Luke Pierce

in CRM since Aug 2026

8 videos · 30-day rolling · period 2026-08-19 → 2026-11-17

Active videos

Scheduled in period

Delivered this mo.

Delivered all-time

Open slots

Undated

Owes now

Findings

Calendar — this client, rolling week

Owes right now — 7

voicenotePaste This Into Claude, Never Write The Same Email Twiceno date

voicenote100 Days Running My Business On AI Agents in 21 Minutesno date

voicenoteThe Automations Agencies Overcharge Fordue 09-23

voicenoteGrok Bot Full Courseno date

voicenoteX Things I'd Automate Firstdue 09-08

voicenote1 AI Expert vs 10 Hatersdue 09-04

voicenoteOne Business, One Daydue 09-21

Upcoming — 9 active

2026-09-12AI Full CourseEsrajul HaqNeeds scripting

script (4.7k chars)

# Research

Warning before pitching this: the huge multipliers on this format come off tiny channel medians, not off huge audiences. Productive Dude's 271x is 898,930 views against a 3,300 median. The number is real but it reads bigger than the audience is. The opposite case is the tell: Nate Herk ran a 2 hour Grok Bots course into 971K subscribers on Aug 31 2026 and it landed at 1.02x, barely his own baseline. Big channels do not get paid for completeness, small and mid ones do. Luke sits in the band that still wins. Second thing to know before he commits: every proof below that cleared 250K views runs past two hours. This format is paid for in runtime, and a 40 minute version of it is not the same video.

Proves title 1: "FULL [Subject] Tutorial For Beginners in 2026! (FULL COURSE)"

271.5x · FULL Claude Tutorial For Beginners in 2026! (FULL COURSE) · @productivedude · 169K subs · 898,930 views vs 3,300 avg · Published Apr 13 2026

youtube.com/watch?v=Xg55nTrbYYY

301.24x · FULL Claude Tutorial for Beginners in 2026! (Become a PRO!) · @ai-foundations · 355K subs · 1,224,884 views vs 4,058 avg · Published Mar 9 2026

youtube.com/watch?v=rRrBbyv3ChM

10.39x · FULL Claude Course for Beginners in 2026! (Become a PRO!) · @iamaimaster · 309K subs · 301,916 views vs 28,480 avg · Published Jul 13 2026

youtube.com/watch?v=Fys4oHlXQmQ

Three channels, three sizes, same sentence. The skeleton survives the swap from Claude to anything, which is why it is title 1.

Proves title 2: "[SUBJECT] FULL COURSE X HOURS: [Outcome] (2026)"

Delivered — 0 all-time

Nothing posted yet.

Setup

Scripts byViewz

Voicenote cadenceRolling

Filming rhythmRolling

CRM statusOnboarded - paying

Open in Notion CRM ↗Money → Finances

2026-09-15Build Your Own JarvisFarhan Alam PushonNeeds scripting

script (3.7k chars)

# Research

Proof top-up still needed on this card. The Velio daily quota ran out part way through this rebuild, so the captions below carry only the fields that were genuinely captured on the earlier pull: multiplier, exact title, channel, views and publish date. Subscriber count, channel average and the @handle are missing and have not been guessed. Two things to fix on the next Velio pass: fill those three fields, and find two more proofs sharing the title 1 skeleton, because title 1 currently stands on one video and the rule is three. Until that is done, title 1 is the strongest of the three but it is not fully proven.

Two proofs that were on this card before have been removed. The AI Workshop Iron Man video used a banned word in its published title, so it cannot sit on a Luke card. The Proton video shares no skeleton with any of the three titles, it was category colour rather than proof.

Proves title 1: "How To Make A Jarvis"

43x · How To Make A Jarvis · Jared Rhod · 152,193 views · Published Jul 7 2026

youtube.com/watch?v=cQqOkx5qnWo

The highest multiplier on the card off the shortest title on the card. Four words, no tool named, no promise. That is the whole skeleton and it is why title 1 barely changes it.

