Call Summary: Jollibee Brand Operating System Technical Briefing

Overview

Wilson Fong provided Gary Sheng with a comprehensive briefing on the Jollibee Foods Corporation brand compliance project. The call covered the project's origin, current scope, technical requirements, and Gary's role as technical lead for building an AI-powered brand compliance portal. Wilson is preparing Gary for an upcoming vendor evaluation call with Jollibee's head of loyalty, head of AI, and head of data analytics.

Key Topics Discussed

Project Origin & History

• Coffee Bean and Tea Leaf Entry Point: Jollibee originally put out an RFP for global brand identity work. Contino (Wilson's agency) submitted a proposal but lost to a larger global advertising agency. However, the proposal was circulated internally and impressed stakeholders.
• Coffee Bean Request: Coffee Bean and Tea Leaf (one of Jollibee's 19 brands) reached out: "We loved what you all did. The way you represented Jollibee's and the fun in the proposal... was really a breath of fresh air for us."
• Brand Work Completed: Contino developed new font, hand-drawn elements, color palette, illustration style, and photography guidelines. "They fucking loved it."
• Trust Built Over Time: "We started out with them four months ago, and in four months they've given us three SOWs... we started with a really small 'Hey, we've got a small budget.' And literally, once we build trust, they were like, here's another 60 grand. Do this. Here's another 80 grand. Do this."

The Core Problem: Brand Inconsistency

• Scale of Issue: Jollibee has 5,000 franchisees across 60+ countries. Coffee Bean and Tea Leaf alone spans 60 countries.
• Visual Chaos: Wilson showed examples of how different franchisees interpret the brand differently. "Look at how the advertising ranges from one place to the next... It's like completely different companies."
• Contrast with Competition: "You go to Starbucks in India and you go to Starbucks in Singapore and you go to Starbucks in the United States. There is a consistency and there is a rigidity that doesn't change. And they're struggling."
• Current Process Failures: Brand standards are shared via "hundreds and hundreds of pages through PDFs." Franchisees hire junior local designers ($25-50/hour) with no context or training. Questions go to Manila customer service, creating inconsistency. "People don't read shit."

Technical Solution: Brand Operating System (BOS)

• Core Concept: AI-powered chat interface that replaces PDF hunting with fast answers on brand compliance.

• Phase 1 Deliverables:

  • Complete brand foundation and standards documentation
  • Simple chat interface for fast answers (similar to Gary's Lumina work)
  • Asset validation built into portal (upload assets, check compliance, return pass/fail with specific feedback)
  • Testing with 10-15 real franchisees
  • Road map to full enterprise platform

• Pilot Brands: Chow King (Filipino-Chinese fast food) or Mang Inasal (Filipino grill) first, then potentially Jollibee mothership brand.

• A Primo Integration:

  • A Primo is their existing digital asset management system ("classic fucking look at this software. It's lame as fuck")
  • REST API available for integration
  • Gary flagged potential deprecation notice but confirmed core APIs still functional
  • Integration would allow franchisees to pull approved assets directly

Technical Requirements

• Enterprise Grade: HIPAA compliant, GDPR compliant portal required
• Compliance Rules Engine: Gary proposed using multiple AI agents/sub-agents focused on different rule categories rather than checking all 50+ rules at once
• Asset Validation: System checks uploaded assets against brand guidelines, returns pass/fail with specific feedback
• Data Ingestion: Will receive 15 stakeholder interview recordings/transcripts plus historical customer service questions (email, phone, chat) for training

Success Metrics (Phase 1)

• Adoption: 75% of pilot users log in and start using it regularly
• Accuracy: 90% of questions get correct answers from the standards
• Speed: Finding brand information drops from hours to under 5 minutes
• Outcome: "Franchisees stop calling Manila for simple brand questions. PDF hunting dropped to near zero."

