No retainers. No lock-ins. Sprint by sprint, we earn the renewal.

It runs the marketing functions you would otherwise hire for: content production, social, email, paid acquisition, attribution, and AI search. Most of the daily output ships from AI agents in production. Our AI architects direct the agents, review the output, and run a monthly strategy session with you with catch-up calls in between. You see weekly output, not a slide deck of recommendations. Three tier sizes, all hours-based: Steady (£3,000 / 15 hours), Compound (£5,000 / 30 hours), Velocity (£8,000 / 50 hours). All three run on your infrastructure, no minimum commitment, hours spent however your function needs them that month.

A fractional CMO is one person with a narrow scope. An agency is a project-based pipeline of human deliverables. We are a marketing function that runs continuously on your infrastructure under our direction — closer to outsourcing the department than hiring a contractor. The cost is lower than a head of marketing (£80–120k plus benefits), the output is wider than a freelancer, and there is no minimum commitment — buy one sprint at a time. The other operational difference: every month, the system gets sharper, because it has another month of context on your business baked into it.

Because the value compounds. An AI marketing system with three months of your data, voice, customer transcripts, and feedback loops produces output that is specific to your business in a way no one-off build can. The first sprint sets the foundation. The second tightens it. By the third, the system knows your customers, your tone, and your numbers well enough that the work stops looking generic. A one-off project hands you a finished build and ends there. Sprints suit teams who want the system to keep getting better, sharper, and more specific to them month after month. Each sprint earns the next.

Everything we build sits on your infrastructure from day one — your cloud account, your CRM, your Notion, your stack. We get access to build, iterate, and run it for you. We never host your system on ours, so nothing is held hostage to a contract. Stop buying sprints and we lose access. You keep the published content, the analytics setup, the workflows, the custom agents, the prompts, the evals, and our internal runbooks for them. If you want to take operation in-house, spend a sprint on training instead of building. Your team takes over, and we step back. No notice period. No clawback. We don't keep clients on lock-in clauses. We keep them by being faster and sharper than running it in-house — and when we stop being that, you stop buying. That's the contract — that there is no contract.

Claude, ChatGPT, Gemini, and others depending on the job. Claude for long-form writing and structured reasoning. ChatGPT for fast, cheap, high-volume drafting. Gemini for some retrieval and grounding tasks. Smaller open models for narrow, repeatable jobs where token cost matters. We do not commit to one vendor — we run a multi-model stack and switch when the economics or output quality change. The /ai-models/ pillar on this site documents which model we pick for which marketing job, and why. Sprint pricing covers the model bill; you do not get a separate API invoice.