Proves title 3: "[Tool]: Zero to Personal AI Assistant (1 Hour Course)"

3.4x · Hermes Agent: Zero to Personal AI Assistant (1 Hour Course) · Nate Herk · 348,757 views · Published May 10 2026

youtube.com/watch?v=gb5TlGw6Uks

Proves title 2: "How I Turned [Tool] Into My Personal Assistant (Full Guide)"

17x · How I Turned Claude Into My Personal Assistant (Full Guide) · AI Edge · 284,942 views · Published May 14 2026

youtube.com/watch?v=CvLkhGWZlvE

Every proof on this card builds an assistant for one person and one inbox. Not one of them builds the assistant that knows a company's pipeline, its margins and its open jobs. That gap is the reason this idea is worth Luke's time rather than being a rerun.

# Concept

Luke builds the assistant that knows his own business, on camera, in X steps, and then puts it on the spot with live questions at the end.

For each step, he covers:

•Which of his own systems he plugs in here, and what that system holds that a stock assistant has no way of seeing

•What this step cost to add, one off or monthly, at the tier he is actually paying for

•The question it can answer after this step that it could not answer before, asked out loud

•What it got wrong the first time he asked, and whether he fixed it or decided to live with it

•Who else in his business is allowed to ask it things, and what he had to lock down before he let them

•The step he built first, threw away, and rebuilt in a different order

The ending reframes it: he asks it the questions it still cannot answer, then says which of those he would not want it answering anyway. The point is not a machine that runs the business, it is that he stops being the only person who knows the numbers.

Production: desk sit down with the screen mirrored, one continuous build, then a live question round to camera at the end. No set build needed.

# Titles

1.How To Make A Jarvis For Your Business

1.How I Turned Claude Into My Operations Manager (Full Guide)

1.Hermes Agent: Zero to Business Assistant (1 Hour Course)

Title 3 keeps the proof's tool name. Swap it for whatever Luke actually builds on and the skeleton still holds.

# Thumbnails

2026-09-19Grok Bot Full Courseno editorNeeds voicenote

script (7.7k chars)

⚠️ SKELETON. Nate is filling this out. The structure, the proof and the honest risk are here. The module list, the runtime and the client outcomes are not, because those are Luke's facts and the skill forbids inventing them.

⚠️ MERGED Aug 31 2026. This card absorbed Grok Bot Is Now Only $20 - 9 Wild Use Cases For Your Business and took its Sept 19 slot. The nine use cases are now the spine of the middle modules, and that card's outlier proof and its title were carried across.

⚠️ TIME SENSITIVE, this is why it took the earliest slot. The Grok Bot category is live right now and decaying. The proof video published Aug 27 and did 416,767 views in four days. The multiplier is currently going to small and mid channels, which is Luke's band: Nate Herk's two-hour Grok course did 1.02x, Matthew Berman's use-case video 0.97x, 9to5Mac 0.62x, all of them below their own channel average. If this slips past September it should be dropped rather than published late, because being early is the entire advantage.

# Research

⚠️ READ THIS FIRST. There is one piece of evidence directly against this idea and it should not be buried.

Nate Herk, "Build & Sell Grok Bots (2 Hour Course)", 47,514 views, 1.02x on a 902K subscriber channel. That is the exact video this card proposes, made by one of the biggest channels in the niche, and it did precisely its own average. Nothing more.

The case for making it anyway is that the two halves of this card are proven separately and the failure above may be a channel-size problem rather than a format problem. Every big channel that has touched Grok Bot has come in at or below its own average: Matthew Berman 0.97x, 9to5Mac 0.62x, Nate Herk's weekly recap 0.64x. Every breakout has gone to a small or mid channel. Luke is in the band that is still winning. That is a real argument, but it is an argument, not proof, and Nate should decide with the 1.02x in front of him.

