Requesting a report
Then waiting until someone has time to build it. And then the follow-up question arrives.
For organisations using AFAS
Just ask your own numbers.
AFASAsk answers questions about your AFAS administration in plain language: revenue, margin, stock, receivables. With the calculation shown, so you can see where every figure comes from. And the routine work is done before you get to the office.
- Read-only connection
- Servers in the EU
- Data processing agreement
- No training on your data
Which installers are ordering less this quarter than last year?
7 installers ordered at least 20% less in Q3 2026 than in Q3 2025. Together that is € 86,400 less revenue. The four largest drops:
| Installer | Q3 2026 | Change |
|---|---|---|
| Montagebedrijf Hoeve | € 28,900 | −53% |
| Poorttechniek Noord | € 31,100 | −36% |
| Installatiegroep Delta | € 27,600 | −30% |
| Hekwerk Service Oost | € 25,800 | −24% |
Last night 06:12
Your morning overview is ready
4 orders drafted, 1 price list processed, 5 technical questions answered. Nothing sent.
Example with invented companies and amounts.
In production since 2025at a Dutch non-profit organisation using AFAS
Plain languageno report builder, no formulas, no exports
Every answer verifiablewith source, period and calculation
The problem
The answers are already in AFAS. Getting to them takes time.
You want to know where things stand. But every new question becomes a job for someone else.
Exporting to Excel
Building pivot tables yourself and hoping the filters are right. Every month again.
The dashboard
Fine for the standard figures. But it answers last month's question, not today's.
With AFASAsk you type the question. You get the answer, the table and the calculation. No waiting.
See it with your own questions
Use cases
Questions you can ask tomorrow.
For management, for sales and the back office, and for finance. Everyone with the same figures and the same definitions.
Directors and management
Management information without waiting for a report.
- How is gross margin developing per product group this year?
- Which customers are at risk of leaving?
- Which products cost us money through heavy support or returns?
- Is our growth coming from new or from existing customers?
Wholesale and technical trade
For sales, back office and purchasing.
- Which stock has not moved for over a year, and what is it worth?
- Which installers are ordering less than last year?
- Which open quotes over € 5,000 have been pending for more than 30 days?
- Which items run out within four weeks at current usage?
Finance and control
Quick answers to the question behind the figure.
- Which debtors pay later than 45 days on average?
- How do costs per cost centre compare with last year?
- Which general ledger accounts deviate noticeably this quarter?
- What is our outstanding receivables position by age?
Alongside your dashboards, not instead of them. Keep Power BI or your standard reports for the figures you want to see the same way every month. Use AFASAsk for the question that comes next.
Verifiable
No black box. Every figure can be checked.
AFASAsk does not guess; it calculates on your own AFAS data. With every answer you see which data was used and how it was calculated.
- Your terms, defined onceWhat you call “margin”, “active customer” or “installer” is recorded in your own knowledge base. Everyone calculates with the same definition.
- The calculation includedSource, period, filters and the database query. Your controller can recalculate it.
- Honest when something is missingIf it is not in the data, AFASAsk says so instead of inventing a number.
-- Revenue per installer, Q3 2026 against Q3 2025
SELECT debiteur,
SUM(bedrag) FILTER (WHERE jaar = 2025) AS last_year,
SUM(bedrag) FILTER (WHERE jaar = 2026) AS this_year
FROM verkoopfacturen
WHERE kwartaal = 3
AND debiteurgroep = 'Installateur'
AND factuursoort <> 'Creditnota'
GROUP BY debiteur
HAVING this_year < 0.8 * last_year
ORDER BY this_year - last_year;
Routine work
While you slept, your agent did this.
Not just answers to questions. PitchAI agents also do the recurring work that comes back every day. You define once how it should be done; after that it happens automatically, and your people decide.
Orders from the mailbox
Four orders that arrived by email or PDF are ready as draft sales orders in AFAS.
A new supplier price list
The price list PDF has been read. Purchase prices are updated and sales prices calculated with your standard margin, ready for review.
Technical questions from installers
Five questions answered from your own manuals, with the page number. Your technical team checks along.
Done. Three items for you.
Nothing has been sent. You approve, adjust or reject. Everything is in the log.
Awaiting your approval
Define it once. Then it runs every day.An example night. You decide which tasks are done.
Discuss your routine workGrows with the PitchAI Agent Engine.
AFASAsk is one agent. With the Agent Engine you get a team of agents that work across the business, monitored and supported by PitchAI.
Reacts as soon as something arrives
An email, an invoice, a form: the agent gets to work right away. No fixed check-in times needed.
Call your agent
A question about orders or stock while on the road? Call your agent and get an answer, in an ordinary phone call.
Work overview and log
Every task with status, exceptions and evidence. You see what was done, and why.
Knowledge stays yours
Corrections and agreements go into your own knowledge base, not into a vendor's chat history.
How it works
Start safe. Start small.
