ai automation for every business.
We design and build custom AI for businesses of every size: agents that read documents, answer the phone, draft plans and answer questions about your own data, plus the automation that connects your systems. It's built into the tools you already use, with a person approving anything that matters.
what we build.
document + invoice reading
Invoices, receipts, forms, contracts and statements read by AI: the key fields pulled out, checked by a person and sent on to your accounts or systems.
ai agents
Agents that carry out multi-step work: look things up, call your systems through tools, draft the result and hand it to a person to approve.
voice + phone agents
Voice agents that answer calls, understand what the caller needs and book it straight into your system.
assistants over your data
Ask questions about your own jobs, customers, stock, rotas or documents in plain English. Read-only by default, so they can look but never change.
ai planning + scheduling
Rotas, routes, schedules and allocations drafted by AI from your rules, then adjusted and approved by a manager.
workflow automation + integrations
The quiet jobs done automatically: reminders chased, data synced between systems, reports compiled, emails and documents sent at the right moment.
speech, text + vision
Voice notes turned into structured records, photos read for the details that matter, and long documents summarised.
ai inside your product
AI features for your own app or platform: smart search, drafting, classification, recommendations and chat, designed for your users.
ai strategy + prototypes
Not sure where AI fits? We find the jobs worth automating, prototype the best one fast and prove it on your real data before you commit to more.
every size
of business.
From a founder's first product to a platform for a large organisation, the approach is the same: you talk to whoever is building it, you get a fixed quote, and we stay on after launch.
founders + startups
A first product built properly: scoped to what matters, launched fast and ready to grow, raise on and scale.
growing businesses
The spreadsheets, group chats and paper replaced with one system that fits how you already work, and grows with you.
large organisations
New products, internal platforms, AI and integrations built alongside your existing systems, teams and processes.
ai reads.
a person decides.
Every AI system we build follows one rule: the AI does the reading and drafting, a person makes the call. Invoices land in a review queue before they reach the accounts. Rotas are drafted, then a manager moves, swaps and publishes. Assistants can answer questions about the business, but they can't change anything.
The cost of being wrong isn't the same on both sides. A missed invoice is annoying; a wrong figure pushed silently into the accounts is a problem you find at year end. Keeping a person on the final step costs a few seconds and removes that risk, and it's why owners who've never used AI are happy to let it do the typing.
Here are three AI systems we've built, drawn node by node. Two are live today; the phone agent is built and not yet live.
a phone booking lands in the same queue as app and portal bookings, then dispatch takes over.
three ways in, one review gate. nothing reaches the accounts without a person tapping approve.
the AI drafts the week across all three sites. the manager moves, swaps and publishes.
the proof.
Products we build and run ourselves, and businesses we've built for. Full case studies are coming soon.


gaffer
gaffer reads your supplier invoices, turns a voice note into a booked job, chases services before they're due and answers any question about the business in seconds. Fewer evenings in the office, less paid admin, nothing forgotten. It runs MW Plumbing & Heating today.
./open-case-study →


nara
nara drafts the rota, reads delivery notes and till reports from a photo, drafts HR letters, screens CVs and answers any question about your sites, so managers spend the shift on the floor, not on paperwork.
./open-case-study →


PingCab
A booking platform for licensed local taxi firms: a rider app, a driver app, a dispatch portal for the firm and a website covering 26 towns. Fixed fares from the council tariff, no surge, fully bilingual.
./open-case-study →
how it works.
Six steps, from the first conversation to keeping it running. More on working with us →
- 01 /talkWe start with the idea, the problem and where you're at. You don't need a finished brief.
- 02 /scopeWe work out what it needs to become, and what it doesn't, and what it has to connect to.
- 03 /fixed quoteYou get a fixed quote in writing, with the timeline, before anything starts.
- 04 /buildDesign and development under one roof. You see progress as it's built.
- 05 /launchWe get it working and into the world: on your domain, in the app stores or in your team's hands.
- 06 /runWe stay on to host, update and support it, so you're never left with something nobody looks after.
questions,
answered.
what can AI automate in a business?
Most repetitive, rules-heavy work: reading documents and invoices, answering routine calls and questions, drafting schedules and plans, turning voice notes into records, triaging requests and moving data between systems. If a person does it the same way many times a week, it's worth looking at.
will the AI change our accounts or records on its own?
Only if you want it to. Our default is that the AI reads and drafts and a person approves, and assistants are read-only. For low-risk jobs we can let automation run on its own, with every action logged.
what happens if the AI gets something wrong?
We design for it. Results are checked at a review step before they count, confidence is shown so doubtful ones get a closer look, and we build a fallback so the work never stops if the AI is unavailable.
can it connect to the software we already use?
Usually, yes, and that's often the point. We connect to accounting, CRM, booking, messaging and in-house systems through their APIs, and we check what your existing tools allow while we scope the work.
which AI models do you use?
We pick the model for the job. The AI systems we run today are built on Claude, with Whisper for turning speech into text, and we can work with other providers where a project needs it.
is our data safe with AI?
Each client runs on their own private database and deployment, and we only send the AI the data a task needs. We'll sign an NDA before you share anything sensitive.
do you work with large companies as well as small ones?
Yes. We work with founders, small and growing businesses, and large organisations. The process is the same at every size: you talk directly to whoever is building it, you get a fixed quote, and we stay on to run and support what we ship.
how much does it cost?
There's no price list, because no two builds are the same. We scope the work with you, then you get a fixed quote in writing before anything starts. We take on projects of every size, from a focused first version to a large, multi-part platform.
ready?
have something you want to build?
It doesn't need to be figured out yet. Bring the idea, the problem and where you're at. That's enough to start.