← All sessionsHomeSearch
AI Catalyst C3·Core Sessions - Week 1·3:10:34

Session 2: How to Plan, Scope and Sell your AI Projects

Harshit Trainer — Catalyst program director (the Basecamp 2-4 trainer); runs an AI agency (name garbled, heard as 'agent device') and the product Glued (AI creative/UGC automation); company registered in UAE; 11 years working with US clients; self-described sales-hater who partners with client-holders instead · Niharika Cohort manager — session intro and close, CSAT poll

Session map

LAUNCHPADFIND & VALIDATEDELIVER & PRICELinkedIn storefrontphoto · banner · one-line offerAEOget cited by AI answersNiche pyramidfunction + industry + problemPain-signal miningReddit · skool · YouTube commentsServices → productthe Glued pathRun more experimentsthe Stanford lessonDiscovery & scopingobjectives · requirements · KPIsDiagram-first designthe flowchart IS the proposalThe atomic pilotfirst win in two weeksQA & error handlingfail loudly to you, never to themPricing is valuereveal the human ROIThree pricing modelsfixed · retainer · performance
LaunchpadFind & validateDeliver & price
click a node — its card pops up (drag it anywhere, × to close)
Concept

The map reads left to right — launchpad flow into find & validate, then into deliver & price. Click any node to open that idea here; every timestamp jumps into the recording.

The short version

  1. The agency startup kit session: everything between 'I have a skill' and 'I have a paying client' — storefront, niche clarity, pain-signal research, outreach, project scoping, diagram-first system design, QA and error handling, and pricing — with templates (proposal, contracts, discovery script, pitch-deck guidelines) shipped via the LMS.
  2. LinkedIn is the storefront: clear photo (AI-generated professional shot is fine), banner carrying the one-line offer ('I help [niche] [verb] [problem] with AI'), case studies as experience links — and the AEO insight: ChatGPT heavily cites LinkedIn posts and Reddit, so consistent niche posting gets you into AI answers.
  3. Niche = business function + industry + the exact problem, mined from where the niche complains: Reddit (title/upvotes/comments as signals), skool.com communities over 1,000 members, YouTube comment sections (outsourced domain expertise), WhatsApp groups — with AI summarizing the noise into painful problems, in the niche's own language.
  4. Delivery discipline: discovery call (objectives, requirements, KPIs — clients state narrow problems; curiosity uncovers the real one) → diagram-first design ('diagram first, then bullet points, no paragraphs') → atomic pilot promised in a month and delivered in two weeks → three QA layers plus error alerts → deployment, which opens the retainer conversation.
  5. Pricing is value, value is business impact: get the client to reveal the human ROI (the $9k/month Vietnamese legal-review team his automation replaced at 5-10x margins), price from AI skills × domain experience × business impact, and choose among fixed price (+training upsell), subscription/retainer, and performance-based (10-25% of the delta, fixed-fee floor).

The concepts

01

LinkedIn as your storefront

0:14:22

Your storefront isn't a website you'll build someday — it's the LinkedIn profile buyers check first, and it's three fixes away from working.

When a stranger evaluates you, they don't trust your website — anyone can type anything there. They trust LinkedIn, where your background, past companies and connections are verifiable. So the storefront work is profile mechanics: a clear professional photo (a selfie run through Gemini or ChatGPT is now legitimate), a banner that states your one-line offer with your domain name placed where the profile picture won't cover it on phones — check the mobile render — and the bio line that does all the selling: 'I help [niche] [automate/build/grow] [problem] with AI.' One line; a stranger should know in seconds what you can do for them.

The under-used surface is the experience section: every entry takes links and media, which is where case studies belong — 'conducted this workshop for a 9-figure D2C brand', linked. The trainer audited his own profile live and admitted the gap: his case studies live buried on his website where nobody digs. Signals of having done the thing before are the biggest credibility currency, and they belong one click from your name.

Worked example · from the session

The live profile audit: his banner, his founder entry linking to the agency site, the media slots — plus the honest 'this is missing from mine, I'm fixing it' on case-study links.

Why it matters

Every outreach message and every piece of content you ever post drives people to this page. Its state decides whether the click becomes a call.

People get this wrong

I need a proper website before I can start selling.

The buyer checks LinkedIn first and trusts it more. Optimize the profile today; the website can follow the first clients.

For your projects

The KB's session pages double as case-study links for exactly this play — 'built a 71-page AI-generated course knowledge base' with a URL is a stronger banner claim than most agencies have.

Go deeper

In one line: Before websites, LinkedIn: a clear profile picture (selfie → Gemini/ChatGPT professional shot), a banner stating the one-line offer plus your domain name (check how it renders on phones), the bio line 'I help [niche] [verb] [problem] with AI', and case studies attached as links/media on every experience entry.

LinkedIn beats a bare website on credibility: background, work history and mutual connections are verifiable; a website says whatever you typed (0:14:22)

The experience section takes links and media — his own profile demoed live, including the gap he admitted (case studies buried on his website instead of surfaced here) (0:20:31)

Same principles port to X, Instagram, YouTube: face visible, offer in the banner, bio says who you help and how (0:26:38)

▶ Watch this taught: 0:14:22

Check yourself

Answer from memory first — the recall attempt is what makes it stick. Then reveal.

Why does LinkedIn out-credential a website?

It's verifiable — background, companies, connections. A website is self-asserted; LinkedIn is socially confirmed.

What are the three storefront fixes?

Clear professional photo (AI-generated is fine), banner with the one-line offer + domain (mobile-checked), and the 'I help X do Y with AI' bio line — plus case-study links on experience entries.

02

AEO: getting cited by AI answers

0:24:35

For a decade the game was ranking on Google; now the game is being the source an AI quotes — and the raw material is LinkedIn posts and Reddit threads.

When people ask ChatGPT or Perplexity for 'best AI agencies for marketing automation', the models don't consult a ranking — they source links whose content answers the query, weighted by credibility and backlinks. The trainer's discovery: ChatGPT cites a LOT of LinkedIn posts. So the content strategy writes itself — post consistently about your niche in your niche's terms ('AI for real estate agents', story after story), and when a few posts get traction, you start appearing inside AI answers rather than under them. He learned this when a client reached out because his agency surfaced in an AI overview.

Reddit is the second citation source — the models constantly surface its threads for sentiment ('how do people feel about X'). Harder to work deliberately, but it means the communities where you're already mining pain (next concept) are also where citations grow. SEO took a decade of tactics; AEO is young enough that consistent niche posting is still the whole playbook.

Worked example · from the session

The 'AI for real estate agents' repetition strategy: same niche phrase, different stories, week after week — until ChatGPT's training and retrieval associate you with the query.

Why it matters

Buyers increasingly ask AI instead of Google. Being the cited answer is the new page-one — and right now it's cheap, because most competitors aren't posting for it.

