For the Virani Family · Confidential · 2026

CREATIVEFUEL.

What we'll
build at Rajkot.

A 2-day AI training for Balaji Wafers, structured around 6 hands-on department labs. No theory. No prompt libraries you'll never open. Every participant walks out with a working AI prototype, built live on Balaji files, that runs on Monday morning.

02Days on-site

06Dept tracks

06Working prototypes

100Attendees max

15AI tools taught

30Days office hours

17

Balaji depts covered

4-5

CF practitioners in room

14

Hours hands-on

80+

Balaji prompts in library

100%

On your actual files

Before The Workshop

2-week diagnostic. Half-day with 6 dept leads. We arrive at Rajkot with your data, not generic FMCG case studies.

During The Workshop

2 days, on-site at the factory. Day 1 foundations, Day 2 six parallel labs. Every block has a fixed exercise and a fixed output.

After The Workshop

30 days of office hours. 6 months of private Q&A channel. Workshop recordings for repeat training of joiners.

Optional Phase 3. If the workshop ships well, 4 to 6 prototypes go into production over the next 6 months. Internal Balaji AI cell stood up. Scoped only after Phase 2 closes, with the family in the room. Indicative range ₹75L to ₹1.25 Cr. Page 7 has the full numbers.

Tuesday morning at Balaji, post-workshop

6 working prototypes On internal Drive/Slack. Each dept head with a runbook. Each one usable Day 1 after workshop.

A draft Balaji AI Code Co-drafted with leadership in Day 1 security session. Ready for family review and company-wide rollout.

A Monday Plan, per dept Every dept head writes one before they leave Rajkot. The 3 AI things they ship in the first 30 days.

Creativefuel · Mumbai HQ
creativefuel.io · An AI-native operating company

Edition 1 · 2026
For private review

Section 01 · The Schedule

A workshop you can read before you sign it.

Day 01Foundations + Live Builds

09:00Kickoff. Family + dept heads. Set the room. 15 min

09:15Where AI actually is (Q2 2026). What's real. What's hype. 45 min

10:00LabChatGPT vs Claude vs Gemini. Same 5 Balaji questions. Pick a daily driver. 40 min

10:40TeaChai break 15 min

10:55The Prompt Brick: Role, Task, Context, Constraints. Live rewrite of a real Balaji prompt. 50 min

11:45LabAI inside Excel. SKU velocity analysis on your Q4 numbers, using Copilot + Claude API. 60 min

12:45LunchLunch 45 min

13:30LiveNotebookLM ingests 50 Balaji PDFs in 2 minutes. Q&A live. 45 min

14:15LiveBuild a distributor anomaly query. In front of the room. 45 min

15:00TeaTea + reset 15 min

15:15LabPack-design sprint. 6 pack variants. 3 image models. 25 min. 30 min

15:45LabVoice + video. A 15-sec Hindi+Gujarati launch reel, end to end. 60 min

16:45Data security & the Balaji AI Code. What goes in. What stays out. 45 min

17:30Day 1 wrap. Track assignments for Day 2. 30 min

Day 026 Parallel Dept Labs

09:00Track kickoff. Each track gets a CF practitioner + Balaji data files. 30 min

09:30LabTrack Lab 1 of 4. (See Section 02 for track-by-track detail.) 75 min

10:45TeaTea + cross-track mingle 15 min

11:00LabTrack Lab 2 of 4. Building the data layer. 75 min

12:15LabTrack Lab 3 of 4. Building the AI layer. 45 min

13:00LunchLunch. Mixed seating across tracks. 45 min

13:45LabTrack Lab 4 of 4. Stitch + test prototype. 75 min

15:00SimThe 60-Min Cross-Functional Simulation. Launch a Q-comm SKU. 60 min

16:00TeaTea before presentations 15 min

16:15LiveEach track presents its prototype. 12 min per track. 75 min

17:30Wrap. 30-day office hours scheduled. Monday plan written down. 30 min

What Balaji brings to Day 1

›1 sanitised SKU performance file (Q4 last year)
›1 sanitised distributor report (top 50 dealers)
›10 sample SOPs from any plant line
›1 marketing brief from a recent SKU launch
›Wi-Fi for 100 laptops & an LED screen for projection

What Creativefuel brings to Day 1

›4 to 5 practitioners (data, vision, voice, ops)
›Enterprise licences for all 15 tools, pre-provisioned
›Printed workbook + prompt library, one per attendee
›Pre-built lab notebooks on sanitised data
›NDA, MSA & data handling protocols ready to sign

Balaji AI Training · Creativefuel 01 / 09

Section 02 · The 6 Department Labs

6 labs. 6 prototypes. One day.

