WEBVTT

00:00:00.000 --> 00:00:04.120
So, get this. A client recently paid $2 ,600

00:00:04.120 --> 00:00:09.320
for a custom AI automation solution. That sounds

00:00:09.320 --> 00:00:11.539
like, you know, a big, complicated project. But

00:00:11.539 --> 00:00:14.759
here's the paradox. It only took about two hours

00:00:14.759 --> 00:00:17.359
to actually build. Wow. Today, we're diving into

00:00:17.359 --> 00:00:19.980
why something so simple, we're talking tools

00:00:19.980 --> 00:00:22.480
like Google Sheets, can generate that kind of

00:00:22.480 --> 00:00:25.160
business value. Welcome back to the Deep Dive,

00:00:25.160 --> 00:00:28.519
everyone. Our mission today. is to crack open

00:00:28.519 --> 00:00:31.839
that exact blueprint. We are looking at a system

00:00:31.839 --> 00:00:36.060
that completely automated a client -student onboarding.

00:00:36.320 --> 00:00:38.560
It just took a huge administrative headache and

00:00:38.560 --> 00:00:41.039
turned it into this, well, this hands -off engine.

00:00:41.240 --> 00:00:43.659
Okay, let's unpack that. So this system, it's

00:00:43.659 --> 00:00:46.799
a really tight four -stage workflow. It replaces

00:00:46.799 --> 00:00:49.500
a total mess of spreadsheets. It eliminates something

00:00:49.500 --> 00:00:53.020
like 120 hours of manual work a year. And this

00:00:53.020 --> 00:00:54.740
is the most important part. It makes sure no

00:00:54.740 --> 00:00:57.399
new student ever gets forgotten. So we're going

00:00:57.399 --> 00:00:59.340
to walk through those four stages step -by -step.

00:00:59.380 --> 00:01:01.039
We'll cover the tech stack, which is actually

00:01:01.039 --> 00:01:03.579
deceptively simple. And then we'll get into the

00:01:03.579 --> 00:01:06.400
essential 10x ROI calculation. that made that

00:01:06.400 --> 00:01:10.439
$2 ,600 price tag a total no -brainer for the

00:01:10.439 --> 00:01:13.099
client. Here is the core philosophy, I think,

00:01:13.099 --> 00:01:16.219
that makes that price possible. Clients are not

00:01:16.219 --> 00:01:19.819
paying for your clean code or the fancy AI model

00:01:19.819 --> 00:01:23.480
you picked. They pay because a really painful

00:01:23.480 --> 00:01:25.840
recurring problem just disappears from their

00:01:25.840 --> 00:01:28.900
business. That is the critical distinction. You

00:01:28.900 --> 00:01:31.780
have to quantify the pain first. For this client,

00:01:31.959 --> 00:01:35.780
the old way... Took about 30 minutes per student.

00:01:35.879 --> 00:01:38.680
Per student. Right. And over a year, that added

00:01:38.680 --> 00:01:41.000
up to, what, two and a half hours every single

00:01:41.000 --> 00:01:43.599
week just wasted on manual follow -ups and tracking.

00:01:43.819 --> 00:01:46.299
And the real cost isn't just the time, right?

00:01:46.400 --> 00:01:48.719
It's the inconsistency, the burnout on the team.

00:01:48.879 --> 00:01:51.040
Exactly. And the student experience suffers.

