← AIAP Topic CoursesAll programsHomeSearch
AIAP Topic Courses·Generating Organic Sales·37:35

Organic Sales Playbook, Part 1: Build the System - Offer, Revenue Back-Math, Buyer Research, and the First Content System

Outskill instructor (unnamed) Organic-sales trainer - helps organisations become AI-native; demos every prompt live in Claude ("any LLM would work")

The short version

  1. Organic is a system: goal, process (research, create, distribute, convert), feedback loop. Test every piece: does it move someone one step closer to trusting me? Mistakes: quitting early (pushing through is 60% of the game), vanity metrics, copying influencers, being everywhere. Live AI audit of your last five LinkedIn posts (0:00-0:12).
  2. Define the offer first - what, for whom, which problem, what price - or content is a hobby blog. Back-math: $5,000/month = 2 clients at $2,500 = 10 conversations = ~100 engaged people. Optimise for the right 200, not the wrong 10,000. Live offer-check prompt with a LinkedIn profile summary flagged a price-credibility mismatch (0:10-0:24).
  3. Buyer research in their words: five profile questions; sources - Google autocomplete, People Also Ask, Reddit, your DMs; rank by intent ('how much does an AI consultant cost' vs 'what is AI'). Live prompt yields a confidence-tagged profile and 20 autocomplete queries sorted by intent. Content system: one platform, one idea in six formats, five formats that sell, 2-3 posts a week for six months, 80% value 20% sell (0:22-0:37).

At a glance, three clicks deep

Skim here first: the closed row is the glance, open is the study card with the key points and timestamps, and the ↓ link drops to that concept's full write-up below.

01Organic as a system, and revenue back-math from the offerGoal -> clients -> conversations -> engaged audience.›

Goal -> clients -> conversations -> engaged audience.

System and mistakes (0:00-0:05)

Offer and math (0:10-0:17)

Offer-check prompt (0:17-0:24)

↓ Full write-up of this concept

02Buyer research: five questions, four sources, intent rankingTheir words;›

Their words; four free sources; intent first.

Questions and sources (0:22-0:27)

Prompt (0:28-0:31)

↓ Full write-up of this concept

03The first content system: one platform, six formats, a cadenceOne lane;›

One lane; repurpose six ways; 80/20.

Platform and formats (0:31-0:35)

Cadence (0:36-0:37)

↓ Full write-up of this concept

The concepts in full

01

Organic as a system, and revenue back-math from the offer

Five thousand a month is two clients. Two clients is ten conversations. Ten conversations is a hundred people who actually engage. Now you know what to post for.

A system has a goal, a process (research, create, distribute, convert) and a feedback loop; random posting has none. Test: can this move someone one step closer to buying/trusting? Five mistakes: quit too early, chase vanity metrics instead of replies and leads, copy 100k-follower accounts, try every platform, post without a buyer. Offer first: what am I selling, for whom, what problem, what price ('AI automation consulting for Shopify store owners'). Back-math: $5,000 = 2 x $2,500; 1 in 5 conversations closes -> 10 conversations; 1 in 10 connections converts -> ~100 engaged people. 300 views with two client DMs beats 10,000 views from non-buyers. Live prompt: state offer, buyer, problem, price, goal; attach a LinkedIn profile summary (pulled by a browser-controlling assistant); ask for gaps - Claude role-played a sceptical buyer, a competitor and a mentor.

Why it matters

Turns 'post more' into a number you can plan against.

This idea elsewhereparallelCatalyst Session 30
02

Buyer research: five questions, four sources, intent ranking

Autocomplete and People Also Ask give you twenty or thirty real questions in five minutes. Rank them by how close the asker is to paying.

Use the buyer's words, not jargon. Five questions: job title, industry, company size; the problem they are solving now; where they get information; what they already tried that failed; what they would happily pay for. Sources: Google autocomplete, People Also Ask, Reddit, your own DMs and comments. Rank by intent: 'how much does it cost to hire an AI consultant for customer-service automation' (near buying) vs 'what is artificial intelligence' (curiosity). Exercise: autocomplete sweep, five Reddit threads, five questions you get asked, ranked. Live prompt: 'act as a buyer research analyst; build the five-question profile with best guesses tagged high/medium/low confidence; list 20 likely autocomplete searches sorted into intent vs curiosity.'

Why it matters

Content that sells answers intent questions; this is how you find them.

03

The first content system: one platform, six formats, a cadence

Three solid posts a week for six months beats ten mediocre posts a week for two.

One platform: LinkedIn for B2B and professional services, X for the tech AI crowd, YouTube is evergreen but 'a tough nut' - start with LinkedIn or X. One idea, six touch points: blog post (the home base you own), LinkedIn/X post, short screen recording, thread or step-by-step, newsletter, Reddit/Quora reply. Revenue content teaches something specific ('how I used Claude to cut a client's support tickets 40% in 30 days'), ties news to the buyer's situation, invites a response ('DM me for the template'). Five beginner formats: what-I-did walkthrough, before/after, how-to, mistakes I made, tool breakdown. Cadence 2-3 posts a week; 80% value, 20% sell.

Why it matters

A content plan that fits a solo operator's week.

Tools referenced

ToolCoverageMomentContext
ClaudedemonstratedPost audit, offer check, buyer research
LinkedInexplained
RedditexplainedResearch source

Action items

    Resources mentioned

    Resources
    • docLesson page on the platform

    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
    13 minutes30 (or 15) minutes for the research exercise
    cloud / Cloud co work / cloud researchClaude / Claude Cowork / Claude Research
    chat g b t / ChargeGPT / Chargebee / chat GBTChatGPT
    ATN / n a 10 / Anytimen8n

    True on recording day — verify before relying