For your projects
  • TechOnCall AEO play: consistent LinkedIn posts on 'AI for Connecticut small business IT' — the local-niche phrasing is exactly what gets cited when an SMB owner asks ChatGPT who can help.
Go deeper

In one line: Answer engine optimization — the successor to SEO. ChatGPT/Perplexity/Gemini cite sources by credibility and query-fit rather than rank, and they heavily cite LinkedIn posts and Reddit threads. Consistent niche posting ('AI for real estate agents', repeatedly) raises your odds of appearing inside AI answers.

His own proof: a client found his agency because it 'came up in AI overview' when they searched for AI marketing-automation agencies (0:22:32)

Reddit is the second AI-citation goldmine — harder to game, but sentiment threads get surfaced constantly (0:26:38)

▶ Watch this taught: 0:24:35

Check yourself

Answer from memory first — the recall attempt is what makes it stick. Then reveal.

How do AI engines choose what to cite, versus Google's old model?

Not rank — query-fit plus credibility. If your post/blog actually answers the user's question, it can get cited regardless of domain authority.

What two platforms did the trainer flag as heavily cited?

LinkedIn posts and Reddit threads — which makes consistent niche posting on LinkedIn the highest-leverage AEO move.

03

The niche pyramid: function + industry + problem

0:28:39

'Pick a niche' is useless advice until it's three stacked questions: which function, which industry, which exact problem.

The pyramid builds from the base: first the business function you resonate with — sales, marketing, accounting, finance, HR — the kind of problem you have a knack for. That alone is too broad. Add the industry — healthcare, coaches, real estate, e-commerce — and now 'sales' becomes 'sales for real estate agents': a niche with edges. The peak is the exact problem, which you don't invent — you find it where the niche complains (next concept), stated in their own words.

The payoff is operational, not aesthetic: function+industry immediately tells you where to research ('where do real estate agents hang out?' is a question Perplexity answers), and the completed pyramid produces the storefront's one-line offer. Session 1 taught WHY narrow beats broad (3-5x pricing); this is the HOW of constructing the narrowness.

Worked example · from the session

The session's running instantiation: sales (function) + real estate (industry) + 'lead follow-up is soul-crushing' (problem, found verbatim in a community) → 'I help real estate agents automate lead follow-up with AI.'

Why it matters

Every downstream decision — where to research, what to post, what to build, what to charge — falls out of these three choices. Ambiguity here is why most people stall.

People get this wrong

Niching down means picking an industry.

Industry alone is still too broad. The niche is function × industry, and the sellable offer needs the third layer — a specific validated problem.

1 · The exact problem found in the niche's own words 2 · Industry healthcare · coaches · real estate · e-commerce 3 · Business function sales · marketing · accounting · finance · HR Function + industry = your niche. Add the exact problem and you have the offer line: "I help [niche] [verb] [problem] with AI."
Function + industry = niche; add the exact problem and you have the offer line
Go deeper

In one line: Clarity framework: level 3 — business function (sales, marketing, accounting, HR); level 2 — industry (healthcare, coaches, real estate, e-commerce); level 1 — the exact problem. Function + industry = your niche; adding the specific problem produces the offer line and tells you which channels to research.

Domain experts can skip ahead; career-changers use the pyramid to rebuild credibility in a new space (0:28:39)

Once function+industry is set, research targets pick themselves: 'where are real estate agents hanging out?' is a Perplexity question (0:32:42)

▶ Watch this taught: 0:28:39

Check yourself

Answer from memory first — the recall attempt is what makes it stick. Then reveal.

Name the three pyramid levels in build order.

Business function (base) → industry (middle) → exact problem (peak). Function+industry is the niche; the problem completes the offer.

What does a completed pyramid immediately give you?

The one-line offer for your storefront, and the research targets — you now know exactly which communities and channels to mine.

04

Mining pain signals from communities

how-to0:34:44

AI is a technology looking for use cases — the use cases are sitting in comment sections, stated in the exact words your offer should use.

The biggest question — 'what should I solve?' — has a mechanical answer: go where your niche already complains. Reddit subreddits, triaged by title, upvotes and comment count. skool.com and Facebook communities (pick ones with 1,000+ members; he joined a wholesaling-real-estate community live and surfaced 'I spent 4 hours on the dialer and felt my soul leaving my body' within minutes). YouTube channels serving the niche — where the creator's manual how-to videos are effectively automation specs, and the comments show which parts hurt. Even WhatsApp groups and hyperlocal apps for hyperlocal offers.

Two multipliers make it scale. AI does the reading: a browser extension summarizing a 295-comment thread into 'the painful problems being discussed' turns an afternoon into minutes. And the harvest isn't just the problem — it's the language. When many people resonate with the same soul-crushing complaint, you have the validation signal AND the exact words for your offer line. If your offer doesn't sound like their comments, it won't land.

Worked example · from the session

The live skool.com session: community joined on camera, dialer-pain thread found, comment section mined — demonstrating the whole loop from 'I need a niche problem' to a validated, quotable pain in under ten minutes.

Do it in this order

GotchasCommunities are full of self-promoters and bots — the engagement signals (upvotes, comment volume, authentic replies) are your filter. And the YouTube move is the underrated one: a creator explaining a manual 5-step workflow has handed you an automation spec plus an audience reacting to it.

Why it matters

This replaces the fatal pattern of building from imagination. Empathy at scale — reading the niche's own complaints — is what separates offers people recognize from offers people scroll past.

People get this wrong

Finding a niche problem requires industry experience.

It requires research where the industry complains. Communities plus AI summarization let you borrow domain awareness — though lived expertise still deepens it.

Reddit 3 signals: title · upvotes · comments only high-engagement threads Niche communities skool.com · Facebook groups pick ones with 1,000+ members YouTube comments niche channels = outsourced domain expertise WhatsApp & local groups, gated communities, hyperlocal apps AI summarizes the noise browser extension: "summarize the comments, find the painful problems" Validated pain, in their language "soul-crushing" is the phrase your offer reuses Many people agreeing in a comment section = the signal you've been looking for.
Four places the niche complains — AI summarizes the noise into validated pain
For your projects

Your equivalent goldmines: CBIA groups, Connecticut SMB Facebook groups, MSP subreddits. One evening of dialer-thread-style mining would hand you the exact language for a TechOnCall AI offer.

Go deeper

In one line: Find the exact problem where the niche complains: Reddit (filter by title, upvotes, comments), skool.com communities over 1,000 members, Facebook groups, YouTube comment sections, WhatsApp groups and hyperlocal apps — with a browser AI summarizing comment sections into 'painful problems people are discussing'.

Live demo: joined a 'wholesaling real estate' skool community and read a thread — 'I spent 4 hours today on the dialer and felt my soul leaving my body' — the exact language your offer should reuse (0:37:02)

295 comments? Don't read them — Claude's browser extension summarizes and extracts the pains (0:41:09)

YouTube niche channels = outsourced domain expertise: the creator knows real estate, you know AI — automate what their videos describe manually (0:47:17)

Comment-section agreement is the validation signal: many people resonating = the problem is real and shared (0:49:20)

▶ Watch this taught: 0:34:44

Check yourself

Answer from memory first — the recall attempt is what makes it stick. Then reveal.

