Portfolio

The Infrastructure Behind the Agent Team

Every tool here is part of the ecosystem that AI agents run on. This is the type of infrastructure we build for your business.

Flagship

Harrier's AI Agent Team

The proof of concept that powers everything Harrier does. A team of AI agents running a real business, every single day.

The Challenge

Running a consultancy means wearing every hat — marketing, sales, research, admin, proposals, content, reporting. There aren't enough hours in the day to do all of it well.

The Solution

Designed and deployed a team of named AI agents, each with a defined role, skill set, and cadence. They work overnight, producing content, research, proposals, and admin output. Every morning, 15 minutes of review.

The Results

  • The founder spends time on client work and strategy — not buried in admin
  • Overnight execution means every morning starts with work already done
  • The exact framework now deployed into client businesses
Built with: AI Agent Framework, Custom Orchestration Meet the full team →
Agent Infrastructure

Business Intelligence & CRM Dashboard

The command centre that agents report into — where all business data, pipeline, and agent output converges.

The Challenge

Projects, tasks, client relationships, and revenue were tracked across disconnected spreadsheets and tools. No single view of the business — and no way for AI agents to read from or write to a unified system.

The Solution

Built a unified dashboard combining project management, CRM, lead pipeline tracking, revenue analytics, and work session logging. This is where agents deliver their output and where the review queue lives.

The Results

  • One place to see everything — no more hunting across spreadsheets
  • Decisions backed by real-time data instead of gut feel
  • Time previously spent gathering data now spent acting on it
Built with: Next.js, SQLite, Drizzle ORM
Agent-Ready Process

Rapid Quotation & Estimation Tool

The kind of process agents now handle — standardised estimation that used to take days of manual work.

The Challenge

Preparing quotes and cost estimates took days of manual work — pulling rates from different sources, building up breakdowns, and formatting the final output. Every quote was built from scratch.

The Solution

Built a reusable estimation tool with standardised methodology. Agents draft proposals automatically, with human review on final output.

The Results

  • Quotes that took days now take minutes — respond to clients faster
  • Senior staff freed from repetitive pricing to focus on winning work
  • Fewer errors mean fewer awkward conversations with clients
Built with: Next.js, TypeScript
Agent Infrastructure

Data Aggregation Pipeline

The data backbone that research agents use to scan, collect, and consolidate information from multiple sources.

The Challenge

Business data spread across multiple websites and databases needed to be collected, cleaned, and consolidated. Manual collection took hours per week and was error-prone.

The Solution

Built automated data pipelines that scrape, clean, and consolidate data from multiple sources into a single, queryable database. Serves as infrastructure for research and monitoring agents.

The Results

  • Hours of weekly data gathering eliminated entirely
  • Team focuses on analysing insights instead of collecting them
  • Opportunities spotted faster — before competitors see them
Built with: Python, Playwright, SQLite
Agent Capability

Competitive Intelligence System

A capability now embedded into the agent team — automated competitive intelligence that runs continuously.

The Challenge

Understanding the competitive landscape required hours of manual research — visiting competitor websites, tracking their offerings, monitoring pricing changes.

The Solution

Built an automated competitive intelligence tool that profiles competitors, analyses positioning, and delivers structured reports. Runs continuously as part of the agent team's monitoring cadence.

The Results

  • Always know what competitors are doing — without spending time looking
  • Spot gaps in the market and act on them before others do
  • Strategic decisions informed by data, not assumptions
Built with: AI, Web Scraping, Automated Analysis
Agent-Ready Process

Automated Financial Reporting

The type of operational process that agents excel at — structured, repetitive, high-value reporting.

The Challenge

Month-end financial reporting was a multi-day exercise. Project costs had to be collated, monthly spend balanced against budgets, and reports formatted for different stakeholders — all by hand.

The Solution

Automated the entire month-end reporting process. Now serves as a template for reporting agents in client deployments.

The Results

  • Finance team spends time on analysis and forecasting, not data entry
  • Month-end closes faster — leadership gets numbers sooner
  • Budget overruns caught automatically instead of discovered too late
Built with: Python, Data Processing, Automated Reporting
Pilot Programme

Your Agent Team Could Be Next

We're selecting 3 Scottish businesses to deploy an AI agent team — reduced rate, full service, and your results featured here as a case study.

Start With the Free Assessment