Internal operations
Research, reporting, monitoring, follow-up, and repetitive coordination work turned into reliable workflows.
AI systems for real operations.
For small operators who need useful systems, not demos. From simple automations to more structured AI systems.
What I build
Systems that actually take work off your plate.
Tool-connected automation
Systems wired into the tools you already use: APIs, docs, dashboards, spreadsheets, CRMs, and messaging surfaces.
Validation and continuity
Memory, handoffs, audit trails, and truth checks so the system stays usable over time instead of drifting into guesswork.
Proof
What I built for myself
Over 6 months, I built an autonomous multi-agent trading research system for Polymarket.
The system used a researcher + engineer architecture, file-based memory, structured handoffs, Chainlink integration, and validation / reconciliation layers to test whether the strategy was real or just badly measured.
The conclusion was honest: the bot did not find durable retail edge in 5-minute crypto markets. The valuable output was the architecture and operating discipline that made that conclusion trustworthy.
One example of the work: the system initially reported 90% accuracy. After rebuilding the measurement layer, the real figure was closer to 60%.
How I work
Simple process, serious implementation.
Map the workflow
We identify where time, context, or decision quality is currently leaking so the project solves a real problem.
Scope the right level of automation
Sometimes that means a simple script. Sometimes it means a more structured AI system. The scope should match the job.
Build and integrate
I connect the system to the real tools, data, and surfaces the work actually depends on, not a demo environment.
Validate and hand over
The output has to be trustworthy, understandable, and usable by you after delivery. Includes 30 days of post-launch support.
Pricing
Typical engagement range
Enough transparency to qualify the conversation, without pretending every project fits one fixed box.
Audit / scoping
From £750
Discovery, workflow diagnosis, and a written implementation roadmap for the highest-value next step.
Implementation
Typically £1.5k–£5k
Build, integration, testing, handover, and support sized to the complexity of the system.
Typical ongoing costs are low and paid directly by the client: hosting is usually modest, and API spend depends on usage.
Contact
Building something similar for your business?
I work with small operators who need serious systems, not demos.