Six department labs run in parallel from 09:30 to 15:00, each led by a Creativefuel practitioner, each shipping a working prototype on sanitised Balaji data. Prototypes presented to the family at 16:15.

LAB 01Plant Floor

The Fryer Line.

Production · Engineering · Maintenance · Packaging

Defect detection. Train a vision model on 30 line photos. Catch burns, breaks, discoloration.Claude Vision · Roboflow
Plant SOP search. 20 SOPs ingested. Query in Gujarati: "how to clean line 3 after a flavour switch?"NotebookLM · Whisper
Yield model. Regression on 6 months of fryer parameters. Predict oil saturation %.Python · Claude in Notebook
Predictive maintenance. Vibration log ingestion. Flag risky bearings 48 hrs out.Pandas · Anomaly detection

You Ship A working defect classifier, an SOP search bot, a yield trend dashboard.

LAB 02Supply Side

Vendors & Sourcing.

Supply Chain · Stores · Purchase (Raw + Mach)

Vendor scoring sheet. 30 vendor rows. Price, quality, dispute frequency, on-time. Auto-rank top + bottom 10.Excel + Claude API · Copilot
Potato price model. Weather + sourcing region + price history. 60-day forecast band.Pandas · Claude reasoning
Contract risk reader. Paste any vendor PDF. AI surfaces clauses, exit triggers, hidden penalties.Claude Sonnet 4.6
Spend conversation. Ask in English: "Where did we overspend Q3 vs Q2?" Get the answer.SQL + LLM glue

You Ship A vendor scoring spreadsheet, a price-forecast notebook, a contract risk reader.

LAB 03Dealer Network

The Dealer Bot.

IT · IT Support (Dealers + CNF) · Dispatch

WhatsApp FAQ bot. 50 dealer questions trained. Replies in Gujarati, Hindi, English.Twilio · Claude API
Helpdesk auto-responder. Common IT tickets resolved without a human.Claude function-calling
Dispatch load optimiser. Order list + truck capacity. Get optimal load plan in 10 sec.OR-tools + LLM
Voice IVR demo. Dealer calls. Voice bot takes order. Sample call recorded.ElevenLabs · Twilio Voice

You Ship A live dealer WhatsApp bot link, a help-desk responder, a dispatch load optimiser.

LAB 04Demand Engine

Distribution & Q-Comm.

Sales · Marketing · MT · Q-Commerce · Ecom

Distributor health scan. Spot which distributors ordered 2σ below pattern this month.Pandas · Claude diagnosis
Q-comm listing generator. Generate 50 SKU listings for Zepto/Blinkit. With Hindi variants.Claude Sonnet + style guides
Retail audit photo scorer. Upload shelf photo. Get shelf-share by brand.Claude Vision
SKU velocity forecast. 6 metros × 60 days. Reorder cycle predicted.Prophet + LLM commentary

You Ship A distributor health dashboard, a listing generator, a shelf-share scorer.

LAB 05Product Lab

R&D & Exports.

R&D · Pack Design · Exports

Regional flavour mining. Scrape Amazon, Reddit, Q-comm reviews. Cluster trending flavour signals.Firecrawl · Claude clustering
Competitor SKU clusterer. Ingredient + nutrition + price across 200 SKUs. Spot whitespace.Embeddings · UMAP
Pack-design A/B. 20 pack variants generated. Attention map ranks them before research.Nano Banana · Midjourney
Export market ranker. Diaspora data + e-comm signals. Rank top 10 country × SKU combos.Trade data + Claude

You Ship A flavour trend dashboard, an SKU clusterer, an export opportunity ranker.

LAB 06Back Office

Finance & HR.

Finance · HR

Invoice to ERP. Drop a folder of supplier PDFs. Output: clean ERP-ready CSV.Claude Vision · Tabula
Expense fraud detector. Anomaly detection on 6 months of claims.Isolation Forest · LLM review
Variance commentary. Drop a P&L. AI writes the board commentary in 60 sec.Claude Sonnet 4.6
Attrition risk model. Anonymised HR data. Predict 90-day quit risk by role.XGBoost · LLM explainer

You Ship An invoice processor, a fraud detector, a variance writer, an attrition dashboard.

