Flareworks: practical AI systems for real business workflows

My AI systems practice helps SME founders and knowledge workers find the repeated work worth improving, build the smallest useful system, and keep human judgment at the important checkpoints.

Flareworks: practical AI systems for real business workflows
Role
Fractional CTO and AI systems advisor
Timeline
2026–present
Published
September 2026
Read
2 min

What it is

Flareworks is my AI systems practice for founder-led businesses, SMEs, and knowledge workers. It was incorporated in 2025 and has been working with clients since 2026: 3 client engagements and 1 workshop so far. It exists for companies that can see the potential of AI but do not need another disconnected tool or a long transformation programme.

The work begins with a real workflow: enquiries arriving, leads being qualified, proposals being written, documents being searched, or repeated decisions being made. We identify where time, money, or quality is being lost, then decide whether AI deserves a place in that process.

How I work

  1. Start with the business outcome

    Define the result in plain language: increase profit, save time, or improve the quality and consistency of the work.
  2. Map the current workflow

    Find the inputs, decisions, handoffs, exceptions, and places where human judgment matters.
  3. Build the smallest useful system

    Test one valuable path before automating the entire operation.
  4. Keep people in control

    Make confidence, escalation, and review visible so the system supports expertise instead of hiding risk.

Why it is different

I bring the perspective of an operator who has built AI products, raised capital, led a product team, and now runs AI-agent workflows inside my own companies. The recommendation is shaped by what will survive daily use, not by how impressive the demo looks.

Selected work

  • MummyReady: a quiz funnel for an AIA advisor team that routes each booking to the right advisor.
  • Insurance knowledge workshop: a hands-on session with two independent insurance advisors, and what it taught me about deciding the questions before choosing the tools.
  • AI agent reply system: an agent team that has published more than 15,000 replies. Routine batches post under rules I approved, and anything uncertain waits for me.

As someone without a technical background, I could focus on the business idea while Tony translated it into a working prototype our team could understand and build on. Even without a development team on our side, he handled the technical side while keeping the focus on the business goal.

Nicholas ChongSenior Director, AIA

Show me the workflow. I’ll tell you honestly whether AI belongs in it.