Tim FinchProducts, systems, AI
HomeProjectsWritingWork with meAbout

Open to product, platform, and AI work.

I build evidence-backed AI products and the systems needed to operate them responsibly, usually where product decisions and real workflows meet.

GitHub ↗LinkedIn ↗Devpost ↗London / Remote

In this case study

  1. 01 Overview
  2. 02 Why it exists
  3. 03 How it works
  4. 04 Built with
  5. 05 Focus
  6. 06 Architecture
  7. 07 Impact
←Back to work
Enterprise data layerConfidential
Claude
MCP
Fabric
Natural-language questions
Reports / analysis / action
Microsoft Fabric AgentsSalesforce data

AI Products

Enterprise AI Data Platform

A conversational enterprise data platform connecting Claude, MCP, Microsoft Fabric Agents, and Salesforce data for reporting and analysis.

2025 / AI Products / Claude / MCP / Microsoft Fabric / Fabric Agents / Salesforce

Quick read

A clear breakdown of the product, the reasoning behind it, and the implementation choices that matter.

Year
2025
Category
AI Products
Status
Confidential professional case study

Overview

A confidential enterprise AI platform conceived, architected, and delivered to give colleagues a more direct way to ask questions of operational data. It combined Claude with an MCP data layer over Microsoft Fabric and Salesforce, alongside Fabric Agents, so teams could explore performance and generate reports in natural language.

Why it exists

Colleagues had to understand complex data systems or wait for analysts to answer everyday operational questions. Fixed queries created a reporting backlog and slowed decisions.

How it works

I conceived the platform, secured stakeholder buy-in, designed the architecture, and led delivery of a governed conversational data layer. Claude and Fabric Agents handled natural-language interaction while MCP connected the experience to trusted Fabric and Salesforce data.

Built with

Claude / MCP / Microsoft Fabric / Fabric Agents / Salesforce

What I focused on

  • Adopted across 8 business teams, with wider rollout under way and interest from executive users.
  • Replaced fixed SQL requests with self-service natural-language access to operational data.
  • Cleared the reporting backlog and freed analyst time for higher-value work.
  • Became a reusable enterprise AI capability that encouraged wider Claude adoption.

Architecture

Colleague question
        ↓
Claude conversational interface
        ↓
MCP data layer
        ↓
Microsoft Fabric Agents
        ↓
Fabric data + Salesforce data
        ↓
Grounded analysis, reports, and operational answers
The platform connected a familiar conversational surface to governed enterprise data and agent capabilities.

The important design decision was to make enterprise data easier to reach without making the underlying governance disappear. The conversational layer reduced friction for colleagues, while the data and agent systems remained explicit parts of the architecture.

Impact

  • 8 business teams adopted the platform, with wider rollout under way.
  • Natural-language access reduced dependence on fixed reports and analyst queues.
  • The capability created a reusable pattern for further enterprise AI adoption.

Where to look

GitHub: Confidential — repository not public