AFAS AI
AFAS Jonas: which AI features do you get in Profit and in SB?
Find out what AFAS Jonas does in Profit and SB. Compare workflow help, summaries and the help assistant in SB, including the necessary human oversight.
Jonas is the collective name for AI at AFAS. In Profit, AFAS describes assistants and workflow automation, among other things; in SB, the AI page emphasises software help. See what the documentation supports and how to try out a feature in a verifiable way.
Jonas is a collective name for AI at AFAS
Anyone searching for AFAS Jonas will come across various uses. The name alone says too little about the task, the set-up required and the product in which a feature is available. So distinguish between AFAS Profit and AFAS SB.
Start with a work question: do you want an explanation of how to use the software, to write a response to a file or to have data from a request processed? These tasks call for different sources and checks. A help assistant can be useful without access to your own administration. A workflow that fills in fields, by contrast, has consequences for the rest of the process.
What does AFAS describe for Profit?
The AFAS Help Center on Jonas describes, among other things, two uses in workflows. Jonas automation analyses workflow data and can fill fields according to the set-up in the workflow editor. A Jonas assistant can automatically run a prompt at a chosen point; the user sees the result in an information block.
The general AI page of AFAS also describes help with responses, summaries and translations, and assistants for which you select your own sources. Think of sources such as a manual or an FAQ. These descriptions are no guarantee that all features are already active in your environment.
It is therefore incorrect to see Jonas solely as a chat window in which every action waits for a separate click. Some tasks can be an automatic part of a workflow. Have the administrator point out where input is used, which fields can change and where employees check before a process continues.
What does AFAS describe for SB?
On the AI page of AFAS SB, Jonas is described as help with using the software. The assistant uses the help documentation and shows sources with the answer. That suits questions such as where to find a setting or how to carry out an action.
It does not follow that Jonas independently runs your entire administration. Nor does it say which other automatic accounting features SB has. Keep software help and financial processing separate when you request a demonstration. Also look at the software profile of AFAS SB to compare the package as a whole.
Compare features by task and source
| Topic | Documented for Profit | Documented on the SB AI page |
|---|---|---|
| Help and text | Including responses, summaries and assistants | Answers to questions about using the software |
| Sources | Own sources can be selected for assistants | Help documentation, with source references |
| Workflow | Automatic prompts and filling fields after set-up | This page does not support such workflow features |
| Question for the demo | Which actions and fields are set up in our environment? | Can a user find the explanation and carry it out? |
A feature missing from this table means that the source cited does not support it. It is not a statement that the feature does not exist anywhere in the product. Ask the supplier for additional documentation if a task is important for your choice.
An example task with a verifiable result
The example below was written by us. We have not carried out this task in Jonas. You can use it to assess the quality of a summary in a demonstration.
Project Koraal receives a draft schedule on 18 November. The project manager discusses it with the client on 20 November. The budget has not yet been approved. The client also asks whether delivery a week earlier is possible; nothing has been agreed about that yet.
Give this task with it:
Summarise the agreements in a maximum of three points. State the two dates with the corresponding action. List open questions separately. Do not add any new agreement or confirmation.
The check consists of four questions: are both dates correct, is the missing budget approval visible, does the earlier delivery remain a question and have no new commitments been added? A neat wording is not enough if a wish appears as an agreement.
Repeat the task with a contradictory date in a second message. Ask how the assistant deals with that contradiction and whether you can trace the source used. Note correction time alongside writing time. That way you assess whether the tool actually saves time in your work.
What must an administrator arrange in advance?
- Availability: record which Profit version, feature and activation the demonstration uses. Check this in the current documentation for your environment.
- Sources: choose an owner for manuals and other documents. Remove outdated versions before an assistant works with them.
- Access: try a question with an ordinary user role. Check which sources and results that role can open.
- Automatic actions: show which fields can change and how you correct an incorrect result. Appoint someone to manage the set-up.
- Data use: discuss which input is processed and stored, which parties are involved and how that fits your own information policy.
Use fictitious data for the first trial. That way you can assess how it works before you decide which real files are suitable. Then start with one clearly delimited task for which an employee can easily check the result.
Do not let the choice between Profit and SB depend on AI alone
A company that only needs accounting software does not choose an extensive ERP system because an assistant knows more kinds of tasks. First compare the ordinary business processes, users and costs. Then assess AI as part of that package.
If you already work with Profit, investigate which existing workflow is suitable for a limited trial. If you are looking at SB, have the help assistant answer a question that a new user really asks. For a broader selection, compare accounting software or ERP systems. Our guide to accounting with AI helps with the assessment of financial tasks.


