Applied AI In Action.
I don’t just advise on AI strategy — my body of work is built on executing it. Each system below was designed, architected, and deployed in production. Together they represent a complete AI intelligence stack: understanding people, making institutional knowledge queryable, surfacing operational data through natural language, and executing autonomously on what it finds.
GoGig | Talent Intelligence
The AI Smart Assessment (AiSA)
Built and managed the cross-functional team of data scientists and engineers to scale one of the first personality-analysis SaaS tools on the AWS Marketplace. Grounded in computational psychometrics and the Big Five Personality Model, the system uses the LIWC-22 framework to evaluate 50+ behavioral traits through real-world communication patterns — delivering ethical, frictionless, AI-driven talent intelligence at scale. Successfully integrated RAG architectures and NLP into production workflows on AWS.
This is for you if
you’re rolling out a new system, strategy, or organizational change and you’re not sure your people are actually ready — or you’ve been through a failed implementation and suspect culture and change readiness were the real culprits.
Ultra Consultants | RAG Architecture
Ultra Knowledge Base — Enterprise AI Knowledge Pipeline
Designed and deployed a production-grade SharePoint-to-Azure AI ingestion pipeline that continuously synchronizes Ultra Consultants’ entire SharePoint tenant (1,061 sites) into three purpose-built Azure AI Search indexes. Built on a queue-based fan-out architecture with Azure Functions and Event Grid, the system extracts and chunks content from six file formats, generates 3,072-dimension vector embeddings using Azure OpenAI text-embedding-3-large, and applies a custom classification engine to route content intelligently — client engagement content, internal knowledge, and operational workforce data — each with a schema purpose-built for its query patterns.
Connected to the Ultra Knowledge Agent in Microsoft Copilot Studio (deployed to Teams), enabling natural language queries against Ultra’s full institutional knowledge base with real-time, citation-backed responses. Agent evaluation confirms ~80–96% pass rate across 50+ test queries.
This is for you if
your organization’s best knowledge lives in the heads of your longest-tenured people, scattered across folders nobody can find, or walks out the door every time someone leaves. You know the answers exist somewhere — you just can’t get to them fast enough.
8 Platforms | 150+ Processes
ERP AI Intelligence System
Developed and architected a proprietary knowledge engine enabling Ultra Consultants to deliver evidence-based ERP AI capability assessments across 8 platforms and 150+ business process friction points. Built on Azure AI Search, the system combines a structured capability index (1,690 SCOR/APQC-classified assessments), a semantic document library, and an automated ingestion pipeline that extracts and scores ERP AI features from vendor research using a multi-factor maturity model — Implementation Pattern × Feature Coverage × Evidence Quality.
Deployed a Microsoft Copilot Studio agent with specialized child agents surfacing real-time, citation-backed insights through natural language queries — positioning Ultra as the definitive AI × ERP advisory practice.
This is for you if
you’re evaluating enterprise software, trying to build an honest picture of what AI actually does inside these platforms, or need to walk into a vendor conversation with evidence instead of the vendor’s own marketing deck.
Ultra Consultants | Operational Intelligence
Projector: Workforce Intelligence
Integrated Ultra’s Ginsu PSA utilization data — 250,981 weekly records spanning 101 consultants and four years of actuals plus forward projections — into a dedicated Azure AI Search operational intelligence index with a purpose-built schema for workforce queries. Built a custom Python indexer that parses the Ginsu pivot export, normalizes seven utilization metrics per consultant per week, computes utilization percentages, generates natural language narratives for semantic search, and loads the full dataset with 3,072-dimension vector embeddings.
Connected to the Ultra Knowledge agent alongside the SharePoint knowledge base, enabling a consultant to ask “who has available capacity for a client engagement next month?” and receive a data-backed answer drawn from live PSA data — the first time operational workforce intelligence and institutional project knowledge have been queryable from a single interface.
This is for you if
you run a professional services organization and your leaders are still answering “who’s available?” with a spreadsheet, a gut feeling, or a Slack message. You have the data — it’s just not talking to your knowledge.