What three signals triage a Reddit thread before you click?

Title (pain-shaped?), upvote count, comment count — only high-engagement threads are worth mining.

What's the YouTube 'outsourced domain expertise' move?

Follow creators who serve the niche daily: their manual how-to content is your automation spec, and their comment sections show which steps the audience finds painful.

Beyond the problem itself, what must you capture?

The language — the niche's own phrases ('soul-crushing'). Offers written in the customer's words are the ones they recognize as theirs.

05

Services first, product later (the Glued path)

0:57:30

Don't guess which product to build — run a services business until one solution sells three times, then you already know.

The services-first logic is evidence-based product discovery. Serving clients puts you inside real workflows, where one solution eventually clicks — for the trainer it was UGC-ad automation for D2C brands, spotted via a YouTube comment section, sold as a service to three clients, then productized as Glued. The clients who bought the service became the product's first users and its proof; one creative agency he eventually stopped charging now onboards her own clients onto it — evangelism worth more than her subscription. Selling the same solution five times is the signal to productize into recurring revenue.

The strategic asymmetry: services pivot at the speed of the technology — n8n workflows yesterday, Claude skills today, same business — while products carry re-engineering overhead every time the ground shifts, and standalone AI products are 'getting crushed left, right and center' by the labs' own releases. Even the investors agree now: Y Combinator's requests-for-startups explicitly seeks AI-native services companies — don't just build AI for lawyers, provide the lawyers too, and charge 10x. Their startup directory doubles as free market research on what's fundable.

Worked example · from the session

Glued's full arc told first-person: comment-section insight → service for 3 clients → product → client team of 10 onboarded → free evangelist client feeding the pipeline.

Why it matters

This resolves the build-a-product anxiety most learners carry: the product isn't a decision you make up front, it's a pattern that emerges from paid service work — with revenue the whole way.

People get this wrong

A product is the serious business; services are what you do until you can afford to build one.

Services are the discovery engine AND a durable model — they pivot freely, fund themselves, and are now what top investors explicitly back in the AI era.

For your projects

TechOnCall is already the services layer; the Glued lesson is to watch which AI solution you deliver three times — that's your productizable asset, with your client base as first users.

Go deeper

In one line: Start with services: solve problems for real businesses, watch for the one solution that clicks, then productize it. His product Glued grew from client services (UGC-ad automation for D2C brands); services pivot instantly as models change, while products carry rebuild overhead — and YC now explicitly invests in 'AI-native services' companies.

His UGC-ads automation: found via a YouTube comment section, sold to 3 clients as a service, then turned into the product — with a client's 10-person team as the first at-scale users (0:43:11, 1:01:42)

The free-client evangelist: a creative agency he stopped charging brings him clients continuously because Glued is embedded in her workflow — 'the best kind of marketing' (0:57:30)

Services pivot with the technology ('n8n workflows yesterday, Claude skills today'); products require re-engineering the back end (1:03:44)

YC signal: their requests-for-startups now lists AI-native services/agencies — 'take care of the law as well... and charge 10x'; the startup directory is a research tool for what's fundable (1:05:47)

▶ Watch this taught: 0:57:30

Check yourself

Answer from memory first — the recall attempt is what makes it stick. Then reveal.

What's the trigger to productize a service?

Repetition: when you've sold the same solution ~5 times, it's validated and templated — turn it into a subscription product.

Why do services survive model churn better than products?

A service swaps tools the day something better ships; a product must re-architect. Services sell outcomes, which are tech-agnostic.

What is an 'AI-native services' company per YC's thesis?

Product + the service humans around it — AI for lawyers plus the lawyers — charging multiples of what a bare tool commands.

06

Discovery calls: scope the real problem

1:11:53

The client will tell you a narrow problem; the paid work is discovering the business around it — with the success metric agreed before anything gets built.

The discovery call's engine is curiosity about the business, not the feature request. Clients arrive with a narrowed symptom — 'we need this generation automated' — because they don't know what's buildable. Ask how the business operates: what's the manual workflow, who touches it, what does each tool produce, what are the high-level goals. Often the real opportunity is one level up — the founder fixating on one automation actually needs the whole content operating system, and proposing that ('forget this branch — what if I automate the entire thing?') is how scope and price both grow legitimately.

The part everyone forgets: metrics. Agree on the KPIs that define success before building — revenue moved for sales systems, leads/reach for marketing — because an undefined finish line guarantees a dissatisfied client. Do the metric research beforehand, and shape your booking intake (Calendly questions) so calls start half-scoped. Session 1's constraint diagnosis runs inside this same conversation.

Worked example · from the session

The lead-qualification client (next concept) who asked for 'lead qualification' — and, after operational questions, ended up with a scored-lead research pipeline plus an optional voice agent, because the discovery uncovered what the sales reps actually did all day.

Why it matters

Scoping is where projects are won or doomed: the right questions surface bigger engagements, and agreed KPIs are the difference between 'delivered' and 'disputed'.

People get this wrong

A good discovery call pitches your capabilities.

It's an interview of their business. Curiosity uncovers the real problem; the pitch comes later, in a diagram.

Go deeper

In one line: Project scoping happens on the discovery call: genuine curiosity about how the business operates (clients state narrow problems; the real one is usually adjacent), clear objectives and requirements, and agreed success metrics/KPIs — plus the questions asked at booking time (Calendly intake).

Clients present symptoms: the founder obsessing over one marketing automation may actually need a content operating system — you only see it by learning the business (1:13:54)

Define measurable success up front: sales = revenue moved; marketing = leads/reach — 'how will you evaluate whether the system we delivered is successful?' (1:15:57)

A discovery-call script ships in the startup kit (1:11:53)

▶ Watch this taught: 1:11:53

Check yourself

Answer from memory first — the recall attempt is what makes it stick. Then reveal.

Why not build what the client asks for on the call?

They present symptoms, narrowed by what they think is possible. Learning the business reveals the adjacent, often larger, real problem.

What three things must leave the discovery call with you?

Objectives, concrete requirements (the manual workflow, tools, data), and agreed success metrics/KPIs.

07

System design: diagram first, no paragraphs

how-to1:16:57

The proposal that closes isn't written — it's drawn: one flowchart that tells the client the whole story of what you'll build.

His design discipline: after discovery, reverse-engineer the requirements into a flowchart — when does the system trigger, what does each step do, what data moves where. The live example: a client asked for 'lead qualification'; the diagram that sold the project showed a LinkedIn lead-form trigger, Apify scraping the lead's profile, company page and website, an LLM extracting data points into their GoHighLevel CRM, a scoring model ranking leads 1-10 (work the 8-10s; dip lower only when volume is thin), and an optional voice agent bolted on top — a bigger, clearer system than the request, visible at a glance.