00:01:51.299 --> 00:01:53.140
Yeah. When a student feels forgotten, they're

00:01:53.140 --> 00:01:55.180
more likely to churn. Those are the soft costs

00:01:55.180 --> 00:01:57.560
that really hurt. Totally. And this is the mistake

00:01:57.560 --> 00:02:00.120
so many people make. They charge hourly for the

00:02:00.120 --> 00:02:03.060
build time. The build was two hours. If you charge

00:02:03.060 --> 00:02:05.599
for your effort, you're cheap. But if you focus

00:02:05.599 --> 00:02:08.719
on the outcome, this system saves the client

00:02:08.719 --> 00:02:13.599
$6 ,000 a year in labor alone. So a $2 ,600 fee

00:02:13.599 --> 00:02:16.960
against a $6 ,000 saving. It's just a good math

00:02:16.960 --> 00:02:19.340
equation for them. It's an easy yes. But how

00:02:19.340 --> 00:02:21.259
do you defend that math? I mean, what if the

00:02:21.259 --> 00:02:23.280
client says, wait, this took two hours, but it

00:02:23.280 --> 00:02:25.699
costs this much? You just explained that you

00:02:25.699 --> 00:02:28.300
sold them $6 ,000 worth of time and reliability,

00:02:28.659 --> 00:02:31.539
not two hours of your time. You sold them the

00:02:31.539 --> 00:02:33.939
permanent deletion of a headache. And that system

00:02:33.939 --> 00:02:36.039
is what we're calling an AI agent. Yeah, which

00:02:36.039 --> 00:02:37.659
is really just a simple way of saying it's a

00:02:37.659 --> 00:02:40.180
workflow that uses logic powered by something

00:02:40.180 --> 00:02:43.750
like. NNN plus a little bit of language AI for

00:02:43.750 --> 00:02:45.930
the personalized stuff. So if the value is all

00:02:45.930 --> 00:02:48.330
about deleting a business problem, how do we

00:02:48.330 --> 00:02:50.569
shift our thinking from two hours of build time

00:02:50.569 --> 00:02:53.789
to $2 ,600 in justified revenue? Focus on the

00:02:53.789 --> 00:02:56.069
pain relief and the annual labor savings, which

00:02:56.069 --> 00:02:59.370
showed a clear 10x return on investment. So now

00:02:59.370 --> 00:03:01.449
that we get the justification, let's look under

00:03:01.449 --> 00:03:03.949
the hood. Let's start with stage one. This is

00:03:03.949 --> 00:03:07.949
the instantaneous first impression. The moment

00:03:07.949 --> 00:03:10.530
a student pays. The system just becomes their,

00:03:10.530 --> 00:03:13.509
what, their digital concierge. Right. Reliability

00:03:13.509 --> 00:03:16.050
from the very start. So the automation kicks

00:03:16.050 --> 00:03:19.189
off three things at once. First, a warm welcome

00:03:19.189 --> 00:03:22.310
email goes out. With the account link? Yep. Second,

00:03:22.370 --> 00:03:24.629
a quick message pops up in the team's Slack channel.

00:03:24.729 --> 00:03:28.370
Just, new student, Jane Doe, paid. So everyone's

00:03:28.370 --> 00:03:31.030
in the loop. Everyone's aligned. And third, it

00:03:31.030 --> 00:03:33.189
creates a crucial entry in the Google Sheets

00:03:33.189 --> 00:03:35.389
CRM. OK, I have a question here about the tech.

00:03:35.689 --> 00:03:38.449
The source says the automation engine calculates

00:03:38.449 --> 00:03:40.729
the follow -up dates. Why not just use a simple

00:03:40.729 --> 00:03:43.960
Sheets function? Like today, plus three. Yeah,

00:03:44.000 --> 00:03:45.939
that's a great question. And you could do that.

00:03:46.039 --> 00:03:49.539
But by having the engine, n8n, handle the date

00:03:49.539 --> 00:03:52.020
math before it writes to the sheet, you guarantee

00:03:52.020 --> 00:03:54.080
those dates are fixed. They're set in stone.

00:03:54.300 --> 00:03:56.400
Ah, so they won't recalculate later. Exactly.

00:03:56.439 --> 00:03:58.719
You remove that variable. It makes the timing

00:03:58.719 --> 00:04:01.180
for the next stages totally rigid and reliable.

00:04:01.520 --> 00:04:03.280
That makes a lot of sense. You're building in

00:04:03.280 --> 00:04:06.159
reliability upstream. And that flows right into

00:04:06.159 --> 00:04:09.000
stage two, the day three automatic follow -up.

