ZIBATISFARZAM / BUILDS
EN
Contact ↗

FARZAM KHAKBAZ / BUILDS

I understand most things by building them.

From early ventures and brand-building to role-specific GPTs, assessment tools, content systems and operational workflows—this is a record of things designed around real problems.

Explore AI systemsExplore the venture path ↘
VENTURES · AI SYSTEMS · GPTs · DIGITAL PRODUCTS

BUILDER MINDSET

An idea becomes serious when it can be tested.

A large part of my professional path was shaped through direct building—from early ventures and real market consequences to digital products and AI systems.

Much of what I know about brand, growth, product and AI came through building, failing, observing, refining and rebuilding.

  1. 01BUILD
  2. 02OBSERVE
  3. 03LEARN
  4. 04REBUILD

ENTREPRENEURIAL EVOLUTION

Before AI, the problem was the same: how do we turn an idea into something real?

This is not a list of job titles. It is an evolution through markets, products, operations and repeatable systems.

01EARLY-STAGE STARTUP

Va3manchand

Group-buying startup

One of my first serious encounters with markets, distribution, customer acquisition and product–market reality.

  • Founder
  • Marketplace experiment
  • Customer acquisition
  • Partnership model
  • Offer economics
LEARNINGA good idea is not enough without distribution and positioning.
02D2C BRAND

Ferengi Clothing

Independent graphic apparel brand

Building identity, product, content and a market for an independent graphic apparel brand.

  • Brand identity
  • Product concepts
  • Creative direction
  • Pricing
  • Sourcing
  • D2C marketing
LEARNINGA brand is not only what we design; it is what the market experiences through the product.
03VENTURE / OPERATIONS EXPERIENCE

Hospitality Venture

Restaurant launch & operations — Tabiat Bridge, Tehran

Launching and operating a physical hospitality business—not presenting it as a software startup.

  • Customer experience
  • Service delivery
  • Team coordination
  • Supplier operations
  • Daily execution
LEARNINGEvery business model eventually has to work in the real world.
04FOUNDER / BUSINESS SYSTEM

Zibatis

Business, Brand & AI Systems Studio

Where business, brand, content, digital experience, growth and AI converged inside a real delivery environment.

  • Business
  • Brand
  • Content
  • Digital Experience
  • Growth
  • AI
  • Operations
LEARNINGIdeas, expertise and execution create value when they become a coordinated system.
Explore Zibatis ↗

Then the tools changed.The problem did not.

The earlier question was:

How can we design a better business process?

Now another question has joined it:

Which part of it can become an AI system?

THE SHIFT

For me, AI did not begin with prompting. It began with the problem.

When language models entered real workflows, the interesting part was not the chatbot. It was the relationship between expertise, data, rules, workflow and human control.

CUSTOM GPT ECOSYSTEM

Systems designed for a specific role.

Instead of one generic GPT, these systems have distinct roles, knowledge boundaries, workflows and guardrails. The number describes systems and structured experiments—not active users or production apps.

20+DOMAIN-SPECIFIC AI SYSTEMSEVIDENCE-CONTROLLED LABEL

9 systems shown

01 / MARKETING OS

Zibatis Digital Marketing & Copy Writing

Marketing strategy, SEO content, copy and publishing workflow.

System detail
Problem
Fragmented production without a connection between search intent, brand voice and conversion.
Role
Connect Strategy → Content → SEO → Copy → CTA → Publishing.
Architecture
  1. Strategy
  2. Content
  3. SEO
  4. Copy
  5. CTA
  6. Publishing
Evidence
Role definition, instructions and workflow are documented in the systems portfolio.
02 / MEDICAL CONTENT SYSTEM

Medical Content Orchestrator

A multi-role medical content planning and quality-control system.

System detail
Problem
Disconnects between research, medical accuracy, search, structure and human review.
Role
Orchestrate roles and make responsibility explicit at each stage.
Architecture
  1. Strategy
  2. Research
  3. Medical Writing
  4. SEO
  5. Schema
  6. Visual Brief
  7. Human Review
Evidence
Role-based workflow and QA checkpoints are documented system artefacts.
Human / clinical boundary
Medical outputs require human review and accountable domain expertise.
03 / SEMANTIC SEO

Schema Optimizer Pro

Structured-data architecture and JSON-LD generation grounded in page and entity understanding.

System detail
Problem
Fragmented schema, unstable IDs and disconnected entities.
Role
Extract entities and design a coherent graph for search and AI retrieval.
Architecture
  1. Page understanding
  2. Entity extraction
  3. Schema graph
  4. Persistent IDs
  5. Structured output
Evidence
Compiler logic, graph relationships and output controls are documented.
04 / CREATIVE INTELLIGENCE

Creative IDE

An environment for developing creative direction—not a generic idea generator.

System detail
Problem
Clichéd solutions and premature closure of the possibility space.
Role
Break patterns, create routes and evaluate creative territories.
Architecture
  1. Tension
  2. Contradiction
  3. Routes
  4. Evaluation
  5. Creative Territory
Evidence
The ideation framework and evaluation logic are documented in the portfolio.
05 / MEDICAL SEO

BrandMaster Medical Writer

Medical content, SEO, HTML-ready publishing and CTA architecture.

System detail
Problem
Medical writing that is accurate but hard to publish—or search-ready but weakly grounded.
Role
Coordinate content, publishing structure and the conversion path.
Architecture
  1. Medical context
  2. Search intent
  3. HTML structure
  4. CTA
  5. Review
Evidence
Role definition and output structure are documented in the GPT ecosystem.
Human / clinical boundary
Final publishing should pass domain and editorial QA.
06 / HEALTHCARE ASSISTANT

Dr. YK Diet Farsi

A Persian bariatric nutrition and patient-education assistant.