Proves title 1: the FULL COURSE skeleton, which is the strongest thing on this card

301.2x · FULL AI Automation Tutorial For Beginners in 2026! (FULL COURSE) · AI Foundations · 355K subs · 1,224,884 views vs 4,058 avg

271.5x · FULL AI Automation Tutorial For Beginners in 2026! (FULL COURSE) · Productive Dude · 169K subs · 898,930 views vs 3,300 avg

10.39x · AI Master · 309K subs · 301,916 views vs 28,480 avg

⚠️ These three are carried over from the AI Full Course card and need their thumbnails and exact published dates pulled when Velio is back. The multipliers come off very low medians, so treat 301x as evidence the sentence works, not as a view forecast.

Proves title 2: "[Tool] Is Now Only $X - Here Are 9 Wild Use Cases"

8.0x · Grok Bot Is Now Only $20 - Here Are 9 Wild Use Cases · Paul J Lipsky · 476K subs · 416,767 views vs 48,641 avg · Published Aug 27 2026

youtube.com/watch?v=UyMJBUCyDIs

Proves the compression variant and the install-and-judge angle

3.11x · How to Master Grok Bot in 16 Minutes · 10x Builder · 2,210 subs · 1,056 views vs 159 avg · Published Aug 28 2026

youtube.com/watch?v=C6eItTdAuNw

3.68x · Grok Bot Is The First AI Agent You Just Install. Is It Worth $200? · Nate B Jones · 317K subs · 142,719 views vs 38,498 avg · Published Aug 14 2026

youtube.com/watch?v=LM7Ft7g8qJw

1.94x · Grok Bot Just Got 70% Cheaper... (I Switched) · Brock Mesarich · 139K subs · 52,984 views vs 26,920 avg · Published Aug 23 2026

youtube.com/watch?v=BFZ7lyKYcg8

Category demand, not a format model

36.5x · Introducing Grok Bot · Grok (official) · 5,910 subs · 158,895 views vs 4,300 avg · Published Aug 11 2026

youtube.com/watch?v=F1_0Lkp16Rc

⚠️ The Grok Bot CATEGORY is hot but the multiplier is going to small and mid channels. Several of the biggest Grok Bot videos underperform their own channel: Nate Herk's 2-hour course did 1.02x, Matthew Berman's use-case video 0.97x, 9to5Mac 0.62x. Luke at his size is in the band that is still winning. Grok companion bots have no usable data at all, one video at 176 views, so that angle stays off the table.

⚠️ Still to pull when the quota is back: proof for the compression variant applied to Grok specifically, meaning the "learn 95% of X in N minutes" shape, and confirmation of whether any small channel has already run a Grok course and broken out. If one has, this card gets much stronger. If several have tried and missed, kill it.

# Concept

⚠️ SKELETON. Nate fills the remaining four jobs, the runtime and the real client outcomes.

Luke teaches Grok Bot end to end for a business owner rather than for a developer, which is the gap. Every other course on this tool is made for people who already know what an agent is. This one assumes the viewer owns a company, has never installed anything like it, and wants to know what it does for the business by the end of the afternoon.

The structure is the thing that makes it different from the ones that missed. Not a feature tour. The course is taught by building, and the middle of it is nine real business jobs run through Grok Bot at the $20 tier, in order, each one a job a company actually pays someone to do rather than a demo. Each job gets the before price and the after price on screen, and each module ends with something running.

Modules, to fill in:

•What it is and what it costs, in one plain paragraph, with the $20 tier on screen

•Install and first run, unedited, including whatever goes wrong

•The nine jobs, one module each or grouped, picked from work Boom Automations already does for clients so every one is a job he can prove. Named so far:

•the quote follow-up nobody sends on day three

•reading a supplier PDF and pulling the numbers into a sheet

•turning a voicenote from a job site into a written scope

•the weekly report a manager builds by hand every Friday

•first-pass triage on an inbox that is not his own

•four more, TBD, Luke picks them from live client work

•For each job: the build, the run, and then what breaks

•The three jobs it should NOT be given, named, with the reason. Three of the nine are not worth it and he says which, because the channel rule is that the failure gets shown or the claim does not get made

He closes on the honest ceiling: at $20 a month this replaces a task, not a person. What it replaces, what it does not, what the viewer should still hire a person for, and that anyone telling a business owner otherwise is selling something.

