🔒 Internal Competitive Edge

Purpose

This section describes structural advantages Go Digit has over any Thai competitor that we use internally to deliver at higher speed and quality — but we do NOT market publicly because:

• They are commercial partner terms, not public claims
• Disclosure could breach partner agreements
• Mentioning them lowers perceived value (clients think "oh, you're using free credits")
• They are structural, not features — clients should hire us for OUR work, not a supplier's credit

💎 The Asset: $200K USD AWS Partner Credit (Annual)

The Numbers

• $100,000 USD/year — Go Digit as AWS Advanced Tier Services Partner
• $100,000 USD/year — Next Brain as separate AWS Partner entity
• Total: $200,000 USD/year ≈ ฿7,000,000/year in AWS credit
• Usable on AWS Bedrock — hosts Claude, Llama, Titan, Cohere, Mistral models
• Renewable annually (subject to partner program status)

What This Means Operationally

We can run Claude Opus, Claude Sonnet, and other premium models for:

• Prototype development during AIRA-X Diagnostics (Day 1-3 builds, not wireframes)
• Engagement delivery (Sprint, Custom Dev, Implementation)
• Internal R&D — testing new AI patterns without burning cash
• Sales POCs — rapid demos without client paying POC fees
• Next Brain product development — training, eval, continuous improvement

Structural Advantage

No other Thai software house has this. Not G-Able, MFEC, Bluebik, or any indie AI shop. Dual-entity structure (Go Digit + Next Brain as separate partners) doubles it.

What This Translates To

→ 5-10x quality/speed advantage on anything touching premium models — paid for by AWS, not us or clients.

🎯 How We Use It (Internal Playbook)

During Diagnostic (Days 1-3)

• Run Claude Opus against client's actual data (anonymized) during Day 2 analysis
• Produce working prototypes on Day 2-3 as part of readout — not mockups, actual running AI
• Why our Diagnostic feels "unusually deep" vs other 3-day assessments

During Sprint / Custom Dev

• Every prompt gets AI code review before commit
• Claude Opus used for architecture review, not just implementation
• Multiple model providers tested per use case (not locked to cheapest tier)

During AI Build & Run Retainer

• Monthly "model upgrade audit" — test newest models against client's existing prompts
• Cost optimization reports leverage our benchmark data from the credit pool
• POCs per quarter use premium models — competitors can't match quality at our price

🗣️ How to Talk About This

✅ Internal framing (team meetings, planning)

• "We have premium model access as a structural advantage — use it aggressively."
• "Don't cheap out on models during R&D — the credit exists for this."
• "This is why we can deliver Sprint in 6-8 weeks when competitors quote 12."

⚠️ External framing (client conversations)

• "We build with premium Claude models by default — we've optimized our stack for it."
• "Our development velocity is higher because we use best-in-class models during iteration."
• "Anthropic / AWS Bedrock are core to our build methodology."

❌ Never say

• "We have $200K of free AWS credits"
• "AWS pays for our development"
• "We get free Claude tokens from AWS"
• Anything implying our advantage is supplier-subsidized

Why: Clients who learn you're running on vendor credits will ask "why should I pay full price?" — and partner terms forbid disclosure anyway.

🛡️ Protection Rules

🧠 The Bigger Pattern — Multiple Structural Advantages

Go Digit has accumulated multiple small structural advantages that together form a moat:

1. $200K AWS Bedrock credit INTERNAL
2. AWS Advanced Tier Services Partner status PUBLIC
3. Cloudflare Partner status PUBLIC
4. n8n Authorized Reseller — only in Thailand PUBLIC
5. Humming Bird Consulting strategic partnership PUBLIC
6. Nomura Consulting partnership INTERNAL UNTIL Q3
7. Proprietary AIRA-X framework
8. Proprietary Next Brain Platform (silent deploy)

No single competitor in Thailand has more than 2-3 of these. We have all 8.

🔐 Credit Management Policy

Tracking

• Finance team tracks monthly Bedrock spend vs credit balance
• Monthly utilization report — shown in exec meetings only
• Forecast annual burn rate quarterly

Allocation Priority

1. Client engagements — always covered (never pay out-of-pocket for Bedrock)
2. Next Brain platform development — covered
3. Sales POCs — covered (max 2 per prospect)
4. Internal R&D — covered within quarterly R&D budget
5. Experimental / exploratory — CEO approval above $5K USD/month

Renewal Risk

• AWS partner tier reviewed annually
• Maintain partner requirements: certifications, case studies, customer references, delivery competency metrics
• If tier is downgraded, credit shrinks — treat as operational risk

⚠️ If This Leaks

If $200K credit information appears in external channels (web, social, press):

1. Do NOT confirm the number
2. Respond with: "We are AWS Advanced Tier Services Partner and build heavily on AWS Bedrock — specific partner terms are confidential"
3. Alert CEO immediately
4. Audit source: who had access, how did it propagate
5. If client-facing: pivot conversation to outcomes, not supplier terms

🎯 Rules of Thumb

1. Use the credit aggressively internally. Don't ration.
2. Never mention the specific dollar amount anywhere public.
3. When asked about AWS, answer with tier name ("Advanced Tier Services Partner") — never credit amount.
4. Treat this page as internal confidential — shredder-grade if printed.
5. Revisit annually on partner tier renewal.

Final note: This credit is one lever. Your talent, framework, platform IP, and partner ecosystem are the real moat. Treat this as one piece of a multi-layer advantage — not the whole thing.