00:04:09.229 --> 00:04:12.590
Exactly. So stage two just runs quietly in the

00:04:12.590 --> 00:04:14.430
background. It checks the sheet every day. And

00:04:14.430 --> 00:04:16.850
it's looking for what? It's looking for any student

00:04:16.850 --> 00:04:20.550
who is still marked as account creation and whose

00:04:20.550 --> 00:04:23.189
day three follow bait has passed. If it finds

00:04:23.189 --> 00:04:26.910
one, boom, a gentle personalized reminder email

00:04:26.910 --> 00:04:29.829
gets sent. It's amazing how much business friction

00:04:29.829 --> 00:04:33.129
is just caused by, you know, simple human memory

00:04:33.129 --> 00:04:35.959
problems. Oh, absolutely. And, you know. Honestly,

00:04:36.040 --> 00:04:37.839
while this sounds simple, getting these schedule

00:04:37.839 --> 00:04:40.100
checks right always requires some pretty rigorous

00:04:40.100 --> 00:04:42.300
testing. Right. Even with these low -code tools,

00:04:42.459 --> 00:04:44.519
I still wrestle with prompt drift myself sometimes

00:04:44.519 --> 00:04:46.939
when I'm trying to get complex logic right. You

00:04:46.939 --> 00:04:49.180
just have to be meticulous. So by eliminating

00:04:49.180 --> 00:04:51.439
that manual calendar work and using fixed dates,

00:04:51.579 --> 00:04:54.060
what's the single biggest win for reliability

00:04:54.060 --> 00:04:56.939
in these first two stages? Eliminating manual

00:04:56.939 --> 00:04:59.959
calendar calculation prevents human error, ensuring

00:04:59.959 --> 00:05:01.980
no student follow -up is accidentally missed.

00:05:02.639 --> 00:05:04.500
OK, so let's move into the really intelligent

00:05:04.500 --> 00:05:08.199
part of this system. Stage three, the human escalation

00:05:08.199 --> 00:05:10.860
on day five. Right. So we sent the welcome. We

00:05:10.860 --> 00:05:13.240
sent the day three reminder. But if after five

00:05:13.240 --> 00:05:15.800
full days, the student still hasn't set up their

00:05:15.800 --> 00:05:18.720
account, the automation stops. It knows its limits.

00:05:18.899 --> 00:05:21.579
That feels like a key design choice instead of

00:05:21.579 --> 00:05:24.480
just spamming them. Exactly. It performs a strategic

00:05:24.480 --> 00:05:27.740
pivot. It sends an urgent red flag alert to the

00:05:27.740 --> 00:05:30.379
team in Slack. and updates the sheet to human

00:05:30.379 --> 00:05:33.019
notified. We call that intelligent escalation.

00:05:33.139 --> 00:05:35.680
That's it. The agent handles the boring, repetitive

00:05:35.680 --> 00:05:38.740
stuff. A human only steps in when it's clear

00:05:38.740 --> 00:05:41.279
that simple automation isn't enough. It's the

00:05:41.279 --> 00:05:44.680
sweet spot. That intelligent handover. That really

00:05:44.680 --> 00:05:47.519
is where the high ROI comes from. And that leads

00:05:47.519 --> 00:05:50.319
us to the high touch part, stage four. The personalized

00:05:50.319 --> 00:05:52.740
AI kickoff. So when the student finally does

00:05:52.740 --> 00:05:54.860
create their account, they fill out a form, right?

00:05:54.980 --> 00:05:57.879
Business name, team size, and their main goals.

00:05:58.100 --> 00:06:00.699
And that data is immediately fed into the AI

00:06:00.699 --> 00:06:04.600
agent. This blueprint uses Claude 4 .5 Sonnet.