Balaji AI Training · Creativefuel 02 / 09

Section 03 · Tools You'll Master

15 AI tools.
One operator's toolkit.

Specific tools, specific use cases, specific blocks on the schedule. Not a generic list. By end of Day 2 every participant has signed in, hands-on, on each one that matters for their function.

Capability	Tools You Use	Where it Plugs In
Chat reasoning	Claude Sonnet 4.6 · GPT-5 · Gemini Pro 2.5	Day 1 daily-driver lab. The tool every dept head picks once and uses thereafter.
Vision / images in	Claude Vision · GPT-Image · Gemini Vision	Lab 01 (fryer defects), Lab 04 (shelf audit photos), Lab 06 (invoice OCR).
Image generation	Nano Banana · Midjourney v7 · GPT Image · Flux	Day 1 pack-design sprint. Lab 05 (R&D pack A/B variants).
Voice synthesis	ElevenLabs (Hindi, Gujarati, Marathi)	Day 1 voice block. Lab 03 (dealer IVR demo).
Video generation	Veo 3 · Pika · Kling	Day 1 launch reel. The 60-min simulation. Festival campaign mocks.
Spreadsheet AI	Microsoft Copilot in Excel · Claude API in spreadsheets	Day 1 SKU velocity lab. Lab 02 (vendor scoring). Lab 06 (variance commentary).
Deep research	NotebookLM · Perplexity Deep Research · Claude Projects	Day 1 NotebookLM live demo. Lab 05 (competitor + flavour mining).
Web scraping	Firecrawl · Apify · Browserbase	Lab 05 (review scraping). Lab 04 (Q-comm price comparison).
WhatsApp + Voice bot	Twilio API · Claude function-calling · ElevenLabs voice	Lab 03 (dealer FAQ bot, voice IVR). The big quick-win.
Agentic workflows	Claude Computer Use · n8n · Zapier AI	Day 1 agentic preview. Lab 06 (back-office automations).
Coding co-pilot	Cursor · Claude Code · GitHub Copilot	All labs. Track lead uses it to ship prototype faster.
Data + analytics	Pandas · Polars · DuckDB · Claude analyst persona	Lab 01 (yield model), Lab 02 (price forecast), Lab 04 (velocity), Lab 06 (attrition).
Vector + semantic search	pgvector · Pinecone · BGE-M3 embeddings	Lab 01 (SOP search), Lab 05 (competitor SKU clustering).
Enterprise admin	Claude Team · ChatGPT Enterprise · Gemini for Workspace	Day 1 data-security block. The Balaji AI Code session.
Translation + OCR	Whisper · Indic-Trans · AWS Textract	Lab 01 (Gujarati SOP search). Lab 06 (invoice OCR in any script).

Balaji AI Training · Creativefuel 03 / 09

Section 04 · Take-Home Artefacts

Tuesday morning, you have:

Files, dashboards, prototypes, bots. Real things, not workbooks. Every track leaves with a folder of working AI artefacts the function will keep using. Plus a 30-day post-workshop window for questions, fixes, and follow-on.

Lab 01 · Plant Floor

Fryer defect classifier.ipynb + Claude prompts
Plant SOP search bot in GujaratiNotebookLM project
Yield trend dashboard for shift managers.xlsx + Claude formulas
Predictive maintenance notebook on sensor logsPython + Pandas

Lab 02 · Supply Side

Vendor scoring spreadsheet auto-ranked.xlsx + Anthropic plugin
Potato price 60-day forecast notebook.ipynb
Contract risk reader for any PDFClaude Project
Spend chatbot in plain EnglishSQL + LLM

Lab 03 · Dealer Network

Live dealer WhatsApp FAQ botTwilio link + 50 Q&As
IT helpdesk auto-responderClaude function-calling
Dispatch load optimiserCSV in, plan out
Voice IVR sample recordingElevenLabs · Twilio

Lab 04 · Demand Engine

Distributor health dashboard with anomaly flagsStreamlit app
Q-comm listing generator, 50 SKUs in 30 secClaude prompt chain
Shelf-share photo scorerClaude Vision
60-day SKU velocity forecast for 6 metrosProphet model

Lab 05 · Product Lab

Regional flavour trend dashboard auto-refreshedFirecrawl + Claude
Competitor SKU clusterer across 200 SKUsEmbeddings + UMAP viz
20 pack-design variants shortlistedNano Banana exports
Top-10 export opportunity rankingTrade data + Claude

