U.S. Small Business Administration Veteran-Owned Certified

SBA-Certified Veteran-Owned Small Business (VOSB) · GCC & Commercial Microsoft 365

Capability Statement

Puget Sound AI LLC · M365 AI & Automation Consulting · GCC and Commercial · SBA-Certified Veteran-Owned Small Business (VOSB)

Company Overview

Puget Sound AI LLC is an SBA-Certified Veteran-Owned Small Business (VOSB) specializing in AI and automation engineering for Microsoft 365, with deep production experience in both GCC (Government Community Cloud) and commercial M365 environments. The firm designs, engineers, and deploys production AI agent systems, Power Platform automation, and M365 governance frameworks for government agencies and commercial organizations. Government clients benefit from GCC-native architecture built around FedRAMP boundaries and CMMC alignment. Commercial clients on M365 Business Premium, E3, or E5 get the same engineering rigor without the government compliance overhead. Available as a VOSB subcontractor or teaming partner to federal primes on GCC engagements.

All work is led by senior engineering: a senior M365/GCC practitioner and U.S. Navy veteran with federal IT background (GS-09, U.S. Army Cyber Command / NETCOM) leads every engagement. Engagements deliver complete documentation, source code, and staff handoff so client teams can maintain all deliverables independently.

Core Competencies

AI Agent Engineering: Copilot Studio agents with generative orchestration, citation-bound retrieval-grounded architectures, RAG pipelines, and custom Model Context Protocol (MCP) servers for policy-compliant automation within the GCC boundary.
Power Platform Automation: Graph API-backed Power Automate workflows and Power Apps eliminating manual processes at scale with full audit trails; designed for government compliance requirements.
M365 Architecture & Governance: Tenant-wide governance across Exchange Online, Teams, SharePoint, Microsoft Purview, and Entra ID. Government GCC solutions align with FedRAMP-authorized boundary requirements and CMMC / NIST 800-171 control objectives. Commercial M365 solutions apply the same governance depth without government-specific compliance overhead.
AI Agent Infrastructure: Local SLM inference pipelines, vector embedding and retrieval systems, Azure AI Foundry integration, and reusable MCP server tooling for cloud-based government AI agents.
M365 Training & Adoption: Hands-on workshops for government and commercial IT staff and end users on Copilot Studio, Power Platform, and AI governance. GCC sessions cover boundary constraints and compliance requirements. Commercial sessions cover the same engineering depth with a wider connector catalog. Delivered by the senior engineers who build production systems using these tools.
AI Readiness Review: Structured, fixed-price gap analysis of GCC or commercial M365 environments against AI deployment readiness criteria, with a prioritized 90-day roadmap. Government reviews include FedRAMP, CMMC, and federal AI governance alignment. Commercial reviews focus on licensing, governance configuration, and automation opportunity mapping.

Differentiators

Production GCC engineering: Systems built inside the agency GCC tenant using only FedRAMP-authorized services, with full documentation, source code, and runbooks transferred at close.
Direct senior-engineering engagement: No account managers or subcontractors unknown to the client; the senior engineering that scopes your work builds and delivers it.
Environment-native design: GCC architectures are designed from the outset for government boundary constraints, not adapted from commercial templates. Commercial M365 architectures are designed for production governance and maintainability from day one.
Audit-ready, governed delivery: Every engagement ships architecture diagrams, documented workflows, source code, runbooks, and staff training for client-maintainable, independently operable deliverables.
SBA-Certified VOSB: Supports VA Vets First VOSB set-asides and primes’ veteran and small-business subcontracting goals.

Capabilities

Classes of system the senior engineering team builds for government GCC and commercial M365 clients.

M365 license reclamation through contextual inference

Python, PowerShell, and Graph API pipeline scanning a GCC tenant for misallocated license assignments on a scheduled cadence with AI-flagged anomaly review. Surfaces reclamation opportunities that rule-based logic cannot detect.

Statutory-retention document classification

Local SLM plus vector embedding pipeline classifying documents against agency retention schedules inside the GCC boundary, with per-document audit records and a records-officer review workflow. Converts multi-year manual backlogs into auditable, automated processes.

Legacy mailbox department inference

PowerShell and local SLM pipeline resolving unowned shared and legacy mailboxes to department owners at scale, with documented classification records. Unblocks data quality and compliance work stalled by missing ownership metadata.

Reusable MCP tooling layers

Python and PowerShell Model Context Protocol server providing compliance-grade Graph API and REST tooling to all Copilot Studio agents in a tenant. Every downstream agent accesses a shared, self-healing automation layer rather than duplicating API tooling.

Citation-bound policy and personnel agents

Copilot Studio agent delivering verbatim, SharePoint-grounded answers to HR and policy questions. No summarization, no paraphrase, no liability from re-interpreted policy text.

Natural-language-to-Graph and PowerShell administrative agents

Copilot Studio agent translating plain-language requests into validated Graph API and PowerShell operations, with Microsoft documentation grounding and runtime self-healing. Tier-3 administrative operations performed safely by non-scripters.

Plain-language rewriting agents

Copilot Studio agent grounded in an organization's own style standards, rewriting staff documents without paraphrase or hallucination, with before/after comparison output for review.