Why this is for his audience

His audience owns a home or professional services business and is being told daily that AI agents will run their company. They have no way to judge that claim and no time to test it. A full course from someone who actually runs an automation agency, that ends with them having built something and being told plainly what it cannot do, is the version of this video nobody in the niche is making. It also feeds Boom Automations directly: the viewer who follows along and stalls on module four is exactly who books a call.

Production: screen recording throughout, real account, real business data with a privacy pass before it ships.

# Titles

⚠️ Arm 3 is unproven pending the Velio pass. Arms 1 and 2 are direct rips of proven skeletons.

1.FULL Grok Bot Tutorial For Beginners in 2026! (FULL COURSE)

1.Grok Bot Is Now Only $20 - Here Are 9 Wild Use Cases For Your Business

1.Learn 95% of Grok Bot in 31 Minutes

# Thumbnails

2026-09-221 AI Expert vs 10 Hatersno editorNeeds voicenote

script (4.7k chars)

# Research

Read before this gets pitched. Four of the five proofs belong to one channel with roughly 11M subscribers running a studio format with a cast, a set and a crew. Those multipliers are not available to Luke and quoting them at him would be dishonest. The proof that actually carries this card is CLASH: a small channel, a 1 vs 10 cast instead of 1 vs 20, a narrow topic, and it still pulled 298,439 views at 5.3 views per subscriber. That is the version Luke can produce. Note that CLASH's multiplier was never captured, only its views per subscriber, so the caption below says so rather than inventing one.

Second thing: "Surrounded" is Jubilee's format name, not a generic tag. Title 3 keeps it because the rip rule says keep the structure, but using it in public would read as a copy of somebody else's show. Treat title 3 as the shape to imitate, not the words to publish.

Proof top-up still needed. The Velio daily quota ran out during this rebuild. Subscriber counts, channel averages and @handles are missing from every caption below and have not been guessed.

Proves title 1: "1 [Position] vs [N] [Opposing Group]"

25x · 1 Liberal Man vs 20 MAGA Women (ft. Dean Withers) | Surrounded · Jubilee · 9,626,266 views · Published Jun 21 2026

youtube.com/watch?v=odBbDQj-GRs

7.9x · 1 Anti-Manosphere vs 20 Manosphere (ft. Mark Manson) | Surrounded · Jubilee · 2,879,243 views · Published Jul 19 2026

youtube.com/watch?v=KevTBc07wm8

Multiplier not captured, 5.3 views per subscriber · 1 Analytics Nerd vs 10 Passionate NBA Fans · CLASH · 298,439 views · Published Jun 19 2026

youtube.com/watch?v=83lkTsCla4k

Three proofs, same sentence. The CLASH one is the one to copy: it proves the structure holds at 1 vs 10, on a niche subject, on a channel Luke's size.

Proves title 2: "[Named Person] vs The [Movement]"

11.8x · Jillian Michaels vs The Body Positivity Movement | Surrounded · Jubilee · 7,343,781 views · Published Mar 15 2026

youtube.com/watch?v=M7K87rGoGps

Proves title 3: "[Named Person] vs [N] [Group] | Surrounded"

3.1x · Tim Pool vs 20 Progressives | Surrounded · Jubilee · 721,756 views · Published Aug 16 2026

youtube.com/watch?v=CO_mqGbyq_Y

Title 2 needs Luke's name to already mean something to the person scrolling, so it is ranked below the position-versus-group version. Title 1 works on strangers, title 2 works on people who already know him.

Why the seat suits him: the anti-AI position is the most loaded take in business right now and nobody has staged it. Luke's brand is calm, low ego, proof over hype, and this format punishes anyone who tries to dunk. He does not win by being louder, he wins by opening a laptop when a claim can be settled on screen.

# Concept

Luke sits in the middle and ten business owners who think AI is oversold take turns trying to change his mind, and every time an argument can be settled he opens a laptop instead of talking.