The diagram works three jobs at once: the client understands exactly what they're buying; you gain a build plan that prevents mid-development chaos; and if you're business-side, it's a precise spec to hand a technical partner. Hence the writing rule: diagram first, then bullet points, and no paragraphs — nobody reads them. Tools don't matter much (FigJam, Canva, Excalidraw, Miro — or ask Claude for an HTML flowchart and screenshot it); the discipline does. The proposal then assembles itself: kit template plus meeting notes into Claude.

Worked example · from the session

The actual client diagram walked node by node on screen — trigger, scrapes, extraction, CRM write, scoring, voice-agent extension — 'this is the image that went in the proposal.'

Do it in this order

GotchasThe diagram is not decoration — skipping it means re-explaining everything mid-development when things go wrong. And write like he does: 2-3 lines max between visuals; 'if I can understand everything from this one diagram, I don't need your paragraphs.'

Why it matters

Every hour here saves days of mid-build confusion and re-explanation. And clients buy what they can see: the diagram converts where prose gets skimmed.

People get this wrong

A thorough proposal is a long written document.

Thorough means legible: one diagram carrying the whole story beats pages of prose nobody reads. The template handles the legal boilerplate.

My principle is diagram first, then bullet points, and no paragraphs. If I can understand everything from this one diagram, I don't need your paragraphs.1:28:13
For your projects

Your KB figures follow this exact doctrine already — the bc/s0x SVGs are 'diagram first' applied to teaching. For client work, the same Excalidraw/Claude-HTML move turns any TechOnCall proposal into a one-picture pitch.

Go deeper

In one line: Design the system as a flowchart the client can read: trigger → each processing step → outputs, drawn in FigJam/Canva/Excalidraw/Miro (or generated by Claude as an HTML page). The diagram carries the proposal — 'diagram first, then bullet points, and no paragraphs.'

Worked example shown: lead-qualification engine — LinkedIn lead-form trigger → Apify scrapes lead + company profiles + websites → LLM extracts data points → GoHighLevel CRM contact → LLM scores leads (work 8-10s first) → optional voice agent (1:18:00)

The diagram serves both sides: client clarity AND your build plan — and lets you hand a precise spec to a technical partner if you're business-side (1:22:06)

Proposal template in the kit: feed it + meeting notes to Claude and the proposal generates itself (1:28:13)

▶ Watch this taught: 1:16:57

Check yourself

Answer from memory first — the recall attempt is what makes it stick. Then reveal.

State the writing hierarchy for proposals.

Diagram first, then bullet points, and no paragraphs — 2-3 lines of prose maximum anywhere.

What three jobs does the system diagram do?

Sells the client (they see the story), plans your build (prevents mid-development chaos), and specs the work for any technical partner.

In the lead-qual diagram, what happened to the client's original request?

'Lead qualification' became a research + scoring pipeline with CRM integration and an optional voice agent — discovery plus design legitimately grew the scope.

08

The atomic pilot: first win in two weeks

1:24:07

'Transform my business' is a three-month promise you can't prove — so sell the two-week slice that proves everything.

Founders arrive with transformation-sized asks. Committing to them cold is how engagements die: trust hasn't been earned, scope is foggy, and the first stumble sinks it. The move is decomposition: find the atomic milestone — the smallest self-contained deliverable you're confident lands in two weeks — and propose it as a paid pilot. Quote a month; deliver in two weeks. Underpromise, overdeliver, on purpose.

The pilot is mutual qualification dressed as a project: the client risks little, you learn whether they're workable, and the early win installs the belief that 'this person actually delivers' — which is what converts the 3-month contract into the 6-month partnership and later the retainer. He opens with it explicitly: 'I want to work on a pilot first.' Confidence, not hedging.

Worked example · from the session

The framing script demonstrated: pilot scoped from the milestone breakdown, priced, quoted at a month — with the two-week delivery pre-planned as the overdelivery.

Why it matters

First-client relationships are trust machines, and trust compounds from small kept promises faster than from big made ones.

People get this wrong

Big engagements are won by proposing big scopes.

They're won by proving small scopes fast. The pilot IS the sales strategy for everything after it.

1 · Scope discovery call: objectives, requirements, KPIs 2 · Design diagram-first — the flowchart IS the proposal 3 · Develop build & test step by step, pilot milestone first 4 · QA & errors 3 QA layers · alerts when things break, not silence 5 · Deploy then iterate, add, improve "How do I maintain this?" → the retainer deployment opens the door to maintenance, updates, support — recurring revenue Underpromise, overdeliver: quote the pilot at a month, land it in two weeks — the first win turns contracts into partnerships.
Scope → design → develop → QA → deploy — and deployment opens the retainer
Go deeper

In one line: Break the client's grand goal into an atomic milestone deliverable in ~2 weeks; propose it as a paid pilot quoted at a month, deliver early (underpromise, overdeliver). The first win converts a 3-month contract into a 6-month partnership and opens the retainer.

His default opener with new clients: 'I want to work on a pilot first — let's see if we can work together' (1:26:10)

The first win establishes the only trust that matters: 'this person actually delivers' (1:26:10)

▶ Watch this taught: 1:24:07

Check yourself

Answer from memory first — the recall attempt is what makes it stick. Then reveal.

What makes a milestone 'atomic'?

Self-contained, valuable on its own, and deliverable in ~2 weeks with confidence — the smallest provable win.

Why quote a month for two weeks of work?

Deliberate underpromise/overdeliver: early delivery is manufactured delight, and it buys the trust that upgrades contracts into partnerships.

09

QA layers and error handling

how-to1:30:18

In the AI era the scarce skill isn't building — it's being the person whose systems fail loudly to the builder and never silently to the client.

AI has made humans negligent: outputs get piped along unread until 'Here is the email for you' appears in a client's inbox. The antidote is layered: an AI QA stage inside the system, a human QA pass, and the founder's own approval — three gates before 'done'. Testing is artifact-driven: ask the client what they'll actually throw at the system, collect 10 real examples, let AI pattern-generate 20 more, and run 10-15 through before delivery.

Error handling is the professional signature. If a system breaks and the client tells you, you've failed twice — once technically, once reputationally. Everything you deploy should alert YOU on failure, and long-running jobs need explicit monitoring: his n8n workflow silently timed out on 500-page uploads while everyone just waited. Security now belongs in the same bucket — compromised mainstream libraries (the Axios incident, the GitHub hack) mean an automated daily audit comparing your stack against vulnerability news is a buildable, sellable habit. When customer counts climb from 10 toward 50, this whole function becomes a hire.

Worked example · from the session

The 500-page timeout, reconstructed: complex workflow, silent stall, client waiting 45 minutes — the exact scenario alert-on-failure design exists to prevent.

