No number pulled from thin air.An honest one instead.
We could print a starting price here that is right for almost nobody. It is more useful to be open about what drives the price and what you can try at no risk. The consulting formats have fixed prices – and a money-back guarantee.
- Fixed prices for consulting
- Money back if it does not help
- No obligation to buy
What drives the price of fabular
There are essentially three. If you know them for your own business, you can estimate the order of magnitude yourself.
How many people work in it
The number of people actually working in the system – not your headcount. A warehouse with five scanner stations counts differently from fifty office workplaces.
Which modules you need
Nobody uses all of them. If you only need purchasing, warehousing and invoicing, you do not pay for production and point of sale. The industry solution sets the sensible core.
How much has to be adapted
The biggest lever. In our experience around eighty per cent is covered by the CORE and the industry solution. What goes beyond that is the part that genuinely costs effort.
AI processes are billed by consumption
An assistant that reads documents, checks reorder proposals or flags deviations consumes computing power – and a different amount from case to case. That part is therefore not licensed as a flat rate but billed with credits: a balance you buy up front that decreases with every run. What an assistant costs depends on its performance tier, and you set that.
- Credits apply to autonomous and semi-autonomous processes – fully automatic or with approval by the person in charge
- Consumption per assistant, department and period visible at any time
- A cap can be set per assistant: once the allowance is used up, the process stops instead of carrying on
- Asking fabAI questions in day-to-day work is included in the licence and consumes no credits
A performance tier per assistant
Low, Standard or Performance – set per assistant and changeable at any time, without rebuilding the process.
Consumption instead of a flat rate
You pay for the runs that actually happen. An assistant that only works at month-end only costs you then.
No consumption without cover
Credits are bought up front. Without a balance no autonomous run starts, which rules out costs quietly spiralling.
Three tiers, the same task
Every assistant runs in one of three tiers. They differ in how much context and computing power a run receives – and therefore in the credits consumed per run. The tier is a setting, not a licence question: what runs in Standard today can run in Performance tomorrow.
Low
- 1.For clearly defined tasks in high volume
- 2.Assigning documents, extracting fields, classifying text
- 3.Lowest consumption per run
Worth it when the same simple decision is made thousands of times.
Standard
- 1.The regular mode for most assistants
- 2.Checking reorder proposals, assessing deviations, drafting replies
- 3.A balanced ratio of consumption and quality of result
The default. Leave it alone and you work here.
Performance
- 1.For cases where a mistake is expensive
- 2.Multi-step processes, many data sources, longer chains of reasoning
- 3.Highest consumption per run, in exchange for the most dependable results
Use it deliberately: for the annual accounts, not for filing documents.
Consulting: here the numbers are clear
The two consulting formats have a fixed entry price and a written result. They work whether or not you buy an ERP afterwards.
Digitalisation consulting
For businesses that know something is not running smoothly, but not exactly where.
- Map the IT landscape and the workflows
- Develop process models with the people who run them
- Action plan with effort and sequence
- A document you could work from without us
AI consulting
For businesses where AI is on the agenda but the viable use case is missing.
- Sort use cases by data quality and impact
- Assess data quality – the most common reason pilots fail
- Clarify governance, data protection and the EU AI Act
- Pilot recommendation with effort estimate and stop criteria
The same applies to both formats. If the result does not help you, we refund the fee. No justification required, and the document we produced stays with you.
How a reliable number comes about
From the first question to a quote usually takes two to four weeks – and costs you nothing along the way.
- Step 01
First conversation
30 to 60 minutes, free of charge
You describe the situation, we say openly whether fabular is the right lever for it. If it is not, we say that too.
- Step 02
Scoping
1 to 2 weeks
We work out which areas the system should cover, how many people will work in it and what is carried over from the legacy system.
- Step 03
Order of magnitude
in conversation
You get a range before anyone has invested effort. That is enough to decide internally whether to continue.
- Step 04
Quote
1 to 2 weeks
A quote with line items, assumptions and what is explicitly not included. The last part is what otherwise makes projects expensive.
About pricing and costs
Why is there no starting price here?
Because it would be right for almost nobody. A business with eight workplaces and a trading company with two hundred are orders of magnitude apart. A number that fits one case scares off the other or creates false expectations. You get the range in the first conversation, before any effort is spent.
What does a consulting engagement actually cost?
Digitalisation consulting starts at €1,400, AI consulting at €2,000. The final price depends on scope and is fixed before we start. The money-back guarantee applies to both.
Are there running costs?
Yes. As with any business software there are running costs for use, maintenance and support, plus infrastructure depending on how you operate it. Running in your own data centre or as SaaS with us shifts these costs between line items. We work both through in the quote so you can compare.
What are credits and when do I need them?
Credits are a balance for AI runs. You need them as soon as an assistant works on its own – processing documents, checking reorder proposals or flagging deviations, whether fully automatically or with approval by the person in charge. Anyone using fabAI only in dialogue, asking questions themselves, needs no credits: that is included in the licence. The balance is bought up front, consumption is visible per assistant, and a cap can be set for each one.
What separates Low, Standard and Performance?
The performance tier determines how much context and computing power a single run receives – and with it the credits consumed. Low is meant for simple tasks in high volume, Standard is the regular mode and the default, Performance for cases where a mistake gets expensive. The tier is a per-assistant setting and can be changed at any time. So you can run the same process cheaply and step it up for the closing.
What does implementation cost on top of the licence?
Implementation is the larger item with any ERP. It covers assessment, configuration, data migration, interfaces and training. How large it turns out depends mainly on how clean your master data is and how far you deviate from the standard. We look at both beforehand rather than discovering them during the project.
Does it pay off for a small business?
Sometimes it does not, and then we say so. fabular pays off where reporting obligations, batches, recipes or multiple sites are involved – not as bookkeeping with a warehouse function. For a pure office operation there are cheaper routes, and we will point you to them.
Do I get a discount?
We do not work with list prices that percentages are then taken off. The quote is what it costs from the outset. If the price does not fit, we talk about scope – not about a reduction on the same scope.
Ask for the order of magnitude
Describe your situation briefly. You get a reliable range before anyone has invested effort – and no follow-up sales email.