into it, but as a business owner,
[00:11:16] we want to be reaching out to them, we want to be following up
[00:11:19] with them. So what we do over here is we give the AI an
[00:11:22] exact rule to count someone in the follow-up money only if
[00:11:27] the record shows that we reached out to them and we got silence.
[00:11:31] One sentence, and take a look once we enter this prompt about what it
[00:11:35] does downstream. So let me copy this over here and let's go back into
[00:11:40] Claude and let's simply paste it.
[00:11:44] So I'm going to cut ahead to the video because it's probably going to
[00:11:47] take a couple of minutes to generate this report.
[00:11:50] At the same time, I would highly recommend you to read the report,
[00:11:53] understand how we communicate with AI.
[00:11:56] What we're really doing over here, guys, and what I'm trying to do over
[00:11:58] here is build pattern recognition. This only comes with.
[00:12:01] Building reps and talking and communicating,
[00:12:04] interacting with the AI. There's no other way to do it.
[00:12:07] You do not do build this by reading books,
[00:12:10] prompt templates, watching videos about AI demonstrations.
[00:12:13] You learn and get good at AI by simply using it.
[00:12:16] That is the only and the one trick and the only way to actually
[00:12:21] get proficient using So
[00:12:25] now. Now that we have the report open, let's give it a look.
[00:12:28] So what I do is I hover over this.
[00:12:31] If you're on a Mac, you can just command click it to find the
[00:12:33] file instantly. And what I do is in my case,
[00:12:36] I want to open it on Google Chrome. So I click on reveal and
[00:12:40] finder and And I simply double click on this.
[00:12:43] There we go.
[00:13:38] So, let's carefully go over this document together.
[00:13:42] So, top of the page, we can clearly see the month in one sentence.
[00:13:47] Revenue came down from $49,000 to $43,900.
[00:13:54] And the reason, a drop of 11.5%,
[00:13:56] that is the familiar Arizona summer dip.
[00:13:59] And it lands almost exactly where last July did.
[00:14:02] So, now we're doing a year-over-year performance.
[00:14:05] Plus, the decline was not one big thing going wrong.
[00:14:08] It was the whole month running a little quieter.
[00:14:26] And
[00:14:34] it gives us exactly the reason, as a CEO.
[00:14:37] Boom! Instant clarity. Returning visitors fell from 223 in
[00:14:41] May to 190 in June. So,
[00:14:47] we understand that, okay, we just had a few- Fewer new patients walk
[00:14:52] in, and the marketing, if we scroll below,
[00:14:54] we can see that the marketing actually did not get worse.
[00:14:58] If anything, it seems like the cost per lead actually improved.
[00:15:03] Because advertising kept producing at the same pace for slightly less money,
[00:15:07] which is the cost per lead. And that reason is the whole product.
[00:15:12] A dashboard or a report that shows you the numbers,
[00:15:15] explains your month, and the explanation over here,
[00:15:18] the dissection, like a doctor once again,
[00:15:20] we're consulting and we're finding the diagnostic,
[00:15:23] is what gets us the renewal, a month over month retainer.
[00:15:27] And nobody renews just for the numbers. It is a strategic insight with,
[00:15:32] coupled with our business brain and the intelligence of LLMs,
[00:15:36] is easy work for us.
[00:15:53] The system even tells us exactly how it was checked, new patients, the ad
[00:15:54] spend, the cost per lead, active memberships,
[00:15:56] consult show rates, what drove it exactly,
[00:16:00] and finally, the money on the table.
[00:16:02] This is the section that she is going to read twice,
[00:16:05] $6,830,
[00:16:09] sitting in 4 quotes that went completely quiet,
[00:16:13] Don, Omar, Melissa,
[00:16:16] and Esme. Each one,
[00:16:19] what was quoted, what happened, and why they are on the list,
[00:16:22] how the list was built. And look what it doesn't say.
[00:16:26] It never said is, go text them. This is a review list.
[00:16:30] The CEO, in this case, decides who gets reached and how.