Do it in this order

GotchasQA is boring, which is exactly why it's skipped — and why 'Here is the email for you' ships inside production emails. The reputational asymmetry is brutal: one silent failure a client discovers outweighs weeks of correct runs.

Why it matters

Delivery quality is remembered as the last failure, not the average run. Loud-failure engineering is cheap, and it's most of the difference between hobbyist and professional automation.

People get this wrong

Testing the happy path before handoff is QA.

QA is adversarial: real client artifacts, silent-failure hunting, timeout monitoring, and security audits — layered, because each layer catches what the previous one missed.

For your projects

Your build-verify habit (rebuild after every session, grep the rendered page) is layer one; the missing piece for client-grade work is the alert-on-failure pattern — worth adopting for any n8n/automation you ever host for TechOnCall clients.

Go deeper

In one line: Three QA layers before 'we're done' (AI-layer QA, human QA, founder approval); test 10-15 artifacts per build (ask the client for 10 example queries, generate 20 more with AI); and error handling that alerts YOU when systems break — the client discovering the outage is the failure mode.

The AI-era QA disgrace: 'Here is the email for you' pasted into the actual email — humans getting negligent about output review (1:30:18)

War story: an n8n workflow silently timing out on a client's 500-page uploads — no error surfaced, people just waited; long-running tasks need explicit monitoring (1:34:22)

Security is now part of QA: compromised libraries (the Axios incident), daily automated audits of your stack, alert-on-news systems (1:36:25, 2:13:02)

Scale rule: going from 10 to 50 customers, hire an engineer/QA engineer; the general hire trigger — when revenue clears their salary (1:36:25, 2:35:22)

▶ Watch this taught: 1:30:18

Check yourself

Answer from memory first — the recall attempt is what makes it stick. Then reveal.

Name the three QA layers.

AI-layer QA inside the system → human QA pass → founder/your final approval. Three gates before telling the client it's done.

How do you build the test set for a client system?

Ask the client for ~10 real example inputs, generate ~20 more with AI in the same pattern, and run 10-15 artifacts through before delivery.

What's the cardinal rule of error handling?

Fail loudly to the builder: automatic alerts to you on any breakage. The client discovering an outage is the defining failure.

10

Pricing is value: reveal the human ROI

1:40:29

Never price from your effort — price from the number the client just told you the manual version costs them.

Pricing confusion is a credibility leak: if you don't know your price, the client concludes you don't know your work. The anchor discipline: pricing = value = business impact — and the cleanest source of impact is the client's own books. His masterclass case: a firm doing contract due-diligence on Fiserv's billion-dollar payment-processing deals paid a Vietnamese legal team $9-15k/month for manual document review. Once that number is revealed, automating the workflow prices at 5-10x margins with total legitimacy — the value conversation is already over. When no manual system exists, research the counterfactual: what would the 2-4 staff needed cost at market salaries?

Your negotiating range comes from three multiplied variables — AI skills (the craft), domain experience (years they can't fake), business impact (the revealed number) — each one a lever in the conversation. Frame impact positively: not 'halve your accounting staff' but 'take per-employee revenue from 5k to 15k' — raise the ceiling, don't lower the floor. And keep deals win-win: 2-3x the manual cost closes; 10x insults; and starting cheap isn't fatal — his first $700 project became a $10k/month retainer because the value showed up.

Worked example · from the session

The Fiserv chain in one line: billion-dollar contracts → mandatory due-diligence → $9k/month manual team → revealed ROI → an automation priced at multiples, justified by their own numbers.

Why it matters

This flips pricing from an anxiety into an arithmetic: one good discovery question ('what does this cost you today?') replaces all the agonizing — and the three-variable frame tells you exactly which lever to strengthen when a client pushes back.

People get this wrong

Price reflects the time and effort the build takes you.

Effort is invisible and irrelevant to the buyer. Price reflects business impact — ideally a number the client themselves revealed.

Pricing power = three variables multiplied AI skills what you're here building × Domain experience years in the industry nobody can fake × Business impact money made or saved — revealed ROI Fixed price one-time deliverable + team training as the upsell simplest — start here Subscription / retainer maintenance · support · consulting hours recurring — the goal of every first win on AI products: tokens crush margins Performance-based "50k → 150k MRR; we take 10-25% of the delta" hybrid: fixed-fee floor + reduced % high trust — they must open the books Get the client to reveal the human ROI ("we pay $9k/month for that manually") — then price the automation against it.
Three multiplied variables set the power; three models collect the money
Pricing is value. Value is business impact, and business impact is going to command the ROI. Get them to reveal how much they are paying for the manual workflows.1:48:34
For your projects

The revealed-ROI question ports straight into MSP sales: 'what does the manual version cost you?' asked about ticket triage, onboarding docs, or report assembly prices AI add-ons off THEIR number, not your hours.

Go deeper

In one line: Clear pricing signals confidence; fumbled pricing signals amateur. The guide: pricing = value = business impact. Get the client to reveal what the manual process costs them (the $9k/month offshore legal-review team his automation replaced — chargeable at 5-10x), or research comparable staffing costs when no system exists. Pricing power = AI skills × domain experience × business impact.

The Fiserv story: contract due-diligence for billion-dollar payment-processing deals, manually done by a $9-15k/month Vietnamese legal team — an automation of it prices from that revealed number, not from build effort (1:44:31)

Raise the ceiling, not lower the floor: sell 'per-employee revenue 5k → 15k', not 'cut your accounting staff in half' (1:52:37)

Win-win discipline: 2-3x the manual cost clears; 10x ('$90k for what cost $9k') kills the deal; his first project was $700 — and became a $10k/month retainer (1:54:38)

Countermove awareness: 'competitors charge half' — answered by your domain expertise and market knowledge, not discounting reflexes (1:52:37)

▶ Watch this taught: 1:40:29

Check yourself

Answer from memory first — the recall attempt is what makes it stick. Then reveal.

What's the single best pricing question?

Get them to reveal the human ROI: 'what does the manual version of this cost you today?' — then price the automation against that number.

State the pricing-power formula.

AI skills × domain experience × business impact — three multiplied variables, and three separate levers in negotiation.

Why 'raise the ceiling' instead of 'cut costs' in the pitch?

'Cut your staff in half' threatens; 'lift per-employee revenue from 5k to 15k' inspires — same economics, opposite reception.

11

Three pricing models: fixed, retainer, performance

1:54:38

Three ways money arrives: once (fixed), monthly (retainer), or as a share of growth you caused (performance) — and the craft is knowing which fits the client and how to protect your downside.

Fixed price fits defined deliverables — 'build this pipeline, hand it over' — and carries a natural upsell: training their team, billed hourly, which he sells constantly. Subscription/retainer is the destination: maintenance, support, consulting hours, the recurring revenue every first win should be steered toward (the trend-chasing content client who needs a new automation weekly is the archetype). One caution for productized AI: tokens crush subscription margins — heavy users cost you real money per interaction, so price from unit economics (cost per interaction plus margin) or sell credits.

