Bloodstone Projects

Building an AI Consultancy That Actually Delivers

When I started doing AI work, it was pure freelance — ad hoc projects, custom builds, one-off automations. The problem was obvious: every engagement started from scratch. No repeatable processes, no scalable delivery model. I was selling my time, not my expertise.

The Productisation Shift

Bloodstone was born from the decision to package what I knew into structured offerings. Instead of saying "I can build anything with AI," I defined clear service tiers: workflow automation audits, AI integration sprints, and ongoing optimisation retainers. Each one has a defined scope, timeline, and deliverable.

The tech stack is deliberately boring and reliable:

• n8n for workflow automation — self-hosted, no vendor lock-in, endlessly flexible
• Claude API for the AI layer — best reasoning model available, and I know it inside out
• Next.js + Supabase for any client-facing tools or dashboards

How We Work

Every engagement starts with a process audit. Most businesses think they need AI when what they actually need is better automation. We map their workflows, identify the highest-ROI opportunities, and build solutions that integrate with their existing tools. No rip-and-replace nonsense.

The key differentiator is that we focus on practical implementation, not hype. I've sat through enough pitch decks promising "AI transformation" to know that most of it is vapourware. Our clients get working systems within weeks, not slide decks about the future.

Challenges and Reality

The hardest part of running an AI consultancy is managing expectations. Every prospect has read the headlines and thinks AI will replace their entire workforce by Tuesday. The real work is education — showing what AI can genuinely do today, where the limitations are, and how to build systems that improve over time.

Pricing was another puzzle. I experimented with hourly, project-based, and retainer models before landing on a hybrid: fixed-price sprints for initial builds, monthly retainers for ongoing optimisation. The retainer model is what makes the business sustainable.

What's Next

We're building an internal library of reusable automation templates and AI prompts that accelerate delivery for new clients. The goal is to reduce time-to-value from weeks to days while maintaining the bespoke feel that clients expect.