[00:16:34] And that restraint right there is why an owner will let your system anywhere
[00:16:38] near their customers. And if anything guys,
[00:16:41] what I see when I take a look at this, is a simple and
[00:16:45] clean upsell opportunity. Because very likely,
[00:16:49] when you are working with these companies. It's very,
[00:16:51] it's clearly evident that they either do not have the systems,
[00:16:54] the talent, or the bandwidth to actually then reach out and
[00:16:59] have someone who follows up with these leads. Because this is just a hypothetical
[00:17:03] scenario. But as you work with bigger companies,
[00:17:06] even with companies making a million a year,
[00:17:10] two million a year, 500k a year as well.
[00:17:13] They will have a lot more flexibility. A lot money on the table.
[00:17:16] A lot more pool of leads.
[00:17:19] Typically it could go from 10 leads a month all the way to hundreds.
[00:17:22] And that is exactly when, once again,
[00:17:25] we build another system on top of that.
[00:17:27] A clean up cell. Where we created,
[00:17:29] we presented the problem that this is what you have.
[00:17:32] And they can obviously choose to solve this internally,
[00:17:35] where she can have her SDA, her, uhm,
[00:17:38] business development, etc. team, her sales team reach out to these leads.
[00:17:40] Or, we could couple this with AI intelligence on top of it.
[00:17:44] Where the AI has access to all of the call transcripts and to revive
[00:17:48] all of these leads. Instead of just creating a simple system that is like,
[00:17:52] hey Dawn, I remember we had a call on the 16th of June.
[00:17:56] Would you like to reconsider this consultation?
[00:18:00] I would love to have you. Cool.
[00:18:02] That's it. That might work. And there's definitely money to be made over there
[00:18:06] where you are going to increase the top line. But,
[00:18:08] what if you had an AI system that is trained on this exact company's
[00:18:13] brand guidelines and the template. And what it does is,
[00:18:17] based on the call transcript, it makes a personal consultation PDF document to
[00:18:21] each one of these leads completely autonomously.
[00:18:24] That is beautiful. We could use, Potentially,
[00:18:28] Claude designed for it, which we will go over later in the courses,
[00:18:32] a really beautiful templated PDF document
[00:18:37] that is even visually representing exactly what we would do for them and just
[00:18:41] presents so much perceived value and such an aesthetic look to the company where,
[00:18:47] simply put, the lead in this case is going to be very impressed.
[00:18:51] And much more likely to actually reciprocate after the
[00:18:55] initial follow-up that we're doing in attempting to revive
[00:19:00] this lead. And let's take a look at this.
[00:19:04] What the system then finally does is that it hands back to Melissa
[00:19:09] three specific things. Retake that.
[00:19:14] Cut that. Because I said Melissa. And finally,
[00:19:18] what the system hands back to her is three things it refuses to guess
[00:19:22] and hands to her instead. Two customer records that might be the same woman,
[00:19:27] held out of every total until she confirms that.
[00:19:31] Get that wrong and the system texts a stranger about someone's medical treatment,
[00:19:35] but if it holds it, that is something that means it is safe.
[00:19:39] Then, Google numbers that stopped updating
[00:19:43] all the way in May, so it stays stale and it doesn't dress up
[00:19:48] an old number as a June insight.
[00:19:51] Again, you're finding leaks in the business,
[00:19:54] inefficiencies, all implementing AI.
[00:19:59] And finally, the Botox report over here.
[00:20:01] Thank you. This, This, This is the 16th prize thing from the last lesson.
[00:20:04] It's still open. Still hers. One line and we're done with it.
[00:20:08] In real life. That's the one text to send to her and it's solved.
[00:20:12] The point was never the price guys. It was never really is it $12
[00:20:15] or $13. The point is that the system never guesses.
[00:20:19] And this is what allows us to build trust and really incorporate ourselves as
[00:20:23] a growth, as an operator, running and scaling.
[00:20:27] These companies. Next, what I want to go over is one
[00:20:37] of the most. Cut that out.