For each owner, he covers:

•Their objection in their own words, and whether he has heard that exact one from a paying client

•His answer, and whether it is something he can put on screen in under two minutes or only assert

•The build he opens when the answer needs proof, named tool, real account, real client data with the names off

•What that build costs to run and what breaks in it, said out loud before the owner gets to say it

•Whether the owner sits down or stays standing, and what changed their mind if they moved

•The one objection in the room he agrees with

The ending reframes the whole thing: he closes on the objections he could not answer, names the two things the owners were right about, and names the one job he tells owners not to automate at all. He loses on purpose, in public, because that is the only version of this format that his audience would trust.

Production: this one needs a shoot. One room, ten chairs in a ring, one seat in the middle, a laptop on the table wired to the screen so builds go up live. The real cost is casting: ten genuine business owners who hold the position, not actors reading it, or the format collapses on the first answer.

# Titles

1.1 AI Operator vs 10 Skeptical Business Owners

1.Luke Pierce vs The Anti-AI Movement

1.Luke Pierce vs 10 Skeptical Business Owners | Surrounded

# Thumbnails

2026-09-26X Things I'd Automate Firstno editorNeeds voicenote

script (2.4k chars)

# Research

Driver: a numbered list where every item is a real thing working on screen, not a tool review. The count in the title sets the promise and the viewer stays to see all of them.

"Set Up Claude Cowork better than 99% of people"

Systems Made Better, 616,015 views, 666x outlier, Published Apr 10, 2026

youtube.com/watch?v=pl90LATQlHI

"How I Use Claude Cowork to Automate 99% Of My Life (Full Guide)"

Paul J Lipsky, 371,621 views, 16.6x outlier, Published Mar 8, 2026

youtube.com/watch?v=cJ2QCyr3yzU

"8 Insane Claude Cowork Use Cases! (automate anything)"

Go9x, 234,469 views, 105x outlier, Published Apr 7, 2026

youtube.com/watch?v=gp3d7RAgFME

"How to Use Claude Cowork – Full Workflow Automation Guide 2026"

AI Master, 207,237 views, 13.1x outlier, Published Apr 16, 2026

youtube.com/watch?v=SNo_recKZyY

"How To Scrape Any Website With Claude Cowork"

Andrew Dunn, 104,897 views, 219x outlier, Published Mar 8, 2026

youtube.com/watch?v=lsI5hP3Nq28

"Fully Automate Social Media with Claude Code in 5 Minutes"

NexGens AI, 76,543 views, 214x outlier, Published Apr 24, 2026

youtube.com/watch?v=IR4buVTRpEg

Why it transfers: six channels, 13x to 666x, all inside two months. Only one of them puts a count in the title, so the count is not what is carrying these, the specificity is. Every version is a generic list for anyone. Luke's is ordered by hours returned and pulled from real Boom installs.

# Concept

Luke names the 10-20 automations he installs first in any business, ordered by hours returned.

For each automation he covers:

•What it does and who it replaces

•What it costs and how long it takes to build

•Hours a week it gives back

Counts down so the biggest hours saved lands last. Ends on the one most owners skip and shouldn't.

Funnel: Top.

# Titles

1.The First X Things I'd Automate In Any Business

1.X Insane Automations Every Business Owner Should Steal

1.Set Up Your Business AI Better Than 99% Of Owners

# Thumbnails

2026-09-29Paste This Into Claude, Never Write The Same Email Twiceno editorNeeds voicenote

script (3.5k chars)

⚠️ THIS CARD FAILS THE PROOF BAR. Read before scheduling it.

The skeleton "Paste This Into [tool], Never [do thing] Again" is proven ONCE, by Sharbel A. at 9.55x, 124,613 views on Aug 19 2026. One video. The rule is that title 1 needs at least three proof videos sharing its exact skeleton, and this has one. Velio ran out of quota on Aug 31 before more could be found, so the card is sitting on a single data point.

It may well be fine. Sharbel's is his single best-performing recent upload and the mechanic is obvious: the title promises one transferable object rather than a lesson. But one proof is not a validated formula, it is one channel getting lucky once, and the card should not be treated as proven until the skeleton is found on two more channels.

