Ground Sourcing: How We’re Making AI Work for Us

It’s hard to imagine a productive office today that isn’t touched by AI. Whether drafting, analyzing, or brainstorming, AI has become part of the daily workflow—not as a novelty, but as a quiet, constant companion that helps teams move faster and think broader.

At Merchant Edge, we’ve gone through our own trial-and-error phase with AI. What we’ve found most valuable isn’t flashy automation or abstract “intelligence,” but something far simpler and more grounded. We call it ground sourcing.

Ground sourcing is the process of building an AI workspace around our own material. We create dedicated projects, upload internal documents—proposals, models, studies, reports—and let AI start connecting dots we might not have time to link ourselves. It summarizes, draws parallels, and surfaces insights that tie our work across sectors together.

We’ve all seen AI “hallucinations,” moments when a model confidently invents quotes, references, or entire studies. That’s exactly why we narrow the AI’s sources to our own documents. By keeping it grounded in our work, we reduce noise, avoid fabricated data, and get results that genuinely reflect our firm’s knowledge and experience.

It’s not about replacing people or producing content for its own sake. It’s about seeing our collective expertise reflected back in a structured, searchable, and surprisingly creative way. In many ways, it’s like training our own mini language model—one that speaks our language, understands our tone, and builds on our real-world experience.

The results have been quietly powerful. A proposal for a manufacturing project in the Levant now references benchmarks from a financial model we built for a completely different sector. A strategy deck on insurance transformation finds useful parallels with our work in education reform. Connections that might have stayed buried in folders or forgotten versions now appear in seconds.

For us, that’s the real value of AI in professional services: not just generating words, but revealing relationships that already exist in our work. Knowledge is rediscovered, contextualized, and made useful again.

We’re still experimenting, refining prompts, adjusting structures, and learning where human context matters most. But one thing is clear: firms that ground their AI in their own work will get the most out of it. It’s less about outsourcing intelligence and more about ground sourcing it.

And yes, I used AI to help me write this piece. It seemed only fair!

Abdul Kader El Haraki
Managing Director

‍