Business Model & Long-Term Vision

• Repeatable System: "If we do this correctly, we may be one of the first people that is creating a compliance template, and we can take this to all their competitors and all the different global brands."
• Target Market: "Franchising is huge. In the hospitality and service in the food business... Dunkin Donuts of the world, Kentucky Fried Chicken, all of these people."
• Scale Economics: "Once we've done it once, it could be potentially boring work, but this is work that just gets you paid... Each one of these integrations is just fucking profit. The more we do it, the more profitable we will become."
• Pipeline: If successful with pilot brands, 17+ more brands behind them. "That might be a year's project."

Gary's Role & Compensation

• Position: Technical lead for AI/portal implementation
• Wilson's Framing: "What I want to do with you and I is establish this workflow so that it's consistent for us for the next 12 months. But it's not going to take your full time, and it'll be like, oh, this paycheck will just take care of my living and anything I lay on top will basically be bonus."
• Expectation: "When you come into this call with them, Gary, you have to be super leadership and chilled like you've done this a million times before. But you're still challenged with the Jollibee system."

User Lifecycle & Query Patterns

• New Franchisees: Lots of initial onboarding questions
• Established Franchisees: Seasonal campaign support for "tent poles" (New Year's Eve, Valentine's Day, Mother's Day, Father's Day, Thanksgiving, Christmas)
• Pattern Recognition: "You're gonna know the patterns, dude... and these patterns, Gary, they can apply to any fucking food business."

Asset Generation Discussion

Gary raised the potential for AI-generated assets as a future extension:

• Gary's Insight: "Right now, what it feels like is there'll be a lot of franchisees getting rejected for their designs. But then it's like... that's totally where AI is going."
• Wilson's Clarification: Franchisees aren't creating designs from scratch. They pull approved product shots from A Primo but may misuse them (wrong colors, logo placement, etc.). Phase 1 focuses on stopping the worst 80% of infringements.

David Thompson / Signal Pattern Update

• David (Wilson's other potential partner on this project) declined to participate: "They really needed governance templates from you guys. There's no AI in there that I could see."
• Wilson disagrees with David's assessment but prefers working with Gary anyway
• David is focused on Signal Pattern, his AI-powered strategy platform with "36+ specialized AI agents"

Partnership Philosophy

Wilson emphasized the collaborative approach:

• Communication: "This is a practice for you and me in excellent communication... What you notice and you're unsure of, you and I can confer."
• Trust Building: "Underpromising and overdelivering. Dude, I'm gonna set you up so this path, if we do this right, is gonna fucking make us excited about working together."
• Brownie Points Strategy: If Gary discovers fixable issues (like A Primo metadata problems), they can fix them free to build goodwill.

Key Quotes

• "I'm going to prepare you to be a slow, confident technical leader."
• "This is, like, the perfect use case for AI because it's such beneath human work to look at every single rule and just be like..."
• "Dude, if we do this right, we work directly with the fucking owners. There's like 25,000 employees at Jollibee's."
• "As long as we have trust in each other, and we give each other space to make mistakes, and communicate when we do make a mistake. You and I are gonna fucking kill it, man."
• "I'm excited to hopefully fucking get in this boat and fucking rocket with you."

Key Stakeholders for Upcoming Call

• Sandeep: Primary contact, Bain Consulting background, 17 years at Unilever (Global Marketing Manager of ice creams), now handling compliance and governance at Jollibee
• Adam Diamond: Sandeep's superior
• New Introductions: Head of Loyalty, Head of AI, Head of Data Analytics (supporting technical vendor evaluation)

Action Items / Next Steps

1. Immediate: Gary to review Wilson's AI-generated preparation notes (texted) and help draft response to Sandeep's request for "use cases and more technical details regarding the BOS implementation"
2. Tonight: Send preliminary response to Sandeep
3. Before Friday: Schedule vendor evaluation call for Monday (not Friday due to Thanksgiving)
4. Gary's Prep: Research A Primo REST API capabilities, understand enterprise compliance requirements (HIPAA, GDPR)
5. Deliverable Focus: Demonstrate asset validation capability similar to Lumina work from previous day