← Content Library (LinkedIn Growth Mastery)All programsHomeSearch
Content Library (LinkedIn Growth Mastery)·LinkedIn Growth Mastery - AI Content Creation Mastery·50:27

Four Ways to Never Run Out of Ideas: Expertise Mining, the 5-Year Reflection Interview, Audience-Question Mining in Perplexity, and Competitor Analysis

Uttam Gupta Trainer - former director of growth at Growth School, founder of Agent Valley (agentvalley.ai)

The short version

  1. Batch creation: block two hours, produce ~30 posts. The five-step workflow: ideas -> hooks -> posts -> schedule -> weekly review (0:00-0:03).
  2. Method 1: a 'positioning canvas' doc feeds a 50-idea prompt (20 how-to, 15 mistakes, 10 frameworks, 5 controversial), run in ChatGPT and again in Claude Opus with the doc uploaded. Method 2: seven reflective pillars (mistakes, challenges, insecurities, achievements, lessons, skills, routines) mined by an AI interview answered by voice, 10 ideas per answer (0:04-0:35).
  3. Method 3: Perplexity finds the top 50 questions your audience asks across LinkedIn, Reddit, Quora and Google, then 30 ideas per list. Method 4: 5-10 competitors x 5-10 best posts, improved with your own view - 'Steal Like an Artist' (0:35-0:50).

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.

01Batch creation, the five-step workflow, and the positioning canvasWeekly batch;›

Weekly batch; canvas doc as standing context.

Five steps (0:03)

Canvas build (0:04-0:08)

↓ Full write-up of this concept

02Method 1: the 50-idea expertise-mining prompt50 ideas in four buckets;›

50 ideas in four buckets; top-10 pass.

Prompt (0:08-0:12)

Claude Opus repeat (0:13-0:16)

↓ Full write-up of this concept

03Method 2: the 5-year reflection interview (seven pillars, by voice)Seven pillars x interview x 10 ideas each.›

Seven pillars x interview x 10 ideas each.

Seven pillars (0:16-0:27)

Voice interview demo (0:28-0:35)

↓ Full write-up of this concept

04Methods 3 and 4: audience questions in Perplexity, competitor content improvedPerplexity questions -> ideas;›

Perplexity questions -> ideas; competitor posts -> improved versions.

Perplexity prompt (0:36-0:43)

Competitor procedure (0:43-0:50)

↓ Full write-up of this concept

The concepts in full

01

Batch creation, the five-step workflow, and the positioning canvas

Two hours on Sunday, thirty posts, and a one-page document the AI reads before every prompt.

Five steps: generate ideas, write hooks, write posts, schedule, review weekly and iterate - in one weekly block instead of daily writing. Before any of it, ask the persistent chat to summarise the pillar work into a 'positioning canvas' (story, mission, one person, pains, offer, frameworks), save it in Google Docs, and upload it to any fresh chat so prompts stay specific instead of generic.

Why it matters

Context is what separates 50 usable ideas from 50 generic ones.

02

Method 1: the 50-idea expertise-mining prompt

Twenty how-tos, fifteen mistakes, ten frameworks, five hot takes - from the canvas, in one run.

Prompt: 'I need 50 content ideas for my LinkedIn based on my expertise. My background: [niche, audience, expertise, problems I solve, transformations, frameworks]. Create 50 ideas: 20 how-to, 15 mistakes, 10 frameworks, 5 controversial. Format each as hook style, headline, content type, key insight.' Follow-up: 'Pick the 10 with the highest potential to go viral and say why.' Run in ChatGPT, then repeated in a fresh Claude (Opus) chat with the canvas uploaded for a second set.

Why it matters

A hundred specific ideas in ten minutes.

03

Method 2: the 5-year reflection interview (seven pillars, by voice)

Mistakes, challenges, insecurities, achievements, lessons, skills, routines - talk for a minute on each and get seventy ideas nobody else can post.

Pick a window (2, 5 or 10 years). Seven pillars with seed questions ('most expensive mistake?', 'imposter-syndrome moment?'). Prompt: 'Help me mine my past 5 years for LinkedIn content using the 7 reflective pillars. Interview me one pillar at a time. After each answer, generate 10 specific content ideas.' Answer in voice or dictation mode rather than typing - demoed live, with a voice-mode glitch and a switch to dictation. This feeds the 25% vulnerability pillar with content that cannot be found generically.

Why it matters

The highest-viral-potential source, and the most personal.

04

Methods 3 and 4: audience questions in Perplexity, competitor content improved

Ask Perplexity what your buyers are actually asking on Reddit at midnight. Then read your ten best competitors.

Method 3: 'What are the top 50 questions [audience, e.g. local businesses in real estate, cleaning, manufacturing doing $50k+/month] are asking about [topic]?' - Perplexity searches LinkedIn, Reddit, Quora, Google and forums and sorts by most frequent, most painful, most misunderstood; then 'create 30 ideas: hook, angle, 3-5 points, CTA'. The trainer had heard one generated question from a real prospect that day. Method 4: pick 5-10 growing competitors, collect 5-10 best posts each (Supergrow analytics optional), write a better version with your own opinion and experience - inspiration, not copying ('Steal Like an Artist').

Why it matters

Grounds authority content in real pain and proven formats.

Tools referenced

ToolCoverageMomentContext
ChatGPTdemonstratedCanvas, 50 ideas, voice interview
Claude OpusdemonstratedSecond idea set with canvas uploaded
Google DocsdemonstratedPositioning canvas
PerplexitydemonstratedAudience questions
SupergrowmentionedCompetitor post analytics

Action items

    Resources mentioned

    Resources
    • docSteal Like an Artist (Austin Kleon)

    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
    Claude Fordunresolved - inside a quoted competitor post
    Lignin / lean inLinkedIn
    GrowthCool / GrowSchoolGrowth School
    agent valueAgent Valley
    Sutta / Utham / AutumnUttam (trainer)

    True on recording day — verify before relying