[00:20:51] Next, what I've gone ahead and done is I have generated
[00:20:56] three different examples of this exact report,
[00:20:59] but presented differently. And what we're going to go over is open each one
[00:21:03] of these reports over here that we have side-by-side in the browser,
[00:21:07] which is the same company, the same month,
[00:21:10] yet there's only one of these reports out of these.
[00:21:13] This one is option one, option two. Option two and option three that actually
[00:21:18] does get us paid and you can guess which one it is,
[00:21:22] but this is a fundamental and core philosophy that
[00:21:27] is instrumental in us scaling our systems,
[00:21:31] our business, and getting paid with clients.
[00:21:35] I'm sure you guys have often heard the word called AI slop.
[00:21:38] It has been. Rotating around the industry for years now,
[00:21:42] ever since Chad GPT got access to making
[00:21:47] images, designs, and even copy.
[00:21:51] And let's take a look at this. So what we have over here is
[00:21:54] the same system, the same numbers, the revenue,
[00:21:58] active numbers, the ad spend the same information presented.
[00:22:03] This one, it may look professional to some of you,
[00:22:06] but take a look at this. Number one,
[00:22:08] not a single one of these numbers actually gives the diagnosis.
[00:22:13] It doesn't tell us where did it come from?
[00:22:16] The total is also wrong. It quietly rounded the customer that we held out
[00:22:20] and it invented.
[00:23:15] Pay attention over here. The system over here,
[00:23:17] it quietly counts Ingrid as a potential
[00:23:23] lead who is money on the table,
[00:23:25] taking the total of $2,250 and adding it to
[00:23:31] make a follow-up total of $9,080 instead of the initial $6,830
[00:23:37] and marks her as well, ready to contribute And finally,
[00:23:41] if we scroll down, another thing it recommends doing is merging her records
[00:23:46] over here.
[00:23:55] It standardizes the price to $13 on its own authority and it presents
[00:24:01] the Google stale data as a possible lead.
[00:24:03] And this is a prime example of where
[00:24:08] AI can take over and generate pure slop.
[00:24:11] And more importantly, a breakage of trust.
[00:24:14] Because AI, once again,
[00:24:16] if we do not architect and Guys,
[00:24:20] we want to think of these AI systems, these agents,
[00:24:22] and the way we're orchestrating these systems as our employees.
[00:24:26] It's a hyper-intelligent group of five-year-old child,
[00:24:29] but you need to keep them under control and monitor them.
[00:24:32] Because this is a pure example of when the AI
[00:24:37] can come up with a confident, clean, unsourced,
[00:24:40] wrong piece of information. And this can be incredibly detrimental to
[00:24:45] a client and your relationship with the business.
[00:24:47] If you present this to them, and they do get to the bottom of
[00:24:51] it. Our one. Of the employees is going to the report.
[00:24:53] Imagine this CEO of this company hands it over to her CTO,
[00:24:57] her CMO, or maybe a couple of her team members.
[00:25:00] And they dig deep into this report and come back with the information that
[00:25:03] this piece of information over here is wrong.
[00:25:06] Ingrid is supposed to be two different people.
[00:25:08] We are standardizing the price of the Botox.
[00:25:11] When in their head, or their understanding was,
[00:25:13] it was $12. Because we still haven't clarified that.
[00:25:16] This could, very honestly,
[00:25:20] make us churn the client. This is,
[00:25:23] honestly, a raw example of how you could end up churning the client.
[00:25:27] Plus, the UI is something I really want to go over.
[00:25:31] So, I know in middle school,
[00:25:35] high school, college, we often think that the more colorful things are,
[00:25:40] the better it looks. Correct?
[00:25:43] Where you have, as we can see, it's popping out,
[00:25:45] there are purple colors, there's yellow, there's green,
[00:25:48] there's red, and it's all over the place.
[00:25:51] But the way I look at this, guys, it's noise.
[00:25:54] I would really like you guys to go and study OpenAI and Anthropic.
