Our Mission
We replace the systems nobody owns - the spreadsheet that quietly became a database, the legacy server whose access list no one has reviewed in years, the tool you rent forever and leave with nothing - with one system of record your business owns outright: the source code, the data and the hosting, in your name.
A mission is only worth publishing if a client can catch us failing it. Everything below is written as a commitment, the mechanism that enforces it, and the place you can go to check whether we did it.
Most mission pages list virtues. A virtue cannot be tested. These three parts can be.
Worked example
Why we exist
The most expensive system in a company is usually the one everyone says still works. Customer records sit in three spreadsheets, the same order is entered twice, and the one person who understands the legacy database is a single resignation away from taking that knowledge out of the building. Ending that situation is the whole reason this company exists.
None of this shows up as a line on the P&L. It shows up as overtime, as a quote that goes out at the wrong price, and as a decision made a week late. We are not against the tools your team uses. We are against the version of your business that only exists in someone's head.
Our commitments
A value is only worth printing if a client could catch us breaking it. Each commitment below names what we do, the mechanism that enforces it, and a page where you can see it in practice. If one of them matters to you, quote it back to us and we will write it into the contract.
The source code, the database and the hosting accounts are in your name from the first phase. If you replace us, nothing stops working and nothing has to be bought back.
How it is enforced
Ownership is a clause in the contract, not a line on a values page. At the end of every phase we hand over the repository, the database schema and the credentials, with written instructions for running the system without us.
You get software your team can actually use in weeks, instead of a twelve-month plan that is out of date before it lands.
How it is enforced
Scope is cut into phases that each end in something running in production. You get a working end-to-end path on your own data inside 30 days and a reviewable release every week after that, and on most engagements a first phase is live within 90 days.
You should never have to ask what a change costs, who is working on it, or what is holding it up.
How it is enforced
Fixed price where the scope is known, time and materials where it is not, and which one applies is agreed in writing before work starts. One board shows what is done, what is next and what is blocked, and you have a login to it.
Payroll, margins and customer records are visible to the people whose job needs them, and to nobody else.
How it is enforced
Roles and permissions are built in the first phase, not added after an incident. Record changes are written to an audit trail, and backups are restored on a schedule so we know the restore works.
Some problems are process problems, and a new system only makes them happen faster. We say so before you spend the money.
How it is enforced
Discovery ends in a written scope that names what is out, what is a process change and what we recommend you do not build. If an off-the-shelf tool is the cheaper answer, that is the answer you get.
The system uses your terminology, your job numbers and the handoffs your team already trusts.
How it is enforced
Before a single table is designed we map how a job really moves through your business, from first enquiry to cash, and we write that map down where your team can correct it.
AI shortens delivery and takes repetitive work off your staff. It does not make the decisions you are paying a firm to be accountable for.
How it is enforced
We name every place AI is used, what data it can see and which person signs off the output. If a result cannot be checked by someone on your side, it does not go into production.
Clients who have held us to these are the best evidence we can offer. Their words are on the client testimonials page, and the numbers behind an engagement are on the pricing page.
AI, honestly
AI does two useful things on a project: it compresses the work of building software, and it removes repetitive work from your staff. It does not own architecture, judgement or accountability, and a firm that implies otherwise is selling you the demo. Saying exactly where the line sits is what makes the claim checkable.
| What AI does | What a named person owns |
|---|---|
| Drafts and refactors code, writes first-pass tests, generates migration scripts. | Chooses the architecture and the data model, and reviews every line before it merges. |
| Reads PDFs, emails and scanned forms and turns them into structured records. | Approves anything that moves money, changes a price or edits a customer record. |
| Summarises, classifies and answers questions over data you already own. | Decides which data a model may see, and what it is never given. |
| Handles repetitive internal lookups and first-draft writing for your staff. | Owns the answer that reaches your customer. |
| Shortens the build and lowers what a phase costs to deliver. | Signs off the release and carries accountability for what shipped. |
Every AI feature we ship names the data it is allowed to read and the person who reviews what it produces. Where an output cannot be checked by someone on your side, we do not put it into production. That rule costs us the occasional impressive demo, and we are content with the trade.
Our vision
A vision nobody can check is decoration. Ours is dated and testable: by 2030, every business we build for should be able to operate, extend and audit its core system without calling us, and keep calling us anyway because they want to, not because they are stuck.
Readable code, a documented schema, and an environment that starts from the repository and a short setup file.
Reporting sits on one system of record, so a manager filters and exports instead of asking someone to write a query.
Every change carries who made it, when, and what the value was before. That turns an audit or a dispute into a five-minute job.
Those three tests are how we grade our own work at the end of an engagement. If you want to see the standard applied to a finished system, the Construction Quotes case study walks through the build, the handover and what the client kept.
The other half of a mission
Any firm can list what it believes in. The part that costs something is the work it turns down. These are five jobs we decline, including profitable ones, and you are entitled to hold us to them during a proposal.
If the right people are not free, we tell you when they will be or we point you elsewhere. An engagement that starts short-handed never catches up.
If the problem is a process, two integrations or a training gap, that is what we quote, even when the rebuild would have been the larger invoice.
No exit fee, no export charge, no account registered in our name. You can leave at the end of any phase with everything that was built.
If the work does not need a model, the proposal will not contain one. A pilot that cannot show hours removed is not a pilot, it is a demo.
A number given before discovery is a guess with a decimal point. You get a range, the assumptions behind it, and exactly what would move it.
If one of these rules means we are the wrong firm for a piece of work, we will say so in the first conversation rather than the third invoice. Where we do fit, the shape of an engagement is set out in our pricing structure and in how we test what we ship.
Questions before you call
These are the six questions buyers actually ask once the introductions are over: who owns the result, what happens in a disagreement, whether we fit a company your size, and how AI is used on the work. The answers below are the same ones you will get on a call.
You do. The source code, the database and the hosting accounts are in your name, and ownership is written into the contract rather than promised on a values page. At the end of every phase we hand over the repository, the schema and the credentials, so there is never a moment where leaving us costs you the system.
You decide, and we write it down. You get the recommendation, the risk we see and the cost of both options in writing, and then we build what you choose. Recording the decision means that six months later nobody is arguing about what was agreed or why.
Almost certainly. Since 2014 we have delivered 720+ projects for 680+ clients across 21+ countries, from single-location contractors to multi-site operators. The smallest sensible engagement is one phase with one clear outcome, not an open-ended retainer.
We use AI to compress delivery work such as code, tests, data migration and documentation, and to remove repetitive work from your staff, such as reading documents or classifying records. Architecture, judgement and accountability stay with named people, and every AI feature we ship names the data it can see and the person who reviews its output.
Most engagements start within one to two weeks of signing. You get a working end-to-end path on your own data inside 30 days, in a staging environment you can log into, and a reviewable release every week after that. On most projects a first phase is running in production within 90 days, because scope is cut into phases that each end in something your team uses.
Then we will tell you to keep it. If the real problem is two tools that do not talk to each other, a report nobody can produce, or a step that is done twice, we would rather quote the integration or the reporting layer than a rebuild you do not need.
Drop us a line - we're here 24/7 to answer your questions and discuss your project ideas.
