ARHAT GROWTH OS • FINAL ZERO-COST PILOT
Find what Arhat should own.
The system first scans the whole market, then removes repetition, finds white-space, ranks opportunities and only then turns one selected opportunity into content.
0research sessions
0topics selected
0content kept
0content killed
0taste rules
What this engine protects against
Repeated topics, confirmation bias, fake certainty and generic content.
Market-wide
Broad discovery
Research spans procurement, infrastructure, channel, workplace IT, pricing, finance, cyber, operations, technology shifts and buyer behavior.
Diversity
Territory control
The council groups duplicate ideas and actively looks for different territories before producing the Top 5.
Editorial
Quality gate
“Promising” is allowed during discovery. “Skip” is reserved for the final publishability decision.
Current selected opportunity
Nothing is forced. The engine can recommend more research instead of a post.
No opportunity selected yet.
STEP 01 • MARKET DISCOVERY
Research Inbox
Do not give the AIs a topic. Let them discover the market first. Each model has a different job, so the three research streams are intentionally not the same prompt.
Important: Paste the answers from the tools here, not the prompts. If a model returns another prompt or asks for clarification, run it again until you have actual findings, sources, observations or strategic analysis.
Perplexity — Evidence Scout
Current developments, primary sources, competitor activity, market signals.
Gemini — Buyer & Market Analyst
Buyer pain, emerging behavior, implications, contradictions and opportunities.
ChatGPT — Strategy & White-space
Independent strategic scan: what is interesting, under-owned and credible for Arhat.
Perplexity prompt
Gemini prompt
ChatGPT prompt
STEP 02 • OPPORTUNITY COUNCIL
Topic Council
The council sees the full research pack. It creates an opportunity universe, clusters duplicates, checks territory diversity, applies cooldown memory, and returns only five useful topic cards plus a compact audit.
Council Prompt
Paste Council Result
Paste the council answer. It may contain the 5 cards and the audit. The dashboard hides the long working by default.
Top 5 Opportunities
Each card is deliberately different where the evidence allows it.
STEP 03 • DEEP DIVE
Deep Dive
Only after you choose one topic do we spend attention on exhaustive verification and a hostile critique. This prevents research effort being wasted on weak ideas.
Selected Opportunity
Choose a topic first.
Deep Research — ChatGPT
Verify facts, counter-evidence, current examples, source quality and what is opinion vs fact.
Adversarial Critique — Claude
Try to kill the angle: genericity, weak claims, competitor saturation, credibility risk, buyer indifference.
Deep Dive Result
Save the research + critique before moving to Content Studio.
STEP 04 • MASTER CONTENT
Content Studio
Now the Master Brain can finally write. It receives the selected opportunity, the market pack, deep verification, critique and Arhat Taste.
Selected Opportunity
Choose a topic first.
Master Content Prompt
Paste Final ChatGPT Output
The prompt requests a compact final package. The dashboard extracts only finished assets; full reasoning stays hidden.
Final Assets
Only the publishable pieces are shown here.
MARKET → BUYER → PIPELINE
Growth & Buyers
A serious B2B growth system eventually needs conversation and buyer data, not just posts.
Buyer map
Target people
CIO / CTOIT HeadProcurementIT ManagerSystem IntegratorFounderStartup
Intent
Signals to watch
Questions, complaints, implementation friction, vendor pain, infrastructure decisions and buying triggers.
Conversion
Real success
Relevant attention → conversation → target account → website → enquiry → pipeline.
Buyer Conversation Log
LEARNING LAYER
Arhat Taste
Three separate things: what you love, what you hate, and what you want the AI to obey as a permanent rule. Rules never contaminate Love/Hate metrics.
❤️ Lovedrag ↔
❌ Hatedrag ↔
📌 Permanent Rulesnever scored as hate
Recent Feedback
30-DAY LEARNING
Learning Ledger
The learning ledger tracks what you select, what you reject, which territories you cover and what style becomes unmistakably Arhat.
| Signal | Current | Interpretation |
|---|---|---|
| Research sessions | 0 | Market scans completed |
| Topics selected | 0 | Opportunities you approved for development |
| Recent territories | — | Used to avoid repetitive topic clusters |
| Love signals | 0 | Positive taste evidence |
| Hate signals | 0 | Negative taste evidence |
| Permanent rules | 0 | Instructions, not rejection signals |
Topic Cooldown Memory
QUALITY GATE
System Test
This checks the core logic and parsers instead of merely checking whether buttons exist.
Ready.