00:06:04.879 --> 00:06:07.420
which is just, it's excellent understanding context

00:06:07.420 --> 00:06:09.980
and following really specific instructions. And

00:06:09.980 --> 00:06:12.740
the agent's job is to generate a fully personalized

00:06:12.740 --> 00:06:16.399
welcome email, not a template. Right, it references

00:06:16.399 --> 00:06:19.319
their specific goals. But this is where the builder

00:06:19.319 --> 00:06:22.480
needs to be smart. You instruct the AI to format

00:06:22.480 --> 00:06:26.079
the output as a structured JSON object. Okay,

00:06:26.139 --> 00:06:28.100
what does that mean in simple terms? It just

00:06:28.100 --> 00:06:30.339
means it's a clean technical structure with clear

00:06:30.339 --> 00:06:32.519
labels, like a key for subject and a key for

00:06:32.519 --> 00:06:36.019
body. It's mandatory for reliability. Whoa. I

00:06:36.019 --> 00:06:38.160
mean, imagine scaling that. That level of detailed,

00:06:38.319 --> 00:06:40.879
personalized attention to a billion queries without

00:06:40.879 --> 00:06:43.120
ever needing a human editor. It's true automation

00:06:43.120 --> 00:06:45.639
at scale. Yeah. By telling the AI you need that

00:06:45.639 --> 00:06:48.060
JSON structure, you guarantee the automation

00:06:48.060 --> 00:06:50.680
tool gets clean, usable data every single time.

00:06:50.920 --> 00:06:53.100
The Gmail node isn't going to get some weird

00:06:53.100 --> 00:06:56.259
broken text. The data is pristine. So when designing

00:06:56.259 --> 00:06:58.399
high touch content, how does using a structured

00:06:58.399 --> 00:07:01.560
output like JSON prevent errors in the final

00:07:01.560 --> 00:07:04.019
execution of the workflow? Structured output

00:07:04.019 --> 00:07:06.720
guarantees the automation tool receives usable

00:07:06.720 --> 00:07:09.779
and correctly formatted email components for

00:07:09.779 --> 00:07:12.259
the final send, eliminating formatting errors.

00:07:12.699 --> 00:07:14.399
Let's pull back and just look at the blueprint

00:07:14.399 --> 00:07:16.899
itself for a second. This whole system, which

00:07:16.899 --> 00:07:20.360
is handling thousands of dollars in value, it's

00:07:20.360 --> 00:07:24.279
built on a really lean, low code stack. NEN is

00:07:24.279 --> 00:07:27.519
the engine. Google Sheets is the CRM. Gmail sends

00:07:27.519 --> 00:07:31.019
the emails and Slack for alerts. That's it. The

00:07:31.019 --> 00:07:33.519
simplicity is the entire point. You don't need

00:07:33.519 --> 00:07:35.959
a computer science degree. You just need to understand

00:07:35.959 --> 00:07:39.680
that simple sequence. Trigger A actions. Update

00:07:39.680 --> 00:07:42.060
status. It's like stacking Lego blocks. And the

00:07:42.060 --> 00:07:43.899
source really emphasizes you have to start with

00:07:43.899 --> 00:07:46.439
the mindset, not the tools. It says, don't start

00:07:46.439 --> 00:07:49.240
by opening NAM. Start with the pain. Step one

00:07:49.240 --> 00:07:51.519
is always identifying the friction. You have

00:07:51.519 --> 00:07:53.399
to look for tasks that are repetitive, annoying,

00:07:53.600 --> 00:07:55.740
and, you know, that people mess up all the time.

00:07:55.879 --> 00:07:58.459
Like chasing invoices. Exactly. Or an e -commerce

00:07:58.459 --> 00:08:00.139
store constantly answering, where's my order?

00:08:00.199 --> 00:08:01.920
Right. If someone on the team is complaining

00:08:01.920 --> 00:08:04.279
about forgetting something or chasing someone

00:08:04.279 --> 00:08:07.879
or checking five different tabs for data, that's

00:08:07.879 --> 00:08:10.060
your target. That's your high ROI opportunity

00:08:10.060 --> 00:08:12.220
right there. And that leads to the mindset shift.

00:08:12.360 --> 00:08:14.959
The wrong thinking is believing you need advanced

00:08:14.959 --> 00:08:17.740
code or the newest, fanciest language model.