Lab 06 · Back Office

Invoice-to-ERP pipeline, PDF in, JSON outClaude Vision + Tabula
Expense fraud detector flags 2-3%Isolation Forest
Variance commentary writer, 1 min vs 1 weekClaude Sonnet
Attrition risk dashboard by roleXGBoost + LLM

For All Participants · Every Track

Balaji prompt library, 80+ prompts across the 17 departments, indexed by function and use caseNotion + .md export
The Balaji AI Code, a 1-page data security & usage guideline, co-drafted with leadership for family sign-offPDF
30-day office hours, 5 sessions of 90 min each, post-workshop for follow-on questions and fixesZoom + Slack
Workshop recordings, the sanitised portions of every session, available offline for repeat training of new joiners.mp4
Internal Q&A channel, a private Slack or WhatsApp group with Creativefuel practitioners for 6 monthsLive

Balaji AI Training · Creativefuel 04 / 09

Section 05 · The Day-2 Simulation

Launch a Balaji SKU
in 60 minutes. Live.

At 15:00 on Day 2, every track re-shuffles into 6 cross-functional teams of 10 to 15 people each. Each team has 60 minutes to deliver a complete launch package for a brand-new Balaji wafer or namkeen, using only AI. At 16:15, every team presents to the family.

The Brief

Balaji is launching a brand-new premium wafer SKU, Q-commerce-first, targeted at urban Gen-Z and millennials. Zepto, Blinkit, Instamart. Premium positioning. 60 minutes. AI does the work. Mixed teams: R&D with marketing, finance with sales, manufacturing with brand. Each team presents a complete launch package at the end.

01 · CONSUMER

Buyer persona built with NotebookLM on competitor Q-comm reviews. Who buys this SKU. Why. How often.

02 · BRAND

Name, tagline, pack design. 6 variants generated across 3 image tools. Shortlisted to one.

03 · UNIT ECONOMICS

MRP, trade margin, Q-comm listing fee, P&L per pack. Live modelled in Excel with Claude.

04 · CREATIVE

15-second launch reel. Script, voiceover in Hindi + Gujarati, video, end-to-end with AI. Watchable.

05 · DEMAND

Day-1 forecast across 5 metros. Re-order cycle. Q-comm inventory plan by zone.

06 · FOUNDER PITCH

3-minute pitch to the Virani family. No slides. Just AI artefacts and clarity. Family votes the winning team.

60 min

Total run-time

6 teams

15-17 people each

6 outputs

Full GTM per team

1 winner

Voted by family

Why it works. The simulation forces every dept to use AI on a real Balaji decision, with real time pressure, in front of the family. Awareness, capability, and team-bonding land in the same hour. The winning team's deck often becomes the actual launch deck six weeks later.

Team Formation Rules

›Every team has 1 leader from each: R&D, Marketing, Finance, Supply Chain
›1 promoter family member per team. Not as leader, as participant
›No two team members from the same dept (forces cross-functional thinking)
›Teams are pre-assigned, posted at lunch on Day 2. No swapping
›Each team has 1 CF practitioner observing, not participating

How Family Picks the Winner

›Clarity of the founder pitch (would you green-light this?)
›Quality of unit economics (does the P&L hold up?)
›Creative landing (is the pack & reel something Balaji would ship?)
›Forecast realism (do the numbers make distribution sense?)
›Hard rule: No slides. Only AI-generated artefacts allowed in the pitch

Balaji AI Training · Creativefuel 05 / 09

Section 06 · Investment

Three phases.
Exit-able at every stage.

You commit to Phase 1 only. Phase 2 unlocks if the diagnostic surfaces opportunity. Phase 3 unlocks only if the workshop ships prototypes worth productionising. No hidden funnel. No 12-month lock-in.

Phase What it is Duration Investment

01

The Diagnostic Half-day with each of 6 dept leads. Sanitised data review. We arrive at workshop with your numbers, not generic FMCG examples. 20-page brief on your desk.

2 weeks

₹ 3,00,000

02

The 2-Day Training On-site at Rajkot. 100 attendees. 6 dept labs. 6 working prototypes shipped. 30-day office hours. Workshop recordings. Internal Q&A channel for 6 months.

2 days

₹ 12,00,000

03

The Transformation (Optional) 6 months. 4 to 6 prototypes from Phase 2 productionised. CF engineers embedded. Internal Balaji AI cell of 3 to 5 staff trained up. Scoped only after Phase 2 ends.