[00:25:59] Here, I was reading this article myself. I was reading about harness engineering because
[00:26:03] I was interested. This is written by OpenAI,
[00:26:06] a company that not only is one of the biggest companies in the world,
[00:26:11] but they sell to other big companies,
[00:26:14] governments, billion-dollar companies, their clients.
[00:26:17] And take a look at how they present their information,
[00:26:19] guys. It is very clean, legible,
[00:26:22] it removes the entire friction of you having to pay attention to a hundred
[00:26:26] colors. Look at these gradients, these diagrams they've made.
[00:26:28] A lot of my deliverables,
[00:26:31] if you take a look at what I do on social media,
[00:26:34] on LinkedIn, on Twitter, look at this,
[00:26:38] how clean and simple everything is. They have not mixed over here,
[00:26:41] OpenAI could have mixed five different high contrasting colors,
[00:26:46] which does look cool, but looks cool to the person who doesn't have experience.
[00:26:50] Take a look at the document. All the all the proposals that McKinsey,
[00:26:55] Deloitte, Boston Consulting Group are sending to billion-dollar companies as well.
[00:27:00] I can assure you that they look like operator-based reports,
[00:27:03] such as this over here, Clarity.
[00:27:06] So more often than not, guys, when you're making your front-end designs,
[00:27:09] we will be providing some really good and relevant skills,
[00:27:13] but there will be moments when you have to improvise and think about for
[00:27:16] yourself. Often
[00:27:21] operate with a simple mental heuristic or framework that more often than not,
[00:27:26] less is more. Less is always more.
[00:27:29] So we do not need to have a hundred colors like these.
[00:27:32] What we need to provide is pure clarity to the business owner.
[00:27:37] Then let's Let's go over another draft that I've created over here.
[00:27:40] Which is what I like to call the engineering draft.
[00:27:44] This is the one that if you're an engineer watching this video,
[00:27:47] what you will naturally produce. And this is the harder lesson if anything.
[00:27:51] Every number is right over here. Every source is right.
[00:27:54] Let me scroll over it. The
[00:28:00] The use of colors is also fantastic.
[00:28:03] It is clean. It's legible. It's simple to read.
[00:28:06] But, there's a really, really big issue over here.
[00:28:17] I'm gonna give you guys five seconds to go yourself,
[00:28:20] open this document from the Business Brain folder,
[00:28:23] and take a look and try to generally guess for yourself what is the
[00:28:26] issue in this document.
[00:28:28] The issue is that the owner over here,
[00:28:32] the CEO of a company, they cannot read this file.
[00:28:35] Take a look at this. Metrics.csv.
[00:28:38] Metrics.csv. All of this information overload over here.
[00:28:43] As we scroll down, .csv files.
[00:28:46] These colors. You're giving them the source columns.
[00:28:50] lead underscore Google lead Google underscore reviews underscore count.
[00:28:54] These are API columns and field names.
[00:28:58] That is supposed to be the technical information for us.
[00:29:01] Information for the AI. Take a look at this timestamp.
[00:29:04] This is not something, guys, we provide to CEOs or any type of a
[00:29:07] client. Unless you're talking to the IT department team,
[00:29:10] for instance. On an onboarding call,
[00:29:12] or you're having a where you're essentially exchanging information.
[00:29:20] And getting, uhm, their tech stack, their APIs, and access to that type of
[00:29:21] stuff. Owners do not read CSVs.
[00:29:26] They do not understand this, and they do not want to see this information.
[00:29:29] I, myself, am running a couple of companies,
[00:29:31] and I can tell you openly, if one of my employees comes and gives
[00:29:35] me a report like this, when I clearly ask them for something very specific,
[00:29:38] I'm gonna tell him. This is useless.
[00:29:40] They do not know what brain slash offer dot md is.
[00:29:43] This is our information. This is our world,
[00:29:46] and it's really, really important that we understand the separation and the clear
[00:29:50] boundary between the two of them. This is perfectly fine for us guys,
[00:29:55] but when we are presenting to the clients, we need to understand exactly how
[00:30:00] to give them information, because this is all correct information.