00:08:17.939 --> 00:08:20.759
That's just chasing tech for its own sake. It

00:08:20.759 --> 00:08:23.319
is. The right thinking is realizing the goal

00:08:23.319 --> 00:08:26.540
is just to solve real problems and deliver measurable

00:08:26.540 --> 00:08:29.360
ROI. This agent worked because it was simple

00:08:29.360 --> 00:08:32.100
and it solved a $6 ,000 problem. It delivered

00:08:32.100 --> 00:08:35.940
high value by getting rid of 120 hours of manual

00:08:35.940 --> 00:08:39.559
work. The 10X ROI is right there. So the takeaway

00:08:39.559 --> 00:08:42.299
for anyone listening is to just prototype fast.

00:08:42.559 --> 00:08:44.779
Yeah. Sketch the workflow out, build a really

00:08:44.779 --> 00:08:47.059
simple version, maybe just an email follow -up

00:08:47.059 --> 00:08:50.100
agent, test it with some fake data. And record

00:08:50.100 --> 00:08:52.889
a quick three -minute loom video. Clients don't

00:08:52.889 --> 00:08:54.990
care how complicated the back end is. They just

00:08:54.990 --> 00:08:57.070
care that the friction in their day disappears.

00:08:57.190 --> 00:08:59.210
If you can show them that in three minutes, the

00:08:59.210 --> 00:09:01.450
sale is practically done. Outside of onboarding,

00:09:01.450 --> 00:09:03.009
if you're walking into a new small business,

00:09:03.330 --> 00:09:05.889
what is the single easiest bottleneck to identify

00:09:05.889 --> 00:09:08.210
that suggests an immediate automation opportunity?

00:09:08.730 --> 00:09:11.470
Any repetitive task that involves someone complaining

00:09:11.470 --> 00:09:13.950
about forgetting, chasing, or manually checking

00:09:13.950 --> 00:09:17.610
data is the prime low -hanging automation target.

00:09:17.889 --> 00:09:20.070
Okay, let's recap the big idea from this deep

00:09:20.070 --> 00:09:23.950
dive. The success of this $2 ,600 agent came

00:09:23.950 --> 00:09:26.990
down to a perfect fit. It was simple technology

00:09:26.990 --> 00:09:30.070
NEN and Google Sheets eliminating a massive and

00:09:30.070 --> 00:09:33.429
easily quantifiable pain point. That's it. 120

00:09:33.429 --> 00:09:36.649
hours of manual labor a year. Gone. And the key

00:09:36.649 --> 00:09:39.230
is the pricing philosophy. You have to anchor

00:09:39.230 --> 00:09:42.370
your fee to that 10x ROI for the client. You

00:09:42.370 --> 00:09:44.629
turn the sale into a simple math equation for

00:09:44.629 --> 00:09:46.570
them, showing them they're buying back time and

00:09:46.570 --> 00:09:49.659
scale, not just... you know, lines of code. It's

00:09:49.659 --> 00:09:52.240
all about solving those inefficiencies that businesses

00:09:52.240 --> 00:09:54.820
just quietly live with every day. Yeah. If you

00:09:54.820 --> 00:09:56.720
focused on making those problems disappear, the

00:09:56.720 --> 00:09:58.879
gap between two hours of your work and a significant

00:09:58.879 --> 00:10:01.080
fee starts to feel very small. That's right.

00:10:01.320 --> 00:10:03.820
And the blueprint is built on that elegant handover.

00:10:03.919 --> 00:10:06.340
The AI handles repetition. The human handles

00:10:06.340 --> 00:10:09.419
the empathy. That balance is everything. So here's

00:10:09.419 --> 00:10:11.519
a final provocative thought for you to chew on.

00:10:11.679 --> 00:10:14.200
If your ultimate goal is to scale your business

00:10:14.200 --> 00:10:16.179
without needing to hire a new administrative

00:10:16.179 --> 00:10:19.240
employee, are you diligently tracking the manual

00:10:19.240 --> 00:10:22.159
time cost that a simple AI agent could eliminate

00:10:22.159 --> 00:10:24.700
before the need for that new hire becomes critical?