Questions, answered plainly.

We never let it act alone. The model only handles the steps that need judgement, and everything it produces passes through code before it touches your systems: checked against your rules and your data, and stopped for a person when something is off. A wrong answer is caught there, not in your CRM or your accounts.

Use them where they fit, and we will tell you when they do. They stop being enough when the work needs your own rules and data, has to update your systems, and cannot afford a confident wrong answer. That is what we build: your process, your checks, your tools, and a person in the loop where it matters.

Almost always. We connect to what you run on, from Gmail, HubSpot, Xero or Shopify to your own database or an older internal system. Mapping those connections is part of the first step, so there are no surprises once the build starts.

The prices are on this page: a paid mapping and test first, then a fixed-price pilot, then a monthly fee to keep it running. On the first call we size it against the hours it removes, and if the numbers do not work for you, we say so before you spend anything.

To a server in the EU that we set up for you, and to the AI provider under business terms that forbid training on your data. We sign a confidentiality agreement and a data processing agreement before any data moves, and test data is deleted within thirty days if you do not go ahead.

Because you do not have to take our word for anything. You see it work on your own data before you commit, the pilot has a fixed price and a date, and if we miss the date you get your money back. You own the code from day one, so it stays yours whatever happens to us.

Then you find out in the first step, for a fixed fee that comes off the build if you go ahead, and we say so plainly. Some processes depend on judgement that is not written down anywhere, and no build fixes that. The test exists so neither of us has to guess.

You do: the code, the configuration, the data and the deployment, in an account you control. If you stop working with us, it keeps running. You can hand it to your own team or another vendor.

We build with one layer between your software and the model, so a model can be swapped without rebuilding anything. When a better or cheaper one appears, switching is a small change, not a new project.

No. We build it, host it and keep it running. You need one person who knows the process well and can answer questions within a day. If you do have developers, we work alongside them and hand over cleanly.

Mapping and testing takes about a week. A pilot covers one process, with a price and a date we fix after the test. Larger systems and custom software are planned as a series of fixed-price steps, each with its own date, so you see progress every few weeks.

It depends on your process and your data, so we measure it on your own sample in the first step rather than quote a number on a call. What we can promise is how it behaves when it is wrong: it shows what each answer was based on, flags what it is unsure about, and stops for a person instead of guessing.

Yes. Web apps, internal tools and mobile apps on both stores. They are scoped and priced after a call, as fixed-price steps.

Yes. You contract with Sifartek in Estonia under EU law, which is usually straightforward for companies in the Gulf or the UK. Hosting can be in the EU or in your region if your rules require it. We work in English, and our tools handle Arabic and French properly.