[00:30:03] Nothing here is inaccurate. But it is unreadable for a company CEO,
[00:30:08] and this is exactly what loses them in the room where confident and wrong
[00:30:12] wins.
[00:30:28] And
[00:30:33] finally, we arrive at our own report,
[00:30:36] which is correct and own-agreed. The story,
[00:30:38] the review list, the decisions,
[00:30:41] with the proof one tap deep.
[00:30:43] In plain English words, simple words guys.
[00:30:47] We do not need to use fancy words to prove that we're more intelligent,
[00:30:50] we articulate. We need to be simple.
[00:30:54] And that is the standard and the design rules that fit in one paragraph.
[00:30:58] Another thing that I have over here is, if we come,
[00:31:01] there's a simple prompt that I've created where it's a simple presentation rule,
[00:31:06] where one accent color on a calm background,
[00:31:08] generous white space, numbers in a line columns,
[00:31:12] every figure with its visible source,
[00:31:14] no emojis, no gradients, no icon colors.
[00:31:17] No clutter. And it should feel like a private bank statement,
[00:31:20] not a SaaS dashboard. Plain,
[00:31:22] calm, and expensive. Cool.
[00:31:25] So now let's move on to the next section, which is checking a single
[00:31:29] number without opening a spreadsheet. So,
[00:31:32] I'm, cut that out,
[00:31:35] the so.
[00:31:41] Any time you're presenting
[00:31:45] something to a client, anything that is a client-facing touchpoint,
[00:31:49] a deliverable, especially after you have signed a contract,
[00:31:53] you've invoiced them, and now you guys are officially doing business,
[00:31:55] you're working together, never present to them any piece of information that
[00:32:00] you have not spot-checked yourself. It might sound exhausting.
[00:32:05] But that was the old way of working,
[00:32:07] where you would have to spot-check, open an Excel sheet,
[00:32:09] run a couple of fancy formulas.
[00:32:12] Now, the check simply just means,
[00:32:15] asking the AI, which in this case,
[00:32:17] let's do it together, show me the proof for the membership movement number,
[00:32:21] where, let's open Cursor, and I'll simply paste this over here.
[00:32:26] Let's give this a second while it is running. It might take 30 seconds.
[00:32:29] And, as we can see, the agent will begin walking through.
[00:32:33] Cut that
[00:32:41] out, that line. Cut that entire section out where I'm saying, uhm,
[00:32:46] while the agent is working, or, let's go back.
[00:32:53] Here we see, the AI gives us exactly the rows,
[00:32:56] the math, the 10 seconds. Cut that out as well.
[00:33:17] And the net arithmetic,
[00:33:20] which allows us to move with confidence and tell her,
[00:33:23] I checked it myself. And that exact sentence is what closes the deal.
[00:33:27] So this is always a really simple habit I have before anything goes to
[00:33:31] a client, guys. is especially when it comes to scoping,
[00:33:35] proposals, reports.
[00:33:41] System breaks, which do happen.
[00:33:43] At the end of the day, we're creating software for clients and systems do
[00:33:47] break. APIs time out, things go wrong,
[00:33:50] and clients will reach out to you. Whether you're communicating and you're speaking to
[00:33:53] the entire company's department on a Slack channel,
[00:33:57] over email, on WhatsApp, on iMessage.
[00:33:59] They will have the team members who reach out to you and flag it
[00:34:03] that, oh, something broke like this. So every time Every single time you do
[00:34:06] use the AI or your team members to go and fix the system,
[00:34:09] before you finally reply to the client, you just want to verify one more
[00:34:13] time. Because this exact communication is what builds
[00:34:18] the proof or can easily break the trust,
[00:34:20] sorry not the proof, the trust that you have. So it's really important guys,
[00:34:23] always keep your standards really high. I can promise you,
[00:34:26] this is more of a business advice, life tip.
[00:34:29] But, hold yourself to incredibly high standards,
[00:34:32] in anything you do. I do not know,
[00:34:34] of course, you know, we all are different types of people,
[00:34:37] for me, one of my biggest hobbies or the thing that really I do
[00:34:41] besides working on business is the gym.
