AI Build
Bring a brief or bring a problem. We build either way.
Workflow automations, document pipelines, custom AI assistants. Working software, not strategy decks.
Path A · Have a brief
You've done the thinking. You know what you want. Hand us the brief and we implement it.
Path B · Have a problem
You know the outcome you need. We scope it with you, write the plan, then build.
Built for
How we work
No agreed plan, no build.
Every engagement starts with a written plan, either yours or one we write with you. This protects you: you approve exactly what gets built before we start.
A clear plan means we aren't guessing mid-build. It means you can challenge scope before it becomes work. It means both sides know when a project is done.
Path A clients bring their own brief. Path B clients get a scoping phase before any build begins. In both cases the plan is approved before development starts.
What we build
Four kinds of systems.
Work that stops being manual
Email triage, data entry, report generation, scheduling. The processes where a person is currently acting as the connector between two systems.
Typical result: A task that took hours a day becomes a review step that takes minutes.
Documents that read themselves
Contracts, invoices, applications, reports. Extraction, classification, summarisation and routing at volume.
Typical result: The six fields someone pulls out by hand come out automatically, with a human confirming the exceptions.
Internal assistants your team actually uses
Domain-specific, context-aware, connected to the systems you already run. Not a chatbot bolted on the side.
Typical result: Institutional knowledge becomes searchable by the people who need it, not just the people who wrote it.
Existing tools, made sharper
CRM enrichment, support escalation, proposal generation, meeting summarisation. AI inside the workflow you already have, rather than another tool to adopt.
Typical result: Nobody changes how they work, the work just takes less time.
Process
How it works
- 1
You share a brief or detailed description
- 2
We review scope and confirm feasibility (1–2 days)
- 3
Delivery in agreed stages, with demos
- 4
Handover with documentation and training
- 1
You describe the problem or outcome you need
- 2
We scope it with you over 1–2 weeks
- 3
You approve the plan before build begins
- 4
Delivery in agreed stages, then handover
How the work runs
Once we know the scope, you get a plan you can see.
We don't quote until the scope is agreed. Once it is, we lay out the stages, what you'll see at each one, and when. This is the shape the work follows:
Stage 1
The riskiest unknown first
Environment, integrations, and the riskiest technical unknown first. If something is going to be harder than expected, we find out while there's still time to change course.
Stage 2
Core build
First working demo as early as possible, running on your real data rather than samples.
Stage 3
The quality lives here
Edge cases, failure handling, and the human review points. This is where most of the quality lives, and where most projects skip.
Stage 4
Handover
Hardening, documentation, handover, and training your team to run it.
Throughout, a demo, a written update, and an honest note on anything that slipped. No status theatre.
What they said
From founders who've shipped with us
“I'd looked at AI tools before and always came away thinking they were built for companies much bigger than us. Grapine started with how we actually work rather than what the technology could do, which is why it stuck. Eight hours a week back on fulfilment, and my team didn't have to change the way they work to get it. I'd point any founder running a lean operation their way.”
“What I'd expected was a strategy document. What I got was something we could actually use, and quickly. They were direct about what wasn't worth doing, which I hadn't had from anyone in this space before. A day a week back on correspondence alone. Worth the conversation for the clarity alone.”
Pricing
Engagement-based pricing.
We work on a monthly retainer. Engagement length is scoped to the work, some builds are a single stage, others run for several months. We quote after reviewing your brief or problem.
From $15,000/mo
Monthly retainer
Agreed
Stages
Scoped
To the work
What the quote accounts for: build effort, integrations, the cost of running the system once live, and handover. We'd rather tell you the real total than the attractive one.
Not sure whether your project fits? Send us a brief and we'll tell you honestly.
Honesty
We turn down work. Here's when.
When there's no measurable outcome
If nobody can say what 'working' means, the project has no end. We'll help you define it, or we'll pass.
When the process underneath is broken
AI applied to a bad workflow produces bad output faster. Sometimes the honest answer is to fix the process first, with no AI involved.
When the volume doesn't justify it
If a task happens four times a month, automating it is a hobby, not an investment. We'll say so.
When you need a person, not a system
Some judgement doesn't transfer. We'd rather tell you that now than after you've paid for it.
This is the same reason our audit and build are separate engagements. Neither of us benefits from building something you don't need.
You don't need an audit before you build with us.
If you already know what you want and have a clear brief, start here. If you're not yet sure where AI fits your business, the AI Opportunity Audit is the better starting point, and it's designed so you can take the report anywhere, including away from us.
Get started
Tell us what you need.
Send a brief, or just a description of the problem. We'll tell you whether it's a fit and quote within 48 hours.
Prefer to talk? . Or email [email protected], we respond within 24 hours.