Microsoft Copilot powers AI across the entire Microsoft ecosystem — Word, Excel, PowerPoint, Teams, Outlook, and GitHub. But Copilot responds to vague prompts with vague output. This 6-band template gives Copilot the structure it needs to deliver precision.
Microsoft Copilot uses GPT-4 under the hood, enhanced with Microsoft Graph data (your emails, documents, calendar, Teams chats). This means Copilot has access to your organizational context — but it still needs you to specify what to do with that context. A vague prompt like "summarize my emails" gives you a vague summary. A structured prompt produces a targeted, actionable briefing.
The sinc-LLM 6-band decomposition applies perfectly to Copilot because the same specification dimensions matter regardless of the interface:
{
"formula": "x(t) = \u03a3 x(nT) \u00b7 sinc((t - nT) / T)",
"T": "specification-axis",
"fragments": [
{"n": 0, "t": "PERSONA", "x": "Executive assistant preparing materials for a C-suite quarterly business review"},
{"n": 1, "t": "CONTEXT", "x": "Q1 2026 board meeting preparation. Company: 500-person SaaS startup. Audience: 7 board members (3 investors, 2 independent, CEO, CFO). Meeting is Thursday at 2 PM."},
{"n": 2, "t": "DATA", "x": "Pull from: Q1 financial reports in SharePoint, sales pipeline in CRM export (attached Excel), customer churn data from last 3 monthly reports, and the competitive analysis document shared in the Strategy channel last week."},
{"n": 3, "t": "CONSTRAINTS", "x": "Must highlight metrics that changed more than 10% QoQ. Must include both positive and negative trends — no cherry-picking. Revenue numbers must match the CFO's official figures exactly. Each section must have a 'So What' interpretation. No section longer than 1 page. Use company brand colors (navy #1B365D, gold #C5A258). Must be ready for print (no interactive elements). Include appendix with data sources for every claim."},
{"n": 4, "t": "FORMAT", "x": "PowerPoint deck: Title slide, Executive Summary (1 slide), Financial Overview (2 slides), Sales Pipeline (1 slide), Customer Health (1 slide), Competitive Landscape (1 slide), Q2 Priorities (1 slide), Appendix."},
{"n": 5, "t": "TASK", "x": "Create the Q1 2026 board meeting presentation deck using the data sources specified above."}
]
}
GitHub Copilot operates in a code context, but the same 6-band structure applies. Instead of business context, you provide technical context. Instead of document format, you specify code format.
{
"formula": "x(t) = \u03a3 x(nT) \u00b7 sinc((t - nT) / T)",
"T": "specification-axis",
"fragments": [
{"n": 0, "t": "PERSONA", "x": "Senior backend engineer following clean architecture principles"},
{"n": 1, "t": "CONTEXT", "x": "FastAPI microservice for user authentication. PostgreSQL with SQLAlchemy ORM. Existing codebase uses repository pattern."},
{"n": 2, "t": "DATA", "x": "User model has: id (UUID), email (unique), hashed_password, created_at, last_login, is_active, role (enum: admin/user/viewer)"},
{"n": 3, "t": "CONSTRAINTS", "x": "Must use bcrypt for password hashing. JWT tokens with 15-min access, 7-day refresh. Rate limit: 5 login attempts per minute per IP. Must validate email format. Must log all auth events. No plaintext passwords anywhere. Type hints required on all functions. 100% test coverage for auth flows."},
{"n": 4, "t": "FORMAT", "x": "Python module with: router (FastAPI APIRouter), schemas (Pydantic models), service (business logic), repository (database queries), tests (pytest). One file per concern."},
{"n": 5, "t": "TASK", "x": "Implement the complete authentication module with login, register, refresh token, and logout endpoints."}
]
}