Digital Signet is an AI studio in London. We build voice, finance, and data AI agents for mid-market companies across the UK, US, and Australia. Twenty years of shipping software backs the engineering. Five live products, an open-source security tool, and 300+ automated content sites back the studio claim.
I have been a developer, tech lead, engineering manager, architect, and CTO across two decades in software. I have delivered platforms across satellite communications, UK media broadcasting, sports media services, global athletic apparel, premier-league football, national rugby, and UK industrial engineering. I have scaled teams, modernised legacy platforms, introduced Agile where there was chaos, and shipped across every role a software delivery team has.
In 2025 I stopped taking full-time roles. Not because I stopped enjoying the work, but because AI changed what a single engineer could deliver. I can now ship, with a small studio and the right tools, what used to take a squad of twelve. The economics of consulting changed with it.
Digital Signet is the AI studio I founded to combine that engineering experience with AI agents that can actually build software. We help mid-market companies ship voice agents, finance agents, and data agents. We use it ourselves to ship our own products on the same stack.
The credibility for client work is the work we have already shipped for ourselves. Bottie is a consumer voice AI with paying users; the same stack powers the voice agents we build for clients. Cageclaw is an open-source security sandbox for AI coding agents, Apache 2.0. Sentaz is secure file sharing for professional services. Signet LLM is a browser-based transformer demo built from scratch in TypeScript. AI Job Impact Calculator is an independent reference site with full source-citation methodology.
Three productised practices: Voice AI for sales calls, customer service, training and induction, AI receptionists. Finance AI for bookkeeping, accounts payable, accounts receivable, contract terms extraction. Data AI for document extraction, internal knowledge agents, RAG infrastructure, AI search. We also take generic AI product builds and AI opportunity assessments for clients whose work does not fit one of the three pillars.
We bring engineering, architecture, design, project delivery, and named specialists into every engagement. Everyone has shipped the kind of work they are doing for you, in production, with real users.
The deliberate choice is to ship, not to bill hours. We size to the engagement, never to a payroll target. The person who scoped your project on the discovery call is the person on the build calls. Never a partner-sells / offshore-builds firm.
Engagements are scoped upfront in writing. No retainer creep, no open-ended engagements that cannot be exited. When we build an agent, you own the code, the configs, and the infrastructure. If we part ways, the agent keeps running. No vendor lock-in.
We run everything human-in-the-loop by default. Every agent action worth reviewing goes through an approval gate until you choose to raise autonomy. Audit logs are standard. Kill switches are yours.
We are honest about what AI can and cannot do. If a pilot engagement shows the agent is not the right fit, we will say so and refund the setup if we agreed that upfront. The goal is clients who come back for the next engagement, not clients who regret the first one.
