AI Readiness Audit
Two weeks inside your operation. We find where the hours leak, rank every opportunity by return, and hand you a roadmap you could build with anyone.
Software that does the busywork, so your people do the work.
AI agents and automations designed with the same care as a product: useful, trusted, on-brand, and measured against a number agreed up front.
- A two-week audit
- One workflow live in 6 weeks
- Monitored and improved monthly
Meanwhile, the internet changed its mind
SearchAsk.
BrowseDelegate.
Hire for itAutomate it.
Customers ask an assistant before they search. Agents compare and buy on their behalf. The busywork inside every company is moving to software. The companies that design for this now will be the ones the assistants recommend.
It’s 2am. A customer emails. Nobody is awake.
This is an AI agent: software that reads, decides and acts across the tools you already use, with guardrails you set and a person in the loop when it matters.
- Email received: "Do you ship replacement hinges to Rotterdam?"
- Classified: spare parts, Netherlands, existing customer
- Customer matched; previous order and model found
- Stock checked: part in the Leeds warehouse, ships Monday
- Reply drafted in your tone of voice, guardrails passed
- Sent. Order link included. Summary posted for the team
Six steps. Six seconds. The customer has an answer before you’ve woken up.
Eight ways in.
Most AI projects fail for design reasons, not technical ones. Nobody trusts the output, the handover to a person is clumsy, and the agent sounds like nobody in particular.
Workflow Automation
The repetitive middle of the business, run by software: intake, routing, quoting, reporting, follow-up. Built on the tools you already pay for.
Customer-facing Agents
Support, sales qualification, booking and phone reception, available at 2am and on-brand at every hour. They hand over to a person the moment they should.
Knowledge Assistants
Your SOPs, contracts, specs and service manuals, answerable in seconds by the people who need them. Private, permissioned, cited back to source.
Agent-ready Websites
People now ask ChatGPT, Gemini and Perplexity before they ever search. We make your site legible to them: structured data, llms.txt, crawler policy, and citation tracking.
Agentic Commerce
Shopping agents now compare and buy on people’s behalf. We get your catalogue, pricing and checkout ready for the new protocols (UCP, ACP, MCP) without giving up the customer relationship.
AI Product Design
Interfaces for features that think: trust, uncertainty, latency and the handover to a human, designed as carefully as the happy path.
Operate
Models change monthly. We monitor, evaluate and improve every agent we ship, and add the next automation when the business is ready for it.
Start small. Prove it. Then scale what works.
We start with an audit of where the hours actually go, prove one workflow in production, then scale what works. You own everything we build.
- Audit
We map the operation and rank what to automate first. You keep the roadmap whatever you decide next.
- Pilot
One workflow or agent, live in production, measured against the number the audit promised.
- Operate
Monitoring, evaluation and the next build. The system gets better every month instead of quietly drifting.
Audit
We interview the team, shadow the work and map every repetitive process. Each opportunity gets an hours-saved estimate and a build cost.
Design the agent
What it may do alone, what it must ask about, how it sounds and how it hands over. Drawn and agreed before any code.
Pilot
One workflow, built on your tools and data, running in production with a person approving its work until the numbers say otherwise.
Scale
Proven workflows earn more autonomy; new ones join the queue in order of return.
Operate
Models change monthly. We monitor accuracy, cost and drift, and keep improving every agent we ship.
The part other studios keep vague.
Working with us is deliberately unexciting in all the ways that matter. Here is exactly what you can hold us to.
A number up front
Every pilot is measured against the hours or revenue the audit promised.
People in the loop
Agents earn autonomy step by step. Nothing customer-facing goes out unchecked until it has earned it.
Your data stays yours
Private deployments, permissioned access, no training on your data, and you own every prompt and workflow.
Not sure where your project fits? Book the call. Half an hour, no charge, and we’ll tell you honestly if we’re the right office for it.
Fair questions.
The things people usually ask on the first call. If yours is not here, ask us.
Especially. The audit is written in plain English, and every agent is designed so your team can supervise it without learning to code.
Whichever fits the job and your data rules: OpenAI, Anthropic, Google, or open models hosted privately. We stay model-agnostic so you are never locked in.
The audit is a fixed fee. Pilots are scoped from the audit at a fixed price. Operate is a monthly retainer sized to what is running.
It occasionally will, so we design for it: confidence thresholds, approval on anything risky, full logs and a one-click route to a person.
Who we do it for.
Sectors where we know the buyers, the constraints and the shortcuts. If yours is not listed, the first call costs nothing.
Your ai project, next.
We need
starting