24 weeks

₹ 75L - 1.25Cr

Σ

Year-1 Total (All Phases) GST extra. Travel + stay at actuals, Mumbai to Rajkot. Each phase invoiced separately. Each phase fully cancellable at boundary.

~7 months

₹ 90L - 1.40 Cr

1%

Yield improvement = ~₹50 Cr / year

3%

Procurement saving = ~₹54 Cr / year

60-70%

Dealer-bot query deflection = hours back daily

1 Qtr

Typical payback period on Phase 2 alone

How to read this. A single-trainer 1-day workshop costs ₹5L in the market today. This proposal is 2.4 to 3x that, but ships 6 working prototypes, runs on your data, brings 4 to 5 practitioners, and includes a transformation pathway. The Year-1 number is roughly what one mid-senior data hire costs Balaji over the same period, with substantially more delivered.

For context: what ₹12L gets you, compared

What you get	Single trainer (~₹5L)	This (₹12L)	Big-4 consulting (~₹50L+)
Trainers in the room	1	4 to 5 practitioners	2-3 consultants
Days on-site	1	2	5+
Working prototypes shipped	0	6, on Balaji files	1-2 mockups
Built on Balaji data	No, generic examples	100%, under NDA	Sample only
Post-workshop support	None	30 days + 6mo Q&A	Paid retainer
Optional transformation path	Not offered	₹75L-1.25Cr, opt-in	₹3-5Cr engagement

Balaji AI Training · Creativefuel 06 / 09

Section 07 · Phase 3 Transformation (Optional)

If Phase 2 lands well,
here is what comes next.

Phase 3 is offered upfront for transparency. It is not signed at Phase 2. We scope it at the end of the workshop, with the family in the room, after you have seen the 6 prototypes and chosen which ones deserve production budget. 4 to 6 pilots typical. Each ships in 8 to 16 weeks.

Dealer WhatsApp + IVR Bot

LAB 03 · 8 WEEKS · BIGGEST QUICK WIN

Production-grade dealer support across 4,000+ distributors. Trilingual. Integrates with order systems. 60 to 70% query deflection target.

Fryer Line Vision QC

LAB 01 · 12 WEEKS · HIGHEST RUPEE IMPACT

Computer vision deployed on selected lines. Defect detection, reject-arm trigger. Targets 1 to 2% yield improvement.

Distributor Health Platform

LAB 04 · 10 WEEKS · HIGH-VISIBILITY

Live dashboard for every distributor manager. Anomaly alerts. Reorder predictions. Trade scheme suggestions. Phone-first UI.

Vendor + Sourcing Intelligence

LAB 02 · 10 WEEKS · CFO-FAVOURED

Vendor scoring goes live. Potato price model integrated to procurement workflow. Anomaly alerts on inbound invoices.

Q-Comm Listing Factory

LAB 04 · 6 WEEKS · FASTEST PAYBACK

Every SKU on every Q-comm platform optimised continuously. A/B testing on listings. Regional language variants per zone.

Plant SOP Knowledge Layer

LAB 01 · 8 WEEKS · OPERATOR-LOVED

Every SOP, every shift handover, every line note ingested. Operators query in Gujarati. 50 years of institutional memory becomes searchable.

4-6

Pilots productionised

24 wks

Total transformation

3-5

Balaji staff trained into AI cell

1 yr

Post-handover quarterly review

Why this matters. Workshops alone have a 90-day shelf life. By Day 91, the printed workbook is in a drawer. Phase 3 converts the prototypes from Day 2 into production tools that run quietly inside Balaji for the next decade. Phase 3 is fully optional and only scoped if the workshop earns it.

How We Sequence Pilots Across 24 Weeks

WAVE 01 · WK 01-08

Visible Quick Wins

Dealer Bot + Q-Comm Listing Factory. Both ship fast, both demo well, both deliver measurable savings within the first 60 days. Builds family confidence in the programme.

WAVE 02 · WK 09-16

Compound Wins

Vendor Intelligence + Plant SOP Layer. These touch core operations. Slower to land but the savings compound quarter over quarter, with no marginal ongoing cost.

WAVE 03 · WK 17-24

Defensible Moats

Fryer Vision + Distributor Platform. Hardest to build, biggest moat. By now the internal AI cell is trained up and Balaji owns the long-term capability.

