If you are searching for what AI automation costs, you have probably met answers that say nothing: "it depends" or "book a call". It really does depend. Anyone quoting a fixed number before understanding your process is guessing. Here we explain what moves the price, why we don't publish a price list, and how to ask for an estimate.

What determines the price of an automation?

The cost sits in the engineering work around the AI model. Four things move it more than anything else.

How complex is the process being automated?

A process with one input, one clear rule and one output is simple to automate. A process with exceptions, edge cases, human judgement calls or steps that change depending on context takes much more mapping and building. That mapping is real work even before any code gets written.

How many systems need to be connected?

Automating something inside a single tool, for example organising data that is already in one spreadsheet, is far simpler than a workflow that pulls from a CRM, cross-references an inventory system and pushes results into an accounting platform. Each additional system usually brings a new API, new authentication and new failure points to cover.

What state is the data in?

Clean, structured data takes little work. Data scattered across PDFs, inconsistent spreadsheets, scanned documents or systems with poor export options adds real work to a project, often more than the automation logic itself.

Does it run unsupervised, or with a person in the loop?

A tool that assists a person (drafts something for review, flags anomalies for someone to check) is simpler than an autonomous system that makes decisions and acts without anyone checking the output. Unsupervised systems need more testing, more error handling and more safeguards before they can be left running on their own.

What types of project are there, from least to most complex?

There are no amounts here. This is a way to place your case.

  • A single, well-defined task automated end to end. For example, generating and sending a recurring report, or classifying and routing incoming requests. The scope is narrow and there are usually one or two systems involved.
  • A broader workflow connecting several systems. It includes some data cleaning and business logic in between. The complexity grows with every integration and every edge case.
  • A more complete integration. For example, automating a whole department's workflow across several tools, with ongoing decision logic and proper error handling. It looks more like a small software project than a quick script.

Team training is a separate case. The aim is to teach your team to use AI tools well, and what you receive is knowledge rather than a specific automation.

A project can move from one type to another once we look at the process in detail. That is why we don't attach amounts to each type.

Why doesn't Dyvab publish a price list?

Because a list would hide the differences that matter. There are no closed packages at Dyvab. Every project is designed for the process in front of us, because two automations that sound similar on paper ("automate our invoicing") can differ a lot in actual work, depending on the systems involved and the state of the data.

We work on a fixed price per project, agreed before we start. That means:

  • You know the total cost before any work begins.
  • The aim is to deliver something that works, without piling up hours.
  • Maintenance or support after delivery is optional and offered at a reduced price.

You also receive full documentation and ownership of the system. And if your case doesn't fit what we do, we tell you.

We can't give you a number on a web page. We can look at your specific process, tell you what makes it simpler or more complex, and give you an estimate.

How do you get an accurate estimate?

Describe the process as it happens today, including the messy parts. Tell us which systems it touches, how the data currently moves between them and what "done" looks like for you. That is enough for a first estimate, which we refine once we look at the process in more detail. Write to us and we start there.