Title 1 was also rewritten Aug 31. It read "Never Write A Proposal Again", which was too narrow: it only lands on businesses that send proposals, and it made the video about one document rather than one prompt. "Never Write The Same Email Twice" is the broader, plainer version of the same promise and it applies to every owner watching.

# Research

Proves title 1: "Paste This Into [Tool], Never X Again"

9.55x · Paste This Into Claude Code, Never Run Out Of Tokens Again · Sharbel A. · 27.8K subs · 124,613 views vs 13,048 avg · Published Aug 19 2026

The single best-performing recent upload on Sharbel's channel. The title promises one transferable object, not a lesson. Zero abstraction.

Proves title 2: "X [Things] You NEED To Install Today"

5.34x · 10 Hermes Agent Skills You NEED To Install Today · Sharbel A. · 27.8K subs · 73,908 views · Published Aug 16 2026

Proves title 3: "X [Hacks] That Will Change Your Life"

6.83x · 7 Hermes Desktop Hacks That Will Change Your Life · Sharbel A. · 32,821 views · Published Jun 21 2026

Why this format, specifically

Sharbel went from 19.6K to 27.8K subs in three weeks on this exact mechanic. Across his top ten the constant is that the payoff is a thing you install or paste, never a concept you understand. Luke's audience is business owners rather than developers, so the artifact has to be a business artifact, but the title mechanic ports directly.

# Concept

One prompt, on screen, that a viewer can copy in the first ninety seconds. Luke pastes it live and it writes a real client proposal from a discovery call transcript, in front of the viewer, with the client's actual numbers in it.

Structure:

•the prompt goes on screen first, before any explanation, so the copy-paste crowd gets what they came for and stays for the reason

•he runs it on a real discovery call he has permission to show, and the proposal comes out priced

•he then breaks the prompt into its four moving parts and explains why each line is there, so the viewer can bend it to their own business rather than just run his

•he runs it a second time on a call that is a bad fit, and it produces a bad proposal, and he shows exactly which line of the prompt caused it

The last third is the part nobody does: what it costs to run, how long it took him to get the prompt right, and the two things he still writes by hand every time.

# Titles

1.Paste This Into Claude, Never Write A Proposal Again

1.10 Claude Prompts You NEED To Install Today

1.7 Automation Hacks That Will Change Your Business

# Thumbnails

2026-10-03100 Days Running My Business On AI Agents in 21 Minutesno editorNeeds voicenote

script (2.2k chars)

# Research

Proves title 1: "100 Days With X in 21 Minutes"

11.54x · 100 Days With Hermes Agent in 21 Minutes · Sharbel A. · 55,478 views vs 4,807 avg · Published Jun 17 2026

His highest multiplier in the top ten. A long duration collapsed into a short number.

Proves title 2: "I Have Spent 1000+ Hours With X. This Is What I Learned"

5.01x · I Have Spent 1000+ Hours With Claude Code. This Is What I Learned · The Coding Sloth · 628K subs · 277,508 views vs 53,296 avg · Published Aug 18 2026

Proves title 3: "Learn 95% of X in 31 Minutes"

6.54x · Learn 95% of Hermes Agent in 31 Minutes · Sharbel A. · 94,237 views vs 14,105 avg · Published Aug 9 2026

The mechanic

Precision is the hook. "100 Days" and "95%" carry the whole title. The generic version, learn this fast, is dead on arrival. Luke has the one thing most people making these cannot produce, which is a real business that has actually been running on the stack for months.

# Concept

A hundred days of Boom Automations running on agents, compressed into a single sitting, told as a log rather than a tutorial.

The spine is a running scoreboard on screen: hours saved, dollars spent, things that broke. It updates as he moves through the hundred days, so the viewer is watching a number move instead of listening to claims.

What gets covered, in order of when it actually happened:

•the first two weeks, where he tried to hand over too much and had to walk it back

•the week something failed silently for four days and cost a client a deadline, with the actual message on screen

•the point where the cost curve crossed and it became cheaper than the alternative, with the invoice

•what he now refuses to let an agent touch, and why, named specifically

He ends on the number of hours it genuinely gave back per week, stated plainly, including the fact that it is smaller than the number people in his niche quote.

