Client Work Research Data Intelligence
Client Case Study

Intelligence at
Speed

Under 2 hours from a single brief to a fully structured, outreach-ready contact database. What happens when you deploy a multi-agent AI team against a real market intelligence problem.

May 2026 Client Engagement 8 min read
46
Named Individuals
70
Companies Scraped
27
Fully Categorised
6
Data Dimensions

The Brief

A client came to us with a specific problem. They were building a UK community portal for drone professionals in the construction, surveying, engineering, and infrastructure sectors. The portal needed members — specifically, professionals with UAV expertise at major organisations who could bring authority, experience, and credibility to the community.

The brief was straightforward but ambitious: find the right people, understand who they are and why they matter, and deliver something the client could act on immediately. Not a list of Google hits. Not a CSV of names scraped from a directory. A structured, categorised, outreach-ready contact database with enough context to make every conversation count.

The client had no time for a multi-week research project. They needed intelligence, fast. We said we could deliver in under 2 hours — and we did.

"Find me the people who matter in UK construction drones — and tell me why they matter."

— The brief, in a single sentence

What We Did

We deployed a three-agent AI team working in parallel, each handling a distinct part of the intelligence pipeline. The work happened simultaneously, not sequentially — that's how we hit the speed target.

Parallel Execution, Not Sequential Discovery

Traditional research flows one step at a time: identify sources → scrape data → clean it → analyse → format. That's fine if you have a week. If you have two hours, parallel is the only option.

Our coordinator agent (Bee) broke the brief into three independent workstreams:

Because each stream was independent, they ran simultaneously. Scraping didn't wait for the taxonomy. The taxonomy didn't wait for a completed dataset. The coordinator stitched them together at the end.

The Taxonomy: Six Dimensions of Fit

Every contact was scored across six dimensions, producing a structured assessment of their relevance to the client's community:

This framework wasn't generic. It was designed specifically for the client's community model. Each dimension answered a practical question: who are they, what do they do, where are they based, and how well do they fit the community's purpose?

What Was Produced

The client received five deliverables, all produced within the two-hour window:

1. Master Contact Database

A structured spreadsheet with 46 named individuals across five tiers — from senior infrastructure drone leads at organisations like National Highways and Network Rail, through to independent surveying companies operating in every UK region. Each record included name, job title, company, website, LinkedIn URL, region, and classification across all six taxonomy dimensions.

2. Company Directory

70 UK drone companies with full contact details — websites, regions, phone numbers where publicly listed — systematically scraped from structured directory sources. Covering the full spectrum from national infrastructure organisations to regional independent operators.

3. Categorisation Framework

The 6-dimension taxonomy document, designed to be reusable. As the client adds more contacts in Phase 2, they can apply the same classification system — no need to redesign the framework each time the database grows.

4. Outreach Playbook

Five persona-specific message templates — each tailored to a different type of community member. In-house UAV leads get a different message than independent surveyors. Engineering consultants get a different message than technology vendors. Every template includes the core value proposition, tone guidance, and suggested channel (email vs LinkedIn).

5. Data Schema & Deduplication Rules

A complete data dictionary documenting every column in the contact database, plus a deduplication protocol for merging records when new data sources are added. Phase 2 can scale without breaking what Phase 1 built.

"Here's what we produced in under 2 hours: 46 named individuals, 70 companies, a 6-dimension taxonomy, and a playbook you can use today."

— The client-facing summary

Methodology

The Data Pipeline

Six structured data sources were systematically mapped and explored:

Each source provided a different lens on the market. Trade associations gave quality-filtered contacts. Structured directories gave consistent, comparable data. Professional bodies added authority and trust signals. Combined, they produced coverage that no single source could deliver.

The Timeline

Methodology

How It Happened

Phase 1

Brief received and decomposed into independent workstreams by the coordinator agent. Directory sources identified and mapped.

Phase 2

Parallel execution: directory scraping (Gav), taxonomy design (Mel), and quality oversight (Bee) run simultaneously across six data sources.

Phase 3

Data assembly: scraped contacts matched against the taxonomy. Fit scores assigned. Gap analysis identifies Phase 2 targets.

Phase 4

Delivery: outreach playbook drafted, schema documented, deduplication rules applied. Client receives five structured deliverables, ready to use.

The Team

This wasn't one AI doing everything. It was a coordinated team of three specialised agents, each working in its domain of expertise, with a human in the loop for direction and quality sign-off.

🐝

Bee

Coordinator

Decomposed the brief, ran individual research on key contacts, cross-referenced data against the framework, and handled quality verification.

🔍

Gav

Research

Systematically scraped six structured directories, extracted company and contact details, and mapped UK regional coverage across 70 companies.

🧩

Mel

Framework Design

Designed the 6-dimension categorisation taxonomy, created persona profiles for the outreach playbook, and drafted the message templates.

The human stayed in the loop for strategic decisions — approving the taxonomy structure, reviewing fit scores on tier-1 contacts, and signing off on the outreach messaging. The agents did the heavy lifting; the human directed the work and validated the output.

What This Means For You

This case study matters because it bridges from "what we do internally" to "what we can do for clients." It's not a demo or a proof of concept — it's real deliverables produced for a real client on a real deadline.

If your business needs market intelligence, contact databases, competitive research, or data enrichment, here's what this project demonstrates:

1

Speed Without Sacrifice

Under 2 hours from brief to delivery. The AI team researched across six data sources, categorised 27 contacts, and produced five structured deliverables — all while the client was still in their morning. Speed doesn't mean shallow; the taxonomy, playbook, and deduplication rules were designed for reuse.

2

Multi-Agent Systems Beat Single Agents

Three specialised agents working in parallel solved the brief faster and more thoroughly than any single model could. One scraped, one designed the framework, one verified quality. The coordinator stitched the outputs together. This is how AI works at scale — not one tool doing everything, but a team each doing what it does best.

3

The Output Is Reusable Infrastructure

The taxonomy, schema, and playbook aren't one-off artefacts — they're infrastructure. Phase 2 can scale to 300+ additional contacts without rebuilding the framework. The client didn't just get data; they got a system for ongoing intelligence gathering.

How We Can Do This For You

This project sits at the intersection of data intelligence and AI-augmented research — capabilities that apply across industries, not just construction and drones.

If you need market intelligence, competitor analysis, contact databases, or structured research from scattered public sources, the same multi-agent pipeline can be configured for your sector. The taxonomy changes. The sources change. The methodology scales.

Want to see what we'd find in your sector? Start a conversation — brief us on your market, and we'll show you what our team can uncover.

Back to Case Studies

Home Services Proof Insights