Video summary

How This 22 Yr Old Made ₹2 Crores With AI Video Making | Vyan Gandhi | #235 The Sanskar Show

Main summary

Key takeaways

Business

Founder + Business Model

  • Vyan Gandhi (22) runs an AI content/video production house (appears as “Augmented/‘Augmented So…’” in the subtitles), positioned as one of the first AI production houses in India.
  • Services
    • AI-made ads
    • Marketing content
    • Long-form + short-form content
  • Production approach
    • Content is generated using client-specific scripts.
    • Delivered in an agency-like model with minimal logistics (implied: laptops + a few people; no camera/crew workflow described).

Revenue + Scaling Claims (Key Metrics/KPIs)

  • Lifetime/episode claim: “Close to ₹2 crores” made using AI video production.
  • Scaling milestone: grew from ₹10 lakhs to ₹1.5 crore in ~10 months
    • attributed “mostly luck” plus a breakout video.
  • Breakout growth driver: “Mahabharat AI trailer”
    • ~1.7 million views (YouTube)
    • generated inbound leads and industry attention, including media coverage.
  • Client pipeline metrics (process KPIs)
    • Talk target: ~2 prospects per day (director production houses)
    • Deal cycle: closes on a script in 10–15 days
    • Client closing time: reiterated as ~15 days
  • Retainer economics / cost structure (unit economics proxy)
    • Example: charging ~₹30 lakhs from a client for AI video usage (context: Pocket FM).
    • ROI logic via unit economics:
      • A prompt engineer can produce ~2 minutes of content per day
      • Tool subscription stated as ~₹30,000 (for the 2 minutes/day assumption)
      • Daily prompt engineer labor: ~₹2,000–3,000
      • Profit rule of thumb: if AI project expense is lower than what the client would pay for physical production, it’s ROI.

Go-to-Market (GTM) + Client Acquisition Playbook

Inbound + Relationship-Driven Sales

  • Leads via YouTube virality and industry reputation, especially with directors/production houses.
  • Outbound: aims to talk to about two people per day using LinkedIn.

Partnership Targeting

  • Works with/has relationships with:
    • Pocket FM
    • People Media Factory (large South Indian production house; worked on a movie titled “Mirai” for AI scenes)

Retainer Conversion Path

  • Start project-based
  • Once quality/throughput is proven, pitch a retainer

Content Strategy: Short-Form vs Long-Form

  • Beginner recommendation: start with short-form
    • long-form is described as more labor-intensive
    • producing 20–30 minutes alone within a month is framed as difficult.
  • Short-form monetization targets (ad unit economics claim)
    • Ads can be priced/earned around ₹20,000–₹30,000 per ad
    • Expense target ~₹10,000 (implied AI tool/production cost)
  • Upgrade path: build skill + team gradually from short-form → long-form.

Breakthrough + Positioning Narrative (What Worked)

  • “Mahabharat AI trailer” was positioned as Game-of-Thrones-level quality realism at Indian mythological scale.
  • Outcome:
    • virality across YouTube + LinkedIn
    • mainstream mentions (Times Now, Hindustan Times)
    • meetings with major tech/content leaders (e.g., Sony CEO, Jio Cinema VP mentioned)

Prompting + Workflow “Playbook” (Execution Framework)

Tool Stack (Mentioned)

  • ChatGPT
  • Claude
  • Freepik (used as a place to access models/assets)
  • Video generation tools mentioned: SeeDance and Nano Banana
  • Kling (used in examples)
  • Seed/SeedDance noted as more expensive; Nano Banana/SeeDance recommended as better options.

Two Production Approaches

  • Image-first → convert to video + stitch
  • Character/background creation → generate different angles/shots
    • described as more expensive (SeedDance-style approach)

Prompt Template (Process Discipline)

A structured prompt should include:

  • Time of day
  • Shot type / camera framing
  • Character description
  • Background details
  • Camera model (e.g., ARRI Alexa)
  • Lens mm / focal length
  • Lighting direction
  • Additional director-like production metrics

Emphasis: tools require a lot of context for realistic output.

Iterative Improvement Loop

  • If output is wrong:
    • describe what output you got
    • specify the exact problem
    • request a fix
  • Claim: typically reach the desired result after 3–4 iterations.

Prompting Best Practice Advice

  • AI prompting is not just “Google search”—it needs:
    • context
    • instructions for consistency
  • Add a meta instruction such as:
    • “Ask me any questions you might have…” to fill missing production details.

Failure Analysis Method

  • Analyze what the AI “misunderstood,” then adjust the prompt:
    • time of day
    • camera angle / shot width
    • realism level
    • randomness
  • Example given: football shape confusion
    • AI outputs the wrong representation
    • fix by specifying the intended type (e.g., “white and black checkered” football)

Learning + Hiring Strategy (How to Build a Team)

Zero-to-First Clients (Portfolio Funnel)

  • Start by copying tutorials, then improving based on gaps.
  • Do free/low-stakes projects (suggested: local brands first rather than massive brands) to build a portfolio.
  • Example portfolio targets mentioned:
    • Haldiram
    • a Nike-style aspirational example
  • Timeline suggestion: within 3–4 months, aim for recognizable-brand-style portfolio ads.

Scaling Operations

  • Scaling and team-building are framed as necessary because:
    • entry barrier is low for ads (many competitors with laptops)
    • long-form requires massive teams + editors

Team Ahead-of-Competitors Claim

  • Claims they are ~7 months ahead of competition due to:
    • client relations
    • team
    • capability

Pricing + Value Justification (Business Tactics)

  • Project pricing → retainer migration once workflows stabilize.
  • Pricing is tied to:
    • quality level
    • number of iterations (subscription costs scale with refinement)
  • ROI argument for enterprise buyers
    • AI reduces costs vs physical production (fewer logistics; mostly subscription + labor)
    • supported by the cost breakdown for:
      • prompt engineers
      • tool subscriptions

Roadmap for Aspiring Builders (Actionable Recommendations)

Skill Acquisition Roadmap

  • Choose a skill aligned to interest (AI prompting/AI ads, video editing, etc.).
  • Avoid paid courses initially:
    • learn basics via YouTube
    • replicate real edits/ads to find weaknesses and gaps

Client Acquisition Roadmap

  • If you lack connections:
    • pitch marketing agencies first (instead of waiting to reach brands directly)
    • start with below-market pricing for initial work, then raise pricing later

Beginner Tool/Process Ramp

  • Start with image-first workflows
    • more learning
    • easier initial control
  • Use a prompt template + an iterative “diagnose + fix” loop

Recommended Learning Strategy

  • “Fail fast, learn fast” and continuously improve.

Presenters / Sources

  • Presenter/Guest: Vyan Gandhi (Founder, AI video production; also interviewed in the episode)
  • Host: “The Sanskar Show” (host name not provided in the subtitles)

Original video