← All sessionsHomeSearch
C7 EST | 14 Day AI Sprint·Day 14 | Graduation·4:08:00

Day 14: Graduation — Skills, Systems, Dollars: the Niche Framework, the First-Dollar Roadmap, Pricing, Moats, and What Happens Next

Harshad Tyagi Monetization mentor - his own path freelancer -> agency -> consultant -> SaaS; runs the two-hour playbook · Uthappa Host - graduation formalities, Catalyst re-enrolment offer, CSAT reveal, hackathon briefing and team rooms
▶LMS lesson

The short version

  1. 'Skills systems lead you to dollars.' The four skill families from the sprint - smart prompts, RAG/document chatbots, automations, customer-support AI - are not businesses; packaged into a system with a working demo and sold into a niche you know from your day job (your 'unfair advantage'), they are.
  2. Find the niche with five questions - who do I help, so that they can achieve what, what is the most unique problem, the most PAINFUL (= willingness to pay), the most LUCRATIVE (= buildable ROI) - and narrow with one or two of seven axes (industry, demographic, geography, platform, price, delivery, 'x without y'). Output: 'I help [who] achieve [what] through [how] without [pain].' 'The riches are in the niches.'
  3. First dollar in a week: days 1-3 pick the niche, days 4-7 build, deliver, collect. Give it free to 2-3 people first. Outreach script: 'I noticed [problem]. I just built a tool that [result]. I'm testing it with 3 [persona] this week for [price]. If it doesn't work as promised, I'll refund immediately.'
  4. Price by revealed value: ask what it is worth, charge a fraction - a lawyer at $200/h saving 5 hours is $1,000 of value, so charge ~$100; SaaS is per-user x users x 12. Then clone the first win to ten buyers at $10-20 for testimonials, standardize onboarding, aim for $500/month, then subscription/agency/community.
  5. Ladder: freelancer -> small agency -> SaaS/service hybrid -> product; consultant only if you already have deep domain expertise. Moats: proprietary data, community, public reputation - 'best product launched in silence' loses to 'first person plus good reputation.' Don't overbuild: 'spend 3-4 days nailing a system, then showcase that MVP.' 10-point opportunity filter, 0-2 each; 14+ means go.
  6. Housekeeping: cohort CSAT 4.76 (sessions) / 4.62 (breakouts), 'highest rated cohort'; a one-time reopening of 10 interview-gated Catalyst seats; weekly Wednesday AI updates for a year, lifetime LMS access, 5% referral; 48-hour hackathon from tonight.

The concepts

01

Skills -> systems -> dollars: what the sprint's four skills become

Nobody buys a prompt. They buy the morning they no longer spend doing the thing the prompt does.

Four monetizable skill families: smart prompts (package 5-10 for one role, give them free to 2-3 people first - Gumroad for the payment page), RAG/document chatbots (build one for one business), automations (automate one weekly task - agencies at scale earn '$10k-$100k a month'), customer-support AI (deploy one bot, charge $1,000-$3,000 setup plus retainer). Each needs proof - a working demo - and a niche you already understand: 'problems you know from your job.' Vertical examples ran across healthcare, legal, real estate, education and ecommerce.

The case study: a client spending $80k/day across 50+ Meta ad accounts paid two VAs $3,000/month each to compile reports by hand; Harshad automated a single cross-account report to Slack/WhatsApp - which became the seed of his own marketing-reporting product.

Why it matters

It is the exact bridge from the KB's how-to records to Paul's agency offers, spoken by the mentor who later teaches Catalyst.

This idea elsewhereextendsCatalyst Session 1 opens with the same skills-vs-business distinction at more depth
Go deeper

In one line: Package one skill into one system for one niche, prove it with a demo, price it on saved hours or reduced no-shows; domain knowledge is the unfair advantage.

'Skills systems lead you to dollars' (l3186087 0:17)

Four families: prompts, RAG chatbots, automations, support AI (l3186087 0:19-0:20)

Domain expertise from your job is the unfair advantage (l3186087 0:21)

First moves: free to 2-3 peers; one chatbot for one business; one weekly task; one support bot at $1-3k setup (l3186087 0:22-0:50)

Case: $6k/month of VA reporting on 50+ ad accounts automated into one report (l3186087 0:40-0:41)

▶ Watch this taught:

02

The niche framework: five questions, seven axes, one sentence

'I help everyone' is a sentence that has never closed a sale.