Performance-based is the high-risk, high-trust play: 'you're at 50k MRR; work with us 3 months, we take you to 150k and keep 10-25% of the delta.' It requires the client to open their books — do it only with people you trust — and the professional structure is the hybrid: a fixed monthly floor that covers your costs, plus a slightly reduced percentage (20-22% instead of 25) on the upside. Session 1's guarantee discipline applies verbatim: his recruiting-era scar of delivering appointments no one closed is why the floor exists.

Worked example · from the session

The creative-agency deal narrated: 50k MRR baseline, 150k target in 3 months, percentage on the delta — with the books-open trust requirement learned first-hand.

Why it matters

Model choice shapes the whole relationship: fixed caps your upside, retainer compounds it, performance multiplies it with risk. Choosing deliberately — and flooring the downside — is what separates pricing strategy from hoping.

People get this wrong

Performance pricing means pure upside share — maximum boldness, maximum reward.

Pure upside repeats the appointments-nobody-closed mistake. The professional shape floors the downside with a fixed fee and takes percentage on measurable, logged results.

Go deeper

In one line: Fixed price (one-time deliverable — with team training as the natural upsell), subscription/retainer (maintenance, support, consulting hours — the recurring goal), and performance-based (a share of the delta: '50k → 150k MRR, we take 10-25% of the increase' — high-trust, protect the downside with a fixed-fee floor).

AI products' subscription margins get crushed by token costs — usage scales, your COGS scales with it; unit economics and credit systems are the countermeasures (1:56:39, 2:31:19)

Performance deals demand transparency: the client must open the books — 'do this only with people you trust' (2:00:43)

The hybrid: fixed fee covers costs, reduced percentage (20-22% instead of 25%) rides the upside (2:23:12)

▶ Watch this taught: 1:54:38

Check yourself

Answer from memory first — the recall attempt is what makes it stick. Then reveal.

What's the natural upsell on a fixed-price build?

Team training — delivered per hour, it converts a handover into an engagement and raises the fixed total.

How do you structure a performance deal safely?

Hybrid: a fixed-fee floor covering costs plus a reduced percentage of the measured delta — and only with clients trusted enough to open their books.

Why do tokens threaten AI product subscriptions?

Usage scales your API costs, so flat subscriptions can invert into losses — price by unit economics or credits instead.

12

Run more experiments (the Stanford lesson)

2:02:44

Stanford's $200k curriculum compressed into one sentence: the people succeeding are just running more experiments — and AI made experiments nearly free.

His GSB mentor's report: the books are the same everywhere; the differentiators are network and one operational truth — winners simply run more experiments. The taught exercise: before building anything, stand up a one-page landing site for the imagined offer, run Google Ads, capture emails. Students who raced five pages found two with real demand and built one. That's validation as a numbers game, not a debate — and AI's deepest business gift is parallelism: five landing pages, five offers, five niches tested simultaneously, then total concentration on whichever clicks.

The Q&A wrapped operating wisdom around it. Inbound beats outbound: when clients find you (content, AEO), you hold the pricing power; when you reach out, they do — so brand building is the structural answer to discount pressure. Niche selection includes buying habits: his worst mistake was building developer tools, because developers build rather than buy — solve rich problems for people who pay. And the 1-minute demo rule disciplines all outreach: if the Loom can't convince in 60 seconds, the 5-minute version won't save it.

Worked example · from the session

His own first client: a posted demo (the EA-copilot Twitter thread, January 2025-era) — inbound from showing work publicly, not outreach — closing the loop between experiments, content and pricing power.

Why it matters

This is the session's operating system: every framework taught works better run five times in parallel than agonized over once. Experiment velocity is the meta-skill.

People get this wrong

Success comes from choosing the right idea and executing it hard.

It comes from testing many ideas cheaply and concentrating only after evidence — the experiment count, not the idea quality, is the differentiator.

People who are succeeding are just running more experiments. That's it.2:04:45
For your projects

The five-landing-pages move costs an afternoon with your stack: five one-page offers for TechOnCall AI services, small ad spend, and the CBIA network as organic distribution — evidence before any build.

Go deeper

In one line: The GSB takeaway: 'people who are succeeding are just running more experiments.' Validate before building — landing page + Google Ads capturing emails for an unbuilt product; 5 pages racing, 2 win, 1 gets built. AI's real gift is parallel experimentation; once one clicks, concentrate everything on it.

Inbound beats outbound on pricing power: when they come to you (via content/AEO), you hold the upper hand; when you reach out, they do — brand building is the systemic answer to discount pressure (2:21:10)

Niche riches: 'developing anything for developers is the worst decision I made' — they build rather than buy; pick niches with money and buying habits (2:21:10)

The 1-minute demo rule: convince in 60 seconds or not at all — the 5-minute version goes only to those who bite; his first client came from a posted demo thread (2:29:17)

Q&A operations: charge for prototypes ('I don't work for free — and neither should you'), take 10-20% upfront, LinkedIn scraping is against ToS ('use Apify, not your own account'), hire when revenue clears salary (2:17:05-2:37:23)

▶ Watch this taught: 2:02:44

Check yourself

Answer from memory first — the recall attempt is what makes it stick. Then reveal.

What was the Stanford validation exercise?

Landing page for an unbuilt offer + Google Ads + email capture — run several in parallel, read demand from the numbers, build only the winner.

Why does inbound command higher prices than outbound?

Whoever initiates cedes leverage: inbound clients arrive pre-convinced by your content; outbound prospects know you need them.

What made developer tools his 'worst decision'?

The niche doesn't buy — developers exhaust free/self-built options first. Niche selection must include willingness to pay.

Every concept, three clicks deep

The same concepts as a quick reference: the closed row is the glance, open is the study card, and every timestamp jumps into the recording.

01LinkedIn as your storefrontBefore websites, LinkedIn: a clear profile picture (selfie → Gemini/ChatGPT professional shot), a banner st…0:14:22

Before websites, LinkedIn: a clear profile picture (selfie → Gemini/ChatGPT professional shot), a banner stating the one-line offer plus your domain name (check how it renders on phones), the bio line 'I help [niche] [verb] [problem] with AI', and case studies attached as links/media on every experience entry.

LinkedIn beats a bare website on credibility: background, work history and mutual connections are verifiable; a website says whatever you typed (0:14:22)

The experience section takes links and media — his own profile demoed live, including the gap he admitted (case studies buried on his website instead of surfaced here) (0:20:31)

Same principles port to X, Instagram, YouTube: face visible, offer in the banner, bio says who you help and how (0:26:38)

02AEO: getting cited by AI answersAnswer engine optimization — the successor to SEO.0:24:35

Answer engine optimization — the successor to SEO. ChatGPT/Perplexity/Gemini cite sources by credibility and query-fit rather than rank, and they heavily cite LinkedIn posts and Reddit threads. Consistent niche posting ('AI for real estate agents', repeatedly) raises your odds of appearing inside AI answers.