# Titles

1.100 Days Running My Business On AI Agents in 21 Minutes

1.I Have Spent 1000+ Hours With AI Agents. This Is What I Learned

1.Learn 95% of AI Automation in 31 Minutes

# Thumbnails

2026-10-06One Business, One Dayno editorNeeds voicenote

script (1.9k chars)

# Research

Driver: watching a real person get diagnosed in public. Not advice, an actual business taken apart on camera while the viewer measures their own against it.

"How I'd Make $100K in 90 Days From Scratch"

MoreMozi, 233,200 views, 66x outlier, Published Jul 26, 2026

youtube.com/watch?v=YF-___fAKAk

"Business is hard until you build systems like this"

Shan Hanif, 230,677 views, 83x outlier, Published Apr 20, 2026

youtube.com/watch?v=4tV_jwspQAs

"turning a bad personal brand aesthetic [full client transformation]"

Casper Verbeek, 142,422 views, 38x outlier, Published May 16, 2026

youtube.com/watch?v=KcEqfGnr9KY

"Watch This If A Copycat Is Stealing Your Business."

MoreMozi, 115,556 views, 46x outlier, Published Jul 1, 2026

youtube.com/watch?v=oGgGkbmAIPU

Why it transfers: these are consulting teardowns, not automation videos, so the format is being carried across rather than copied. Scored down for that. It stays because it is the only concept here that produces a real Boom case study as a byproduct.

# Concept

Luke highlights a real business and installs 8-12 automations in a single day.

For each automation he covers:

•The manual job it replaces and how long that job took

•The build, on screen

•Hours a week it hands back

One of them fails on camera and stays in the edit. Ends on the total hours saved.

Funnel: Middle.

# Titles

1.I Gave This Business Owner 20 Hours A Week Back

1.Business Is Hard Until You Build Systems Like This

1.I Automated A Real Business In One Day [Full Client Transformation]

# Thumbnails

2026-10-10The Automations Agencies Overcharge Forno editorNeeds voicenote

script (2.1k chars)

# Research

Driver: a price you are already paying, cancelled on camera. Every one of these names a real cost and then removes it.

"Run AI Agents Locally? Here's Why Companies Pay $15K for This"

Jinni Doo, 151,239 views, 197x outlier, Published Feb 27, 2026

youtube.com/watch?v=_EkVRn_Hb7Y

"I Got Claude Code Working Completely Free (Here's how)"

Amidia AI, 194,369 views, 80x outlier, Published Jun 17, 2026

youtube.com/watch?v=P3BlPnxmCYQ

"FREE Open-Source WhatsApp CRM (Step-by-Step Setup)"

Lakshit Ukani, 167,054 views, 40x outlier, Published May 16, 2026

youtube.com/watch?v=BbUqyF-cosA

"FREE UNLIMITED Claude Code (No NVIDIA NIM, No Ollama!) - Fastest Responses [2026]"

Danish Sofi, 153,401 views, 19x outlier, Published May 10, 2026

youtube.com/watch?v=TazjcZrTl7Y

"Finally, An AI Assistant That Doesn't Leak Your Data"

Proton, 27,151 views, 6.7x outlier, Published Jul 1, 2026

youtube.com/watch?v=5RIQw_Fdxzg

Why it transfers: four practitioners on small channels beating baseline 19x to 197x, so this carries with no audience behind it. On the price: the $15,000 came from the Jinni Doo title and you were right that copying one number is lazy. Luke should use the real spread Boom quotes, which is what the concept now asks for.

# Concept

Luke goes through 7-10 automations agencies charge thousands for, and builds each one live.

For each automation he covers:

•What an agency quotes for it

•The build, on screen, start to finish

•What it actually costs to run per month

Ends on the ones worth paying for anyway, and why.

Funnel: Bottom.

# Titles

1.I Built The Automations Agencies Charge $15,000 For

1.Why Companies Pay $15,000 For This One Automation

1.Your Client Data Is Training Someone Else's AI

# Thumbnails