Five questions: (1) Who do I want to help? (2) So that they can achieve what? (3) What is the most unique - unpopular - problem I can solve? (4) What is the most painful problem (pain = willingness to pay)? (5) What is the most lucrative problem I can actually build? Then narrow with one or two of seven axes: industry, demographic, geography, digital platform, price point, delivery method ('x but delivered differently'), or the bad-problem axis ('x but without y'). Worked example: busy mothers, get fit, without leaving home, no childcare, cheaper than gym plus babysitter, online, no equipment. Output sentence: 'I help [who] achieve [what] through [how] without [pain].' Mistakes: too broad, fear of excluding future customers, skipping the outreach test out of fear of rejection. Timeline: today Q1, this week all five, this month test with real prospects, this quarter refine.

Why it matters

The same exercise Paul's Catalyst Session 2 runs as the 'niche pyramid'; this is the plain version to hand a prospect.

This idea elsewhereparallelCatalyst Session 2's niche pyramid is the same framework drawn differently
Go deeper

In one line: Who / so-that / unique / painful / lucrative, narrowed by 1-2 of industry, demographic, geography, platform, price, delivery, 'without'; expressed as one I-help sentence.

Five questions ending in painful (pay) and lucrative (buildable) (l3186087 1:08-1:15)

Seven niching axes (l3186087 1:16-1:21)

Template: 'I help [who] achieve [what] through [how] without [pain]' (l3186087 1:22)

Mistakes: too broad, afraid to exclude, never testing (l3186087 1:24)

Outreach script with refund promise (l3186087 1:03)

Cadence: today / this week / this month / this quarter (l3186087 1:25)

▶ Watch this taught:

Check yourself

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

You have 'I help small clinics save staff time with AI.' Which axis is missing to make it sellable?

The 'without' - the pain removed (no-shows, phone tag, after-hours calls) - and a delivery/price axis. 'I help 3-10-staff clinics cut no-shows 30% with automated reminders, without new front-desk hires' passes the test.

03

First dollar in a week, pricing by revealed value, and the growth ladder

A lawyer saves five hours at two hundred an hour. You built that. Charge a hundred - and then sell it ten more times.

Roadmap: days 1-3 find the niche; days 4-7 build, deliver, collect payment. Pricing for custom work: 'get your leads to reveal the value' - ask what it is worth and charge a fraction (value $1,000 -> $100-200); or hours saved x hourly rate / 10 (the lawyer example). Software: price per user per month x users x 12. After the first sale, clone it to ten buyers in the same niche at $10-20 (or free) for testimonials, standardize into repeatable onboarding, target $500/month, then scale by subscription, niche agency or community. The worked clinic package (3-10 staff): prompts (patient summary, insurance letter, post-visit instructions), a 24/7 knowledge bot, automations (booking, 24h reminders, 3-day follow-up, no-show rescheduling), a support bot handling ~70% of inquiries; 1-2 weeks to set up, priced on 10-15 staff hours a week saved and ~30% fewer no-shows.

Ladder: freelancer -> small agency with standard packages -> SaaS/service hybrid ($50-500k range, mostly automated) -> product; skip 'consultant' unless you already have deep domain credibility. Moats: proprietary data (access big players cannot get), community, public reputation - document the build on X/LinkedIn. Anti-overbuild: 'spend 3-4 days nailing a system, then showcase that MVP'; let customers, not ChatGPT, write the roadmap. The 10-point opportunity filter scores 0-2 each; 14+ means pursue now.

Why it matters

This is the pricing logic behind every quote Paul writes - and the explicit permission to charge a tenth of the value rather than an hourly rate.

This idea elsewhereparallelCatalyst Session 1 turns 'clone the win' into a pricing cycleparallelSession 2's pricing models expand the two given here
Go deeper

In one line: Niche (d1-3) -> build/deliver/collect (d4-7); price = fraction of revealed value or hours x rate / 10; clone to 10, standardize, $500/mo; freelancer -> agency -> hybrid -> product; moats = data, community, reputation.