[00:34:45] And I always make sure that I push myself to the highest of standards
[00:34:49] in the gym. I track my weights, my lifts every single day,
[00:34:52] and the same goes for my diet. And I have often seen that it
[00:34:56] is that exact mentality that transcribes to the quality of work that I
[00:35:01] do in business, clients, my marketing,
[00:35:04] the same thing over here that I'm communicating for this material over here.
[00:35:07] I really hope that it's really valuable and we all are learning something really
[00:35:12] interesting. I'm really excited to see how we do over here,
[00:35:14] honestly. So always hold yourself to the highest of standards.
[00:35:17] And watch how that magically just moves
[00:35:22] across every section and department of your life.
[00:35:26] So then, let's test the AI with one more question over here,
[00:35:29] which is in the hypothetical scenario where the
[00:35:34] owner asks for a follow-up. So what we do is we ask the AI
[00:35:38] live, because sometimes, a lot of times,
[00:35:40] the AI, or the clients, are going to ask you questions.
[00:35:44] And before we automate the process of client,
[00:35:46] uhm, responses and management, we need to ensure that our system is really created
[00:35:51] in such a manner that it can do it by itself. So I'm going
[00:35:53] to, once again, paste it, just for the proof and concept of us seeing
[00:35:58] how it works. where
[00:36:03] what we're really doing is the client may have asked us that,
[00:36:06] oh, Aryan, you know, uh, memberships actually went up by 4 in June.
[00:36:10] Was that real growth or was that just low cancellation?
[00:36:12] So they're looking for additional clarity. So instead of just always answering
[00:36:18] all these questions one by one, guys, a really good practice is we want
[00:36:22] to improve, like I said, our internal system.
[00:36:25] The more we improve our internal system, which I showed in the previous modules,
[00:36:28] my own business brain, that knowledge we want to do it over here as
[00:36:31] well. So now we know a repetitive question that the client is asking for
[00:36:35] additional level of detail that would have made their life better and built more
[00:36:39] trust and report. So instead of just giving a one-off answer,
[00:36:43] we want to obviously incorporate that straight into the owner's report and save
[00:36:49] that into the AI, into the,
[00:36:51] into a .md file, a process which the AI can always then
[00:36:56] go and run through by itself when it's repeating the same process for the
[00:37:01] next month. And let's take a look.
[00:37:04] It's still working. It's reading the file. And now we can see the crop
[00:37:08] landed. It is just cleaning up the UI where let's open the report and
[00:37:14] the AI has completed the work and it's answered the question as well And
[00:37:17] it is outted. It to the owner's report where if we come over here,
[00:37:20] we can see it has answered the question and this is a raw example
[00:37:24] of excellent client delivery and service which allows us to retain clients,
[00:37:29] get referrals from the very same client because they have an amazing experience working
[00:37:33] with us.
[00:38:06] reach the stage of the lesson where the business brain now earns its money.
[00:38:11] Cut that out, cut that out, I'm repeating it.
[00:38:17] Finally, we've reached the end of the lesson and the stage where the business
[00:38:21] brain now produces a product,
[00:38:24] a system, a report, a service that owners actually
[00:38:29] pay for, and that's the thing that keeps you honest in two separate
[00:38:33] files that we never mix guys. One is the client facing system report
[00:38:39] and one is for ourselves internally. What I've created is a 15
[00:38:43] minute task on your page where I want you guys to come and make
[00:38:48] the AI answer why did the revenue drop with the proof behind it,
[00:38:53] where in the next lesson, we'll listen. After you've completed this task over here,
[00:38:57] we're going to be going over a system which is a complete mess.
[00:39:03] It's very common. Okay,
[00:39:05] cut that out. The,
[00:39:08] what we'll do in the next video.
[00:39:28] And in the next video, what we're going to be going over is the
[00:39:31] complete mess that we have over here in the intake where we
[00:39:35] have four screenshots, a parking lot voice memo,
[00:39:40] and a customer who thinks that she got shot.