Give it goals and constraints
We set out the users, the problem, the features needed and the constraints, weigh options from several angles, and decide what to build.
PROCESS & PRICING
From first conversation through development and handover. We share scope, deliverables and how results will be checked, so both sides can make informed decisions.
HOW WE WORK
We hear what is getting in the way, who the users are, your current setup and your timeline, and tell you whether we can help.
First consultation free
What we build and how we verify it. Deliverables, acceptance criteria, cost and timeline are fixed before work starts.
Paid research flagged before it starts
We share screens and prototypes to stay aligned while building, and adjust scope if changes are needed.
Decision points along the way
You receive the deliverables, test results and any runbooks you need. Ongoing support can be arranged separately.
Operations and maintenance agreed separately
AI & ENGINEERING
We use Claude and ChatGPT in development
for design, implementation and review.
We set out the users, the problem, the features needed and the constraints, weigh options from several angles, and decide what to build.
One role implements, another reviews, and findings get fixed in repeated rounds. Changes build up on top of working software you can see.
The work is checked against the agreed features and conditions. For complex states and flows, we also consider model-based verification.
Deterministic work such as calculations, format checks and state updates is done in ordinary code. Which information goes to AI, and through which service, is agreed as part of the project.
In practice: building furnaceSCOPE & COST
Two “business apps” can be very different jobs depending on the data, the integrations, the number of users and the quality required. We confirm the points below, then propose a quote and a plan.
Where assumptions are unclear, we may quote research or a prototype as a separate step. Third-party service fees and post-launch operation and maintenance are quoted separately where relevant.
INITIAL CONSULTATION
An outline of the problem, whether we can help, and how we would proceed, including what needs checking first. The first consultation is free.
PAID ENGAGEMENT
Detailed research, tailored design, prototyping, implementation and evaluation. These begin once scope, deliverables, cost and timeline are agreed.
BEFORE WE START
The current pain and what better looks like. If you don’t have numbers yet, describing how things are used is enough.
A person who can explain the workflow and review prototypes and deliverables. We agree a single point of contact for decisions.
The systems and data involved, the environment and your timeline. Confidential information is handled only after we agree how to share it.
FAQ
Yes. Start with what is getting in the way or a rough idea of what you want to build. Knowing the work or system involved, who will use it and your timeline helps us work out what to look at next.
We listen to an outline of the problem and confirm whether and how we can help. Detailed assessments of your current setup, design documents and custom prototypes are done after scope and cost are agreed.
Yes. Focused requests are welcome, such as one feature, one workflow or fixing issues someone has already identified. We quote once scope and acceptance criteria are clear.
We quote each project after reviewing scope, deliverables, integrations, quality requirements and deadline. If key assumptions are unclear, we suggest doing research or a prototype first. Any additional work is also priced and scoped before it starts.
We use Claude and ChatGPT for planning, implementation and review. Building and checking are separate roles; findings are fixed in repeated rounds, and the result is verified against the agreed acceptance criteria. If AI features are to be built into your own operations, purpose, accuracy, cost and data handling are assessed separately.
Of course. Calculations, format checks, data integration and anything else governed by clear rules are handled with ordinary code. AI is one option we consider where it genuinely helps, for example with text.
We decide after reviewing the architecture, access to source code and environments, ownership of rights and available handover information. If an initial assessment is needed, we’ll tell you its scope and cost.
Yes, with access control, personal data protection, logging and monitoring, and fixes for known vulnerabilities. We do not guarantee certification, passing an audit or finding every vulnerability. We agree what safeguards are needed and what we will cover.
Post-launch maintenance is agreed separately: hours, systems covered, scope, contact method and cost. Round-the-clock monitoring and emergency response are not part of our standard offer.
An outline is enough at first. Please don’t enter confidential information, business documents containing personal data, or credentials in the form. When we need materials, we agree confidentiality and how to share them first.
Yes. For example, your firm handles consulting and proposals while we take on development. We clarify the client relationship, who does what and how decisions are made, and find a way of working that suits both sides.
LET’S TALK.
YOUR NEXT CHAPTER.Tell us what’s slowing you down or what you want to build.
That’s where we start.
The first consultation is free. We’ll review what you send and tell you whether we can help and how we’d go about it.
Detailed research, design and prototyping start only once we’ve agreed on scope and cost.