Applied AI · Selected workJacksonville, Florida

From enterprise data
to useful AI.

I develop Copilot Studio agents, build reusable skills, and connect them to the data behind business decisions.

My work brings together AI implementation, finance, and business intelligence—from the first question to a repeatable workflow.

Explore the work Read the portfolio PDF

Copilot StudioAgent skillsMicrosoft FabricClaude & MCPEnterprise automation

Selected work

Built around a business question.

01Deployed implementationEnterprise AI / Market analysis

Find the opportunity
outside your records.

Internal financial data shows where a business already has customers. Understanding the market beyond those relationships requires an external view.

I developed and deployed a Copilot Studio agent that combines general ledger data through Microsoft Fabric semantic models with industry intelligence from Dun & Bradstreet’s D+ MCP.

Reusable skills handle variance analysis, reconciliation, charts and waterfalls, company presentation standards, and executive voice. The framework carries the analysis into other industries with minimal changes.

Read the implementation case PDF · 2 pages

The workflow

Internal context

GL data through
Fabric semantic models

External context

Industry intelligence
through D+ MCP

Copilot Studio agentA reusable skills layer

Analysis · Reconciliation · Charts
Brand standards · Executive voice

Whitespace & saturation analysisOne framework, reusable across industries

Implementation overview. No employer data is displayed.

Illustrative companion

Petroleum: the analysis in context.

Real companies, synthetic commercial figures. A worked example of market coverage, account opportunity, and prioritization.

$5.41Mmodeled whitespace
in the synthetic example

Explore the sample PDF · 6 pages

02Interactive demonstrationFinance / Business intelligence

Revenue reconciliationReview required

$14K difference in Unit B

USD thousands · Synthetic data

Trace the difference to the partner channel before releasing the report.

Read the result.
Find the difference.

A hands-on illustration of the reporting experience: clear numbers, visible exceptions, and an actionable next step.

Try four scenarios—mismatched totals, a clean reconciliation, missing data, and denied access—then inspect the evidence and follow the five-step workflow.

Try the interactive demo

Synthetic data and simulated controls. The demo runs in your browser; it does not call an AI model or connect to employer systems.

03Writing & methodBook in progress

The Skills Pack

Turning Finance Work into Tested AI Skills

A practical book about turning recurring work into reusable, tested AI skills—with explicit rules, bounded judgment, and evidence behind the output.

01 Computation: Make the calculations explicit.
02 Judgment: Define the model’s contribution.
03 House rules: Carry the business context forward.
04 Proof: Check the claims behind the result.

Read the excerpt PDF Book & companion files

Draft manuscript · In progress · Not yet published

The experience behind the work

Business fluency.
Hands-on implementation.

I’m Felix Reyes, a Principal, Financial Planning & Analysis, working at the intersection of enterprise AI, finance, and data.

My experience includes global consolidation of a multi-billion-dollar plan across five business units and multiple sales channels, a Redshift-to-Fabric data foundation feeding Power BI, and administration of a 200-user IBM Planning Analytics environment.

I work with Copilot Studio, Microsoft Copilot Managed Runtimes, production Copilot workflows, and Claude connected to enterprise finance data through MCP.

MBA, University of Florida · BA, Florida International University, magna cum laude

View my experience on LinkedIn

Get in touch

Let’s make useful things.

AI implementation, finance transformation, and business intelligence.

contact@felix.yt

From enterprise datato useful AI.