Days 1-3 niche, days 4-7 build and collect (l3186087 1:26-1:28)

Reveal-the-value pricing; hours x rate / 10 (lawyer: $1,000 -> $100) (l3186087 1:51-1:54)

SaaS: per-user x users x 12 (l3186087 1:50)

Clone to 10 at $10-20 for testimonials; standardize; $500/month (l3186087 1:56-1:57)

Clinic package: 1-2 weeks; 10-15 h/week saved; ~30% fewer no-shows (l3186087 1:53-1:59)

Ladder and moats; 'spend 3-4 days nailing a system, then showcase' (l3186087 1:29-1:45)

10-point filter, 0-2 each, 14+ = go (l3186087 1:40-1:43)

▶ Watch this taught:

04

What the cohort built, what they got, and what came next

Fourteen days, twelve working artefacts - read the list once and you have the table of contents of this whole course.

Recap of breakout builds: D1 RAG on the Tesla manual in Dify; D2 an AI commercial; D3 the n8n website-lead spam checker; D4 the expense-tracker agent; D5 a first MCP server in Claude Desktop; D6 a Retell voice agent booking an orthopaedic clinic via Google Calendar; D8 data storytelling with IDEA; D9 the Apify Google-Maps scraper with cold-email drafts; D10 agent blueprinting; D11 the multi-agent shopping assistant in Lyzr; D12 a Gmail MCP assistant and a custom MCP server; D13 'Jerry' (email, calendar, Fireflies) plus the front-end-connected expense tracker. Then: certificates by Jan 30; a one-time reopening of 10 interview-gated seats in the 6-month AI Catalyst program (phase 1 foundations 4 months, phase 2 scale; 70+ sessions, 24 office hours; Jan 30 kickoff) closing within the session; weekly Wednesday AI-update sessions for a year, content library incl. Kairos business school, lifetime LMS access, 5% referral; CSAT 4.75/4.77 sessions, 4.57/4.66 breakouts - 'highest rated cohort'; hackathon kicks off tonight, submission Jan 27 23:59 IST, demo day Wednesday.

Why it matters

Paul is in that Catalyst program - this is the moment it was pitched to the cohort he is now reading.

This idea elsewhereextendsThe weekly AI-update sessions promised here begin with the February 2026 digest
Go deeper

In one line: Twelve breakout artefacts across 14 days; Catalyst re-enrolment offer; a year of weekly updates; hackathon Jan 25-27, demo Jan 28, certificates Jan 30.

Day-by-day artefact list D1-D13 (l3186087 2:03-2:08)

Catalyst: 6 months, 2 phases, 70+ sessions, 24 office hours; 10 seats, interview-gated (l3186087 2:09-2:22)

Perks: weekly AI updates for a year, library, lifetime LMS, 5% referral (l3186087 2:22-2:51)

CSAT 4.76 sessions / 4.62 breakouts (l3186087 2:38-2:41)

Hackathon: 48 h, submit Jan 27 23:59 IST, demo Jan 28, certificates Jan 30 (l3186087 2:51-2:53)

▶ Watch this taught:

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.

01Skills -> systems -> dollars: what the sprint's four skills becomePackage one skill into one system for one niche, prove it with a demo, price it on saved hours or reduced n…›

Package one skill into one system for one niche, prove it with a demo, price it on saved hours or reduced no-shows; domain knowledge is the unfair advantage.

'Skills systems lead you to dollars' (l3186087 0:17)

Four families: prompts, RAG chatbots, automations, support AI (l3186087 0:19-0:20)

Domain expertise from your job is the unfair advantage (l3186087 0:21)

First moves: free to 2-3 peers; one chatbot for one business; one weekly task; one support bot at $1-3k setup (l3186087 0:22-0:50)

Case: $6k/month of VA reporting on 50+ ad accounts automated into one report (l3186087 0:40-0:41)

02The niche framework: five questions, seven axes, one sentenceWho / so-that / unique / painful / lucrative, narrowed by 1-2 of industry, demographic, geography, platform…›

Who / so-that / unique / painful / lucrative, narrowed by 1-2 of industry, demographic, geography, platform, price, delivery, 'without'; expressed as one I-help sentence.