His own proof: a client found his agency because it 'came up in AI overview' when they searched for AI marketing-automation agencies (0:22:32)

Reddit is the second AI-citation goldmine — harder to game, but sentiment threads get surfaced constantly (0:26:38)

03The niche pyramid: function + industry + problemClarity framework: level 3 — business function (sales, marketing, accounting, HR);0:28:39

Clarity framework: level 3 — business function (sales, marketing, accounting, HR); level 2 — industry (healthcare, coaches, real estate, e-commerce); level 1 — the exact problem. Function + industry = your niche; adding the specific problem produces the offer line and tells you which channels to research.

Domain experts can skip ahead; career-changers use the pyramid to rebuild credibility in a new space (0:28:39)

Once function+industry is set, research targets pick themselves: 'where are real estate agents hanging out?' is a Perplexity question (0:32:42)

04Mining pain signals from communitiesFind the exact problem where the niche complains: Reddit (filter by title, upvotes, comments), skool.com co…0:34:44

Find the exact problem where the niche complains: Reddit (filter by title, upvotes, comments), skool.com communities over 1,000 members, Facebook groups, YouTube comment sections, WhatsApp groups and hyperlocal apps — with a browser AI summarizing comment sections into 'painful problems people are discussing'.

Live demo: joined a 'wholesaling real estate' skool community and read a thread — 'I spent 4 hours today on the dialer and felt my soul leaving my body' — the exact language your offer should reuse (0:37:02)

295 comments? Don't read them — Claude's browser extension summarizes and extracts the pains (0:41:09)

YouTube niche channels = outsourced domain expertise: the creator knows real estate, you know AI — automate what their videos describe manually (0:47:17)

Comment-section agreement is the validation signal: many people resonating = the problem is real and shared (0:49:20)

05Services first, product later (the Glued path)Start with services: solve problems for real businesses, watch for the one solution that clicks, then produ…0:57:30

Start with services: solve problems for real businesses, watch for the one solution that clicks, then productize it. His product Glued grew from client services (UGC-ad automation for D2C brands); services pivot instantly as models change, while products carry rebuild overhead — and YC now explicitly invests in 'AI-native services' companies.

His UGC-ads automation: found via a YouTube comment section, sold to 3 clients as a service, then turned into the product — with a client's 10-person team as the first at-scale users (0:43:11, 1:01:42)

The free-client evangelist: a creative agency he stopped charging brings him clients continuously because Glued is embedded in her workflow — 'the best kind of marketing' (0:57:30)

Services pivot with the technology ('n8n workflows yesterday, Claude skills today'); products require re-engineering the back end (1:03:44)

YC signal: their requests-for-startups now lists AI-native services/agencies — 'take care of the law as well... and charge 10x'; the startup directory is a research tool for what's fundable (1:05:47)

06Discovery calls: scope the real problemProject scoping happens on the discovery call: genuine curiosity about how the business operates (clients s…1:11:53

Project scoping happens on the discovery call: genuine curiosity about how the business operates (clients state narrow problems; the real one is usually adjacent), clear objectives and requirements, and agreed success metrics/KPIs — plus the questions asked at booking time (Calendly intake).

Clients present symptoms: the founder obsessing over one marketing automation may actually need a content operating system — you only see it by learning the business (1:13:54)

Define measurable success up front: sales = revenue moved; marketing = leads/reach — 'how will you evaluate whether the system we delivered is successful?' (1:15:57)

A discovery-call script ships in the startup kit (1:11:53)

07System design: diagram first, no paragraphsDesign the system as a flowchart the client can read: trigger → each processing step → outputs, drawn in Fi…1:16:57

Design the system as a flowchart the client can read: trigger → each processing step → outputs, drawn in FigJam/Canva/Excalidraw/Miro (or generated by Claude as an HTML page). The diagram carries the proposal — 'diagram first, then bullet points, and no paragraphs.'

Worked example shown: lead-qualification engine — LinkedIn lead-form trigger → Apify scrapes lead + company profiles + websites → LLM extracts data points → GoHighLevel CRM contact → LLM scores leads (work 8-10s first) → optional voice agent (1:18:00)

The diagram serves both sides: client clarity AND your build plan — and lets you hand a precise spec to a technical partner if you're business-side (1:22:06)

Proposal template in the kit: feed it + meeting notes to Claude and the proposal generates itself (1:28:13)

08The atomic pilot: first win in two weeksBreak the client's grand goal into an atomic milestone deliverable in ~2 weeks;1:24:07

Break the client's grand goal into an atomic milestone deliverable in ~2 weeks; propose it as a paid pilot quoted at a month, deliver early (underpromise, overdeliver). The first win converts a 3-month contract into a 6-month partnership and opens the retainer.

His default opener with new clients: 'I want to work on a pilot first — let's see if we can work together' (1:26:10)

The first win establishes the only trust that matters: 'this person actually delivers' (1:26:10)

09QA layers and error handlingThree QA layers before 'we're done' (AI-layer QA, human QA, founder approval);1:30:18

Three QA layers before 'we're done' (AI-layer QA, human QA, founder approval); test 10-15 artifacts per build (ask the client for 10 example queries, generate 20 more with AI); and error handling that alerts YOU when systems break — the client discovering the outage is the failure mode.

The AI-era QA disgrace: 'Here is the email for you' pasted into the actual email — humans getting negligent about output review (1:30:18)

War story: an n8n workflow silently timing out on a client's 500-page uploads — no error surfaced, people just waited; long-running tasks need explicit monitoring (1:34:22)

Security is now part of QA: compromised libraries (the Axios incident), daily automated audits of your stack, alert-on-news systems (1:36:25, 2:13:02)

Scale rule: going from 10 to 50 customers, hire an engineer/QA engineer; the general hire trigger — when revenue clears their salary (1:36:25, 2:35:22)

10Pricing is value: reveal the human ROIClear pricing signals confidence;1:40:29

Clear pricing signals confidence; fumbled pricing signals amateur. The guide: pricing = value = business impact. Get the client to reveal what the manual process costs them (the $9k/month offshore legal-review team his automation replaced — chargeable at 5-10x), or research comparable staffing costs when no system exists. Pricing power = AI skills × domain experience × business impact.

The Fiserv story: contract due-diligence for billion-dollar payment-processing deals, manually done by a $9-15k/month Vietnamese legal team — an automation of it prices from that revealed number, not from build effort (1:44:31)

Raise the ceiling, not lower the floor: sell 'per-employee revenue 5k → 15k', not 'cut your accounting staff in half' (1:52:37)

Win-win discipline: 2-3x the manual cost clears; 10x ('$90k for what cost $9k') kills the deal; his first project was $700 — and became a $10k/month retainer (1:54:38)

Countermove awareness: 'competitors charge half' — answered by your domain expertise and market knowledge, not discounting reflexes (1:52:37)

11Three pricing models: fixed, retainer, performanceFixed price (one-time deliverable — with team training as the natural upsell), subscription/retainer (maint…1:54:38

Fixed price (one-time deliverable — with team training as the natural upsell), subscription/retainer (maintenance, support, consulting hours — the recurring goal), and performance-based (a share of the delta: '50k → 150k MRR, we take 10-25% of the increase' — high-trust, protect the downside with a fixed-fee floor).