Ultra Consultants | Consultant Support
Consultant Support Agent
Accepts transcripts, session notes, Excel/CSV exports, and Visio process diagrams as inputs. Visio files are converted to images and parsed alongside current-state process flows. The agent extracts friction points — manual steps, system gaps, redundant handoffs — and stores them against a client-specific knowledge base scoped by engagement ID.
Identified friction points are scored by business value and implementation complexity. The agent maps each to a recommended AI agent and generates three structured outputs: a Gold Box summarizing current-state findings in plain language; an Opportunity Matrix plotting each friction point on a value-versus-complexity grid to support prioritization; and an Agent Recommendation list pairing each high-priority opportunity with a specific agent and deployment rationale. Functions as an early warning system — tracking deliverable status and project health to surface risk before engagements trend off track.
This is for you if
you manage engagements where deliverable status, project health, and early risk signals are spread across emails, spreadsheets, and meeting notes — and by the time something looks off track, it already is.
Freedom3 Capital | Private Equity AI
Analyst Agent — Financial Modeling & KPI Analysis
Designed a document-centric Copilot agent that reads internal financial models and spreadsheets to extract core PE metrics — EBITDA, revenue growth, leverage, ROI, and margins — and supports scenario analysis across base, upside, and downside cases. The agent flags trends and outliers for follow-up, applying the same logic and framing to every file it processes.
Built to eliminate the variation that accumulates when analysts rebuild the same views differently across deals. Same inputs produce the same outputs, every time — reducing manual rework in Excel and improving consistency across the deal team without disrupting existing workflows.
This is for you if
your analysts are rebuilding the same financial views for every deal, slightly differently each time, and your IC prep quality depends more on who ran the model than on what the model says.
Freedom3 Capital | Private Equity AI
Researcher Agent — Market & Competitive Intelligence
Built a role-based research agent that monitors industry trends, competitor activity, and valuation benchmarks — synthesizing external data into concise summaries scoped to active and prospective deals. Designed to give IC and deal teams faster market context without manual research cycles or context-switching between tools and sources.
Deployed as part of a deliberate sequencing strategy — education before automation — shaped by Freedom3’s mixed trust in AI across the team. The agent was scoped narrowly and kept explainable: document-centric, easy to test, and easy to discard if it didn’t fit.
This is for you if
your deal team is doing repetitive market scans from scratch on every new opportunity, and the quality of early diligence context depends on how much time an analyst had that week.
Freedom3 Capital | Private Equity AI
Investor Update Generator — LP & Board Communications
Designed an agent that reads internal financials, risk notes, and supporting documents to draft investor updates, board summaries, and slide-ready content — maintaining consistent tone and structure across every communication. Delivered as part of a three-agent pattern (Analyst → Researcher → Communicator) that mirrors how PE teams already think about their workflow.
Introduced through an onsite Copilot education workshop in Kansas City — hands-on, using real Freedom3 documents — before any agent automation was deployed. The goal was to build comfort first and avoid wasted investment if adoption stalled. The session shifted internal conversations toward using AI earlier in diligence and offering it as part of PE value creation, not IT tooling.
This is for you if
your team is rewriting the same LP and board updates from scratch each cycle, and the consistency of your investor communications depends on who had bandwidth to polish the draft.
Zero-Trust Executive AI
Pansophy: Autonomous Personal AI
The Capabilities: Architected and deployed “Pansophy,” a highly capable, autonomous personal AI agent that serves as a digital chief of staff. Integrated directly with a proprietary “Personal Brain” via Azure AI Search, the agent seamlessly retrieves and synthesizes decades of methodologies, articles, and private operational data. Capable of executing multi-step workflows including Azure infrastructure deployment, CI/CD pipeline management, DNS configuration, and automated API orchestration.
The Security Posture: Designed with a strict Zero-Trust architecture to securely manage private operational data. Operates exclusively over a private Tailscale VPN tunnel with zero public-facing inbound ports. The system is fundamentally isolated from public LLM training datasets, ensuring proprietary methodologies, credentials, and client data remain completely sovereign.
This is for you if
you’re a founder, executive, or operator drowning in context-switching, who needs a thinking partner that actually knows your business — and won’t tolerate your proprietary data sitting on someone else’s servers.