Five questions ending in painful (pay) and lucrative (buildable) (l3186087 1:08-1:15)

Seven niching axes (l3186087 1:16-1:21)

Template: 'I help [who] achieve [what] through [how] without [pain]' (l3186087 1:22)

Mistakes: too broad, afraid to exclude, never testing (l3186087 1:24)

Outreach script with refund promise (l3186087 1:03)

Cadence: today / this week / this month / this quarter (l3186087 1:25)

03First dollar in a week, pricing by revealed value, and the growth ladderNiche (d1-3) -> build/deliver/collect (d4-7);›

Niche (d1-3) -> build/deliver/collect (d4-7); price = fraction of revealed value or hours x rate / 10; clone to 10, standardize, $500/mo; freelancer -> agency -> hybrid -> product; moats = data, community, reputation.

Days 1-3 niche, days 4-7 build and collect (l3186087 1:26-1:28)

Reveal-the-value pricing; hours x rate / 10 (lawyer: $1,000 -> $100) (l3186087 1:51-1:54)

SaaS: per-user x users x 12 (l3186087 1:50)

Clone to 10 at $10-20 for testimonials; standardize; $500/month (l3186087 1:56-1:57)

Clinic package: 1-2 weeks; 10-15 h/week saved; ~30% fewer no-shows (l3186087 1:53-1:59)

Ladder and moats; 'spend 3-4 days nailing a system, then showcase' (l3186087 1:29-1:45)

10-point filter, 0-2 each, 14+ = go (l3186087 1:40-1:43)

04What the cohort built, what they got, and what came nextTwelve breakout artefacts across 14 days;›

Twelve breakout artefacts across 14 days; Catalyst re-enrolment offer; a year of weekly updates; hackathon Jan 25-27, demo Jan 28, certificates Jan 30.

Day-by-day artefact list D1-D13 (l3186087 2:03-2:08)

Catalyst: 6 months, 2 phases, 70+ sessions, 24 office hours; 10 seats, interview-gated (l3186087 2:09-2:22)

Perks: weekly AI updates for a year, library, lifetime LMS, 5% referral (l3186087 2:22-2:51)

CSAT 4.76 sessions / 4.62 breakouts (l3186087 2:38-2:41)

Hackathon: 48 h, submit Jan 27 23:59 IST, demo Jan 28, certificates Jan 30 (l3186087 2:51-2:53)

Tools referenced

ToolCoverageMomentContext
n8nmentionedAutomations as a product line; cloud ~EUR 20-25/mo for 2.5k executions
Make.commentionedAlternative automation platform
GumroadmentionedPayment page for prompt packs
VoiceflowmentionedSupport/voice stack example
TwiliomentionedTelephony in the support stack
TypeformmentionedIntake in the support stack
Harvey AImentionedEnterprise legal AI comparator
HiggsfieldmentionedAd-creative generation example
LovablementionedFront ends for productized systems
SupabasementionedDatabase behind SaaS examples

Session materials

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

md
text-lessons.md
0 KB
md
text-lessons.md
0 KB
md
text-lessons.md
0 KB
md
text-lessons.md
0 KB
md
text-lessons.md
0 KB
md
text-lessons.md
0 KB
md
text-lessons.md
0 KB
md
text-lessons.md
0 KB
md
text-lessons.md
0 KB
md
text-lessons.md
0 KB
md
text-lessons.md
0 KB
md
text-lessons.md
0 KB
lms
Resources
companion page on the platform
lms
Resources
companion page on the platform
lms
Resources
companion page on the platform
lms
Resources
companion page on the platform
lms
Resources
companion page on the platform
lms
Resources
companion page on the platform
lms
Resources
companion page on the platform
lms
Resources
companion page on the platform
lms
Resources
companion page on the platform
lms
Resources
companion page on the platform
lms
Resources
companion page on the platform
lms
Resources
companion page on the platform
lms
Resources
companion page on the platform

Action items

    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
    any 10 / anytime / an itemn8n
    AshishHarshad (host greeting the same speaker)
    Bev / WiproVaibhav (program lead)
    DiffyDify
    LizerLyzr
    Kyra's business schoolKairos business fellowship content
    thirteenth of Janthirtieth of January (certificates)

    True on recording day — verify before relying