At first AFASAsk only reads. Only when the answers have been right for weeks do agents start making drafts. You decide at every step.
Get acquainted
A half-hour demo with your own questions, on example data. Together we look at where the gains are.
Connect, read-only
We connect AFAS through the API connection AFAS provides for this, with read rights only. Nothing changes in AFAS.
Set up with your terms
We record your definitions and test with real questions from your team until the answers are right.
Extend with routine work
Agents make drafts that your people review. Writing to AFAS only with your approval, task by task.
1 · Read only2 · Drafts3 · Routine work
Security and privacy
Your data stays yours.
Your administration is confidential. That is why we start read-only, and put the agreements on paper.
Read-only connection
AFASAsk reads your data and changes nothing in AFAS. Write access only comes up if you later want it for a specific task.
Servers in the EU
The application and your AFAS data are stored on PitchAI servers at Hetzner in Finland, within the European Union.
Data processing agreement
PitchAI processes your data on your behalf and records this in a data processing agreement. You know who does what with your data.
No training on your data
We use business AI models under business terms. Your data is not used to train models.
Sign in with your own account
Employees sign in with their own AFAS account. You decide who gets access.
Privacy filter In development
An option that replaces names, addresses, email addresses and IBANs with pseudonyms on our own server, before anything goes to the AI model.
In practice
We do not name customers without permission. We are happy to tell you more in an introductory meeting.
AFASAsk has been in production since 2025 at a Dutch non-profit organisation. Questions about relations, finance and projects go straight to its own AFAS data.
- ConnectionAFAS data through a fixed read-only connection, refreshed automatically
- KnowledgeIts own knowledge base with the organisation's terms and agreements
- AccessIts own domain; employees sign in with their AFAS account
Packages
Start with questions. Grow into routine work.
Every organisation is different. After the introduction you receive a proposal that fits your team and your AFAS environment.
Introduction
See what AFASAsk does with your questions.
- Demo with your own questions, on example data
- Advice on where best to start
- No connection needed
Free no obligation
AFASAsk
Questions to your own AFAS data, for management and team.
- Read-only AFAS connection, refreshed automatically
- Knowledge base with your terms and definitions
- Setup and testing with your team
- Monitoring, updates and support
Price on request per month
AFASAsk + Agent Engine
Agents that also do the routine work.
- Everything in AFASAsk
- Routine tasks: orders, price lists, technical questions
- Reacts to incoming email and documents
- Call your agent, with work overview and log
Price on request per month
Privacy filter for personal data
Shields names, addresses and other personal data before anything goes to the AI model. An add-on to any package, now in development.
AI usage is passed on at cost, without markup. Beyond that you pay for the product, the monitoring, the support and the advice.
Is AFASAsk a product of AFAS?
No. AFASAsk is an independent product of PitchAI. It works with your AFAS environment through the connections AFAS provides for this. AFAS Software B.V. did not develop, certify or endorse AFASAsk.
Can AFASAsk change anything in AFAS?
No. The connection is read-only by default. Agents that later do routine work first make drafts that your people review. Writing to AFAS only comes up when you approve it, task by task.
Does my data leave the EU?
Your AFAS data is stored on PitchAI servers in the EU. To answer a question, the question and the necessary intermediate results go to a business AI language model. That processing may take place outside the EU, under the provider's business terms. Your data is not used to train AI models.
Want personal data never to reach the language model? For that we are developing a privacy filter as an option. We are happy to discuss whether you need it.
What if an answer is wrong?
With every answer you see which data was used and how it was calculated, so you can check it. If something is not in the data, AFASAsk says so. If a definition needs to be different, we record it in your knowledge base. From then on everyone uses it.
Do my people need to learn something new?
Hardly. You type a question in plain language, as you would ask a colleague. We set up AFASAsk together with a few regular users, so they can help their colleagues get started.
How is this different from ChatGPT or Copilot?
A general chatbot knows nothing about your administration unless you paste data into it yourself. AFASAsk is connected to your AFAS data, knows your terms and shows how it calculates. With the Agent Engine, agents also do work themselves: at set times or as soon as something arrives, with a log and monitoring by PitchAI.
Does it work alongside Power BI?
Yes. Keep your dashboards for the standard figures you want to see the same way every month. AFASAsk is there for the question that comes next, or the question no report exists for yet.
How quickly can we start?
We usually schedule a demo within a week. We then agree the connection and setup with you. That depends on your AFAS environment and the questions you want to ask.
What does AFASAsk cost?
The price depends on the number of users, the size of your AFAS environment and the routine tasks you want done. After an introduction you receive a tailored proposal. AI usage is passed on at cost, without markup.
Request a demo
See it with your own questions.
Tell us which questions are on your mind. In a half-hour demo we show how AFASAsk answers them, on example data. No obligation.
- info@pitchai.net
- +31 6 4047 8645
- Seth van der Bijl, PitchAI