System detail
Problem
Patients need clear, consistent education outside limited consultation time.
Role
Provide education and informational support inside a defined boundary.
Architecture
  1. Approved knowledge
  2. Patient context
  3. Education
  4. Escalation
  5. Human review
Evidence
Scope and usage limits are defined in the system specification.
Human / clinical boundary
Not a replacement for a clinician, diagnosis or clinical decision-making.
07 / DOMAIN KNOWLEDGE

Dental KnowledgeGPT — Dr. Salamat

A domain knowledge assistant for dental content, education and communication.

System detail
Problem
Specialist knowledge is fragmented and difficult to reuse across content and education.
Role
Ground responses in domain context and organise retrievable knowledge.
Architecture
  1. Domain sources
  2. Knowledge structure
  3. Retrieval context
  4. Communication
  5. Review
Evidence
The knowledge model and intended uses are documented in the portfolio.
08 / CREATIVE DIRECTION

Taste Architect

Creative direction, taste calibration and visual concept development.

System detail
Problem
Generative output can be technically correct but lack direction, coherence and perceptual quality.
Role
Define criteria, direct choices and critique output.
Architecture
  1. Intent
  2. Reference logic
  3. Taste criteria
  4. Direction
  5. Evaluation
Evidence
The calibration framework and concept routes are documented.
09 / VISUAL AI

Dental Luxury Image Studio

Dental visual direction with clinical credibility and premium aesthetics.

System detail
Problem
AI imagery often misses material behaviour, treatment realism or brand consistency.
Role
Connect clinical credibility, material logic and art direction.
Architecture
  1. Clinical reference
  2. Material behavior
  3. Composition
  4. Brand consistency
  5. QA
Evidence
Prompt systems, constraints and output examples are documented in the visual portfolio.

SOLUTION SELECTION

Not every problem should become a GPT.

01

GPT

When the work is primarily knowledge, reasoning and structured output.

KNOWLEDGE + REASONING + OUTPUT
02

WORKFLOW

When a trigger, data, transformation and repeatable action must connect.

TRIGGER + DATA + TRANSFORMATION + ACTION
03

PRODUCT

When users need an interface, state, rules, data and a persistent experience.

INTERFACE + STATE + RULES + DATA
AI solution design is the choice between a GPT, a workflow and a product.

DIGITAL PRODUCTS & EXPERIMENTS

Things that moved beyond chat.

Here, the value sits in the interface, state, rule engine, data model or multi-step coordination—not in one prompt.

01DECISION SUPPORT

Smart Assessment Systems

Interactive forms with branching, scoring, lead qualification and personalised output.

  1. INPUT
  2. RULES
  3. SCORE
  4. OUTPUT
02HEALTHCARE PRODUCT

Bariatric Assessment System

Patient inputs, rules, risk flags, method context and a human consultation path.

  1. PATIENT
  2. RULES
  3. FLAGS
  4. CONSULTATION
Intake and decision support—not autonomous diagnosis.
03KNOWLEDGE OPERATIONS

Knowledge & Content Systems

Content OS, medical knowledge structures, SEO architecture and structured publishing.

  1. SOURCE
  2. STRUCTURE
  3. RETRIEVAL
  4. PUBLISH
04DECISION INTELLIGENCE

Management Dashboards

Interfaces that turn data into meaning, priority and action for managers.

  1. DATA
  2. MEANING
  3. PRIORITY
  4. ACTION
05RESEARCH / EXPLORATION

Agent-Ready Website Experiments

Exploring the shift from websites as pages to capabilities callable by agents.

  1. WEBSITE
  2. CAPABILITY
  3. TOOLS
  4. AGENT
Research in progress; the article link activates only after the full source article is published.

AI PRODUCT PRINCIPLES

Four rules for building something that works in the real world.

  1. 01

    Problem before model

    Define the problem, impact and constraint before selecting a model or tool.

  2. 02

    Context before prompting

    A prompt without durable context, sources and output boundaries is not a system.

  3. 03

    Human control before autonomy

    Automation needs explicit approval, escalation and human accountability.

  4. 04

    Learning before scale

    Pilot first; collect evidence and refine before scaling.

SYSTEM MAP

The common thread across these projects is not technology. It is systems.

Ventures, AI systems and digital products converge around a shared systems approach.

EVIDENCE-CONTROLLED SCALE

Name the scale precisely.

20+

Domain-specific AI systems

Documented GPTs and structured experiments—not necessarily standalone production apps.

6

System families

Marketing · Medical · Semantic · Creative · Knowledge · Visual AI

4

Venture chapters

Startup · D2C Brand · Operations · Zibatis

RELATED NOTES

Ideas I think through while building.

Notes and research routes come from the live Insights architecture. No empty article or placeholder route was created to complete this page.

Explore insight routes ↗

BUILDER ↔ DELIVERY ENVIRONMENT

Some of these experiments now become systems inside Zibatis.

What appears here as Farzam’s personal build portfolio can become a product, agent, workflow or digital infrastructure when connected to a real business problem through Zibatis.

Farzam = Builder / Thinker · Zibatis = Delivery Environment

BUILD SOMETHING USEFUL

Have a problem that might deserve a system?

You do not need to know whether the answer is a GPT, agent, automation or full product. Define the problem first; then decide what is worth building.

Start with the problem ↗Explore AI Agent Systems