I Built 5 Products With AI and Never Wrote a Line of Code
After 59 AI development sessions, 94,155 messages, and 3,072 conversations, I wrote the methodology down. It’s called The Re-Anchor Manager.
Mohan Iyer
AI ARCHITECT
50 years as a metallurgist and industrial engineer. 12 years building WordPress solutions — themes, plugins, the whole stack.
At 74, I started working with large language models the week ChatGPT launched. I brought factory-floor instincts with me: quality gates, role separation, shift handovers, and the conviction that systems must not fail.
The hardest problem wasn’t getting the AI to write code. It was getting it to remember what it wrote yesterday.
The solution became The Re-Anchor Manager — a methodology for maintaining continuity across AI development sessions, born from real projects and measured against real outcomes.
The proof? Seekrates AI — a 5-model consensus platform, 164 published posts at 94% quality, and a publishing pipeline that runs without me touching code. All built with the methodology this book describes.
Two ways to work smarter with AI. Use Seekrates to get answers you can trust. Read The Re-Anchor Manager to build AI projects that don't forget.
THE PROBLEM (Why This Exists)
The Single-AI Problem
THE SOLUTION (What Seekrates Does)
Five AIs. One Consensus. Zero Guesswork
Five AIs. One Consensus. Zero Guesswork
HOW IT WORKS (4-Step Visual Flow)
From Question to Confidence in 60 Seconds
Step 1: You Ask
Enter your question Any complex, forward-looking, or high-stakes query. "What skills will marketing managers need by 2028?" "How will AI change my industry?"
Step 2: Five AIs Respond
Independent reasoning, no copying Your query goes to all five models simultaneously. Each answers from its own training data. Different knowledge ecologies. Different perspectives.
Step 3: The Referee Scores
Hallucinations filtered, contradictions flagged Our synthesis engine compares all five responses, identifies agreement patterns, scores each answer on accuracy and coherence, and catches the outliers.
Step 4: You Get Consensus
One answer you can trust A validated synthesis, a confidence percentage, and each AI's individual reasoning—all delivered to your inbox. Transparent. Traceable. Actionable.
The Methodology Behind the Machine
Seekrates didn’t happen by accident. It was built across 59 consecutive AI development sessions — each one starting from zero because the AI forgot everything from the day before. The Re-Anchor Manager is the system that solved that problem. It’s how a 74-year-old industrial engineer kept five AI-built products on track without writing a single line of code. This is the methodology, documented.
If the methodology did not work, this book would not exist. That is the simplest proof I can offer.
The Seekrates Method, Explained
THE ESSENTIALS
| Video | Title | Description |
|---|---|---|
| 1 | What is Seekrates? | “ChatGPT gives you one opinion. Seekrates gives you five.” 60 sec explainer on multi-AI consensus. |
| 3 | How to Ask the Future | “The question you ask determines the answer you get.” Learn the TIME + YOU + SPECIFIC formula. |
GOING DEEPER
| Video | Title | Description |
|---|---|---|
| 2 | Rear-View vs Windscreen | “Google shows where you’ve been. Seekrates shows where you’re going.” Search vs Ask. |
| 4 | 3 Questions NOT to Ask | “These questions waste your query.” Avoid past-dragging, yes/no, and vague questions. |
Browse the Blog
Explore AI career forecasts, SEO strategies for the AI era, prompt engineering tutorials, and timeless wisdom from history’s greatest thinkers—all validated through 5-model consensus.
Built for Bold Thinkers in Every Field
Career & Workforce
- “What skills will project managers need by 2030?”
- “How can nurses transition to health tech roles?”
- “Which MBA specializations will matter in 2028?”
Business Strategy
- “What’s the best pricing model for a freemium app?”
- “How will AI change supply chain management?”
- “Which industries are most resistant to AI disruption?”
Content & Marketing
- “How to optimize content for AI search engines?”
- “What is Generative Engine Optimization?”
- “How do AI crawlers decide which sources to cite?”
Research & Analysis
- “Compare AI predictions across multiple models”
- “Detect hallucinations before they cost you”
- “Get second opinions on complex topics”
Why 5 AIs Beat 1
| Single AI (ChatGPT, Claude, etc.) | Seekrates (5-AI Consensus) |
|---|---|
| One perspective | Five independent perspectives |
| Hallucinations go undetected | Hallucinations caught by disagreement |
| No confidence signal | Consensus % shows certainty |
| Black box reasoning | Transparent: see each AI’s answer |
| “Probably right” | “Validated across 5 models” |
The Herding Cats Reality
Building a functional 5-LLM consensus system is—as one user put it—”taking 5 hallucinating cats for a walk.” Theoretically possible. Practically, a nightmare. We solved it so you don’t have to.
Frequently Asked Questions
Who is Mohan Iyer?
Mohan spent 50 years as a metallurgist and industrial engineer before pivoting to AI architecture at 74. He built Seekrates AI, authored the authored The Re-Anchor Manager, and has published more than 200 consensus-validated articles on this site. He describes getting five hallucinating AIs to agree as “herding cats”—and wrote the playbook so you don’t have to.
What is Seekrates AI?
Seekrates is a multi-agent AI consensus platform. Instead of asking one AI and hoping it’s right, Seekrates queries five leading language models (OpenAI, Claude, Gemini, Mistral, Cohere) simultaneously, synthesizes their responses, and delivers a validated answer with a confidence score.
How is this different from just using ChatGPT?
ChatGPT gives you one AI’s opinion. That AI might hallucinate, have outdated training data, or simply be wrong—and you’d never know. Seekrates cross-validates across five models. When they agree, you have high confidence. When they disagree, you’ve found something worth investigating.
What does the consensus percentage mean?
The consensus percentage shows how much the five AIs agreed on the answer. 90%+ means strong agreement—act with confidence. 60-80% means nuanced topic—read the dissenting views. Below 50% means genuinely uncertain territory where your decisions matter more.
What kind of questions should I ask?
Best results come from forward-looking, specific questions. Instead of “Will AI take jobs?” ask “What skills will marketing managers need by 2028?” Add TIME (by when?), add YOU (my role, my industry), add SPECIFICITY (name your situation).
Is my data private?
Yes. Your queries are processed and not stored beyond the session. We don’t train on your questions. Your results are emailed only to you.
How much does it cost?
Free tier: 3 queries to test the system. Paid plans start at $9.95/month for 30 queries. See our Pricing page for details.
Can I see how each AI responded individually?
Yes. Every results email includes the synthesized consensus answer AND each AI’s individual response. Full transparency—you see exactly what each model said and how they differed.
Why “Seekrates”?
Named after Socrates, the philosopher who believed wisdom comes from questioning assumptions and examining ideas from multiple angles. Seekrates applies that principle to AI—multiple perspectives, rigorous examination, validated truth.