Straight answers about how we work.

By scope and outcome, not hourly. After a short discovery call you receive a written scope with deliverables, timeline and a fixed price. Larger systems are split into phases, each with its own price, so you are never committing to an open-ended budget.

Only if you change the scope. Anything outside the agreed scope is quoted separately in writing before it is built, and you decide whether it belongs in this phase or the next one.

Staged against delivered milestones rather than calendar dates. We invoice from a UK registered company account and accept bank transfer, and other common business payment methods on request.

A single integration is usually days. A workflow automation is typically two to four weeks. A custom business system runs in phases, with the first usable phase normally live within four to eight weeks — we prioritise getting something real into production early rather than a long silent build.

Rarely, and not by default. Most of the value comes from connecting and orchestrating what you already run. We replace a system only when keeping it costs more than moving off it, and we say so plainly.

Meaningfully at the start and lightly afterwards. We need the people who actually run the process for the modelling sessions, then short review checkpoints. If we have to guess how your operation works, the system will be wrong.

A support window is included in every engagement for stabilisation and fixes. After that you can continue on a support arrangement, take it in-house, or call us only when something new is needed. Nothing locks you in.

Often less than people expect. AI earns its place where there is judgement over language or unstructured data at volume — classification, extraction, drafting, retrieval. Deterministic rules are cheaper and more reliable for everything else, and we will tell you when that is the honest answer.

Not on consequential actions. We design an approval gate wherever an action has financial, legal or customer impact. The model proposes; a person confirms; the system records who confirmed and when.

No. We build on your data for your system. We do not contribute client data to model training, and we configure providers to keep your content out of training pipelines where that setting exists.

Grounding on your approved sources, an evaluation set built before launch, confidence thresholds that route uncertain cases to a human, and full logging so any output can be traced back to the source it used.

You do. On final payment, the source code and all project data are yours, deployed in your own accounts and repositories wherever possible. We do not hold your business hostage in our infrastructure.

Yes, routinely and before any detailed discussion if you prefer. Client confidentiality is also why this website shows solution patterns instead of named case studies.

Least-privilege access, credentials held in secret managers rather than documents, environments separated between development and production, and access removed at handover.

English, Arabic and French, from a UK registered company. We work comfortably across European, Gulf and North African business hours.

Yes — through the partner programme, either as a white-label delivery team behind your brand or on a referral basis with a 10% commission.