AI products' subscription margins get crushed by token costs — usage scales, your COGS scales with it; unit economics and credit systems are the countermeasures (1:56:39, 2:31:19)

Performance deals demand transparency: the client must open the books — 'do this only with people you trust' (2:00:43)

The hybrid: fixed fee covers costs, reduced percentage (20-22% instead of 25%) rides the upside (2:23:12)

12Run more experiments (the Stanford lesson)The GSB takeaway: 'people who are succeeding are just running more experiments.' Validate before building —…2:02:44

The GSB takeaway: 'people who are succeeding are just running more experiments.' Validate before building — landing page + Google Ads capturing emails for an unbuilt product; 5 pages racing, 2 win, 1 gets built. AI's real gift is parallel experimentation; once one clicks, concentrate everything on it.

Inbound beats outbound on pricing power: when they come to you (via content/AEO), you hold the upper hand; when you reach out, they do — brand building is the systemic answer to discount pressure (2:21:10)

Niche riches: 'developing anything for developers is the worst decision I made' — they build rather than buy; pick niches with money and buying habits (2:21:10)

The 1-minute demo rule: convince in 60 seconds or not at all — the 5-minute version goes only to those who bite; his first client came from a posted demo thread (2:29:17)

Q&A operations: charge for prototypes ('I don't work for free — and neither should you'), take 10-20% upfront, LinkedIn scraping is against ToS ('use Apify, not your own account'), hire when revenue clears salary (2:17:05-2:37:23)

Tools referenced

ToolCoverageMomentContext
LinkedIndemonstrated0:14:22Storefront anatomy demoed on his own profile: banner, bio line, experience links/media, case-study slots — plus the AEO insight that ChatGPT cites LinkedIn posts heavily
skool.comdemonstrated0:34:44Community platform joined live (wholesaling real estate, 1,000+ members) to mine pain threads — the 'soul leaving my body' dialer post
Redditdemonstrated0:39:06r/RealEstate triaged live by title/upvotes/comments; also flagged as a top AI-citation source
Claude (browser extension)demonstrated0:41:09Summarized a 295-comment thread into 'painful problems people are discussing' — the research multiplier
Y Combinator (requests for startups + startup directory)demonstrated1:05:47Toured live as market-signal research: AI-native services thesis, batch directory filtered by category (B2B marketing etc.); apply anytime, deadlines soft
FigJamdemonstrated1:16:57His team's system-design tool — the lead-qualification engine diagram shown was built in it; Canva/Excalidraw/Miro named as peers
Apifyexplained1:20:04The scraping layer in the lead-qual design (LinkedIn lead + company profiles, websites); also the ToS-safe answer to 'is scraping LinkedIn legal?' — 'not legal; use Apify or another account'
GoHighLevelexplained1:20:04The client's CRM in the worked design — contacts created, data points and lead scores written back
n8nexplained1:34:22Build platform for client automations: largest integration catalog (its edge over Claude skills), the 500-page timeout war story, self-hosting for cost, weekly Google-Sheets re-auth reality
Claude Code / Codexexplained2:19:07His heavy daily pair (via cmux terminal); Claude for proposals/websites/diagrams; Codex recommended as the cheaper rate-limit relief; 'Claude Code is going to be the best thing ever' for most learners
Gluedexplained0:57:30His product — UGC-ad/creative automation for D2C brands, grown from services; the free-client evangelist story
Gemini / ChatGPTmentioned0:16:25Selfie → professional headshot; banner generation; ChatGPT as the AEO citation engine
Perplexitymentioned0:32:42'Where are real estate agents hanging out?' — niche channel research
Vanta / CodeRabbitmentioned2:37:23Compliance certification (SOC 2, HIPAA, GDPR) and code-review/security tooling for client-grade builds
Loommentioned2:31:19The 1-minute demo vehicle: 60 seconds to convince, 5-minute version only for those who bite
Google Adsmentioned2:04:45The Stanford validation engine: ads onto unbuilt-offer landing pages, demand read from email capture
cmuxmentioned2:19:07His terminal for running multiple Claude Code/Codex instances ('not for everyone — I'm an engineer'); session on it promised
Windsurfmentioned2:43:27Learner tool-confusion question (heard as 'BenSurf/WIMServe'); answer: don't tool-hop — 'Claude Code is going to be the best thing ever' for most

Session materials

Archived locally on V: — click to open. Companion pages link to the LMS.

Action items

Resources mentioned

Resources
  • docAgency startup kit (Google Drive via LMS): proposal template, legal contracts (short + long), discovery-call script, Calendly intake questions, pitch-deck guidelines, website structuring guide, NDAs 2:00:43
  • docSession whiteboard + the FigJam lead-qualification diagram (images shared) 1:22:06
  • docY Combinator requests-for-startups and startup directory (market-signal research) 1:05:47
  • docHis UGC-ads automation video (the services→product origin artifact) 0:43:11
  • docPromised: GTM session (post-build), premium-website session, cmux walkthrough, error-handling sessions, margin/niche article (office hours) 0:59:33

Extraction notes

This page was built from an auto-generated transcript, which garbles product and people's names. Those were corrected silently in everything above and logged here for transparency. The warnings flag claims that were true on the recording day but change fast.

Transcript corrections applied

The transcript saysThe trainer actually means
Harsha / HarshidHarshit (trainer)
white codingvibe coding
agent devicetrainer's AI agency name (garbled, unverified — consistent with BC2)
Blued / Glued / Gloop / Bella HybridGlued (his product; 'Bella Hybrid productive workforce' likely a banner tagline garble)
any time / anything / Anytime / editing workflows / Internet workflows / n a 10n8n
Wappy / AP 5 / scrapeifyApify
school dot comskool.com
Charge GBT / ChargeGPD / Chargebee / chat g b t / chat g p dChatGPT
icky guy momentikigai moment
BiroMiro
c mux / CMOX / c maxcmux (terminal)
BenSurf / WIMServe / VimServeWindsurf
OpenFlow / open blockOpenClaw
spectrum and developmentspec-driven development
AMZAMC (annual maintenance contract)
GSBStanford Graduate School of Business
Fiserv / FISERVFiserv (payment processing company, as stated)
3.5 Flash (video model)Gemini video model version as stated (point-in-time)
hundred x b c handlerunresolved garble in a learner exchange
Mythos was first released to all of these banks... posing a threat to cybersectrainer's characterization of Anthropic's Mythos rollout (as stated — see freshnessNotes)

True on recording day — verify before relying