Balaji AI Training · Creativefuel 07 / 09

Section 08 · About the Partner

We don't sell training.
We run a 300-person AI-native company.

Most AI vendors will pitch you with a story. Here are the numbers. Creativefuel is a Mumbai-headquartered AI-native operating company. 300 employees use AI 8 hours a day. We don't teach what we read. We teach what we run.

300+

Employees using AI 8 hrs/day across creative, ops, sales, finance, HR. The same 17 functions Balaji has.

30+

Models routed through our internal LLM gateway. Claude, GPT, Gemini, Llama, Qwen. By use case.

20M+

Business WhatsApp messages indexed and queryable across our partner brand intelligence pipeline.

8M+

Documents embedded for semantic search across briefs, contracts, archives.

660GB

On-premise AI fleet. Llama 3.3 70B, Qwen 2.5 72B, Whisper, image gen. Not a reseller of APIs.

30M+

Aggregate IG follower base across CF-owned brands: Sarcasm, Miss Malini, Daily Over Dose.

What we've built and operate. A unified chat surface. Image generation across 4 model families. Deep research agents. Voice synthesis in Indian languages. OCR pipelines. Retail audit photo intelligence. Real-time ops dashboards. Predictive analytics. Classification systems. Multi-language WhatsApp support bots on production creator networks. All in-house. All running today.

Why this matters for Balaji. The dealer WhatsApp bot? We've built that for our own creator network. The retail audit photo AI? We run vision models on brand grids weekly. Demand forecasting? Our brand intelligence engine does that across millions of signals. We are not coming to Balaji with a curriculum. We are coming with the same playbook we run on ourselves.

A Few AI Systems We Operate Today

Internal LLM Gateway

Single chat surface routing 30+ models. 4 image-gen providers. Deep research, voice synth, code interpreter. Used by every CF team daily.

WhatsApp Intelligence Pipeline

20M+ business messages indexed and queryable. Group classification, sentiment, action extraction. Live for our partner brand operations.

Real-Time Ops Dashboards

Live across 7 brand verticals. AI-summarised commentary, anomaly alerts, predictive forecasts. Bloomberg-style ops console for content teams.

On-Prem AI Compute Fleet

660 GB across Mac Studios + minis. Llama 3.3 70B, Qwen 2.5 72B, Whisper, image-gen, embeddings. No cloud dependency for sensitive workloads.

Brand Intelligence Engine

Vision models on millions of social posts daily. Trend detection, competitor tracking, hashtag whitespace analysis. Powers our creator strategy work.

Multilingual Voice + OCR Stack

Hindi, Gujarati, Marathi, Tamil, Telugu voice + OCR pipelines. Production-grade. Same stack will power Balaji's dealer support bot.

Balaji AI Training · Creativefuel 08 / 09

"Workshops promise transformation.
We promise prototypes."

Six working AI prototypes, on Balaji's actual files, in two days at Rajkot. One full-day live simulation that ends with a launch pitch to the family. Thirty days of office hours after, six months of Q&A. If Phase 2 lands well, Phase 3 turns six prototypes into six production systems. If it doesn't, we shake hands and part with no obligation.

The Ask

A half-day on-site at Rajkot with three frontline leads, Production, Supply Chain, and IT Support, before any commitment. We arrive with the diagnostic framework. You arrive with what is actually broken. We leave with a draft Phase 1 brief, on Balaji letterhead, that the family can review in private.

If we go forward: the next 30 days

D 00-02

NDA signed. Diagnostic kickoff call.

D 03-14

Half-day with 6 dept leads. Data shared.

D 15-21

Workshop brief delivered. Family review.

D 22-30

Workshop dates locked. Pre-work begins.

Three things this engagement is explicitly NOT

Not a lecture.

No slide-by-slide AI awareness session. Every hour of Day 1 has a hands-on exercise. Day 2 is a build, not a tutorial.

Not a sales funnel.

Phase 2 is standalone-priced. Phase 3 is optional, scoped after, exit-able. No upsells embedded. No required follow-on.

Not generic.

Nothing in this proposal is template content. Every track, every exercise, every prototype is built around Balaji's actual 17 departments and operating reality.

Creativefuel

AI-Native Operating Company · Mumbai

creativefuel.io
300 employees · 30+ models in production · 660 GB AI compute fleet

Balaji AI Training · Creativefuel · For the Virani Family · Confidential 09 / 09

What we'llbuild at Rajkot.