search
search cart 0

wals.pro AI 4 weclapp

AI for your weclapp via MCP: with Claude, ChatGPT and Copilot

Take AI out of the text window and into your weclapp.

wals.pro AI 4 weclapp connects AI assistants such as Claude, ChatGPT or Microsoft Copilot Studio to your weclapp tenant via MCP, hosted in the EU. MCP (Model Context Protocol) is the open standard AI assistants use to access other systems. Answers from your ERP in minutes, every write action with preview, approval and log. Setup in 0.5 to 3 hours, with the AI assistant your team already uses.

wals.pro AI 4 weclapp comes from wals.pro — official weclapp partner with daily project practice.

In productive use, permanently free entry tier, no credits. Read first, write later.

Start for free
AI for your weclapp via MCP: with Claude, ChatGPT and Copilot
Here's what it looks like

Everything by chat: every change with preview and approval, every action traceable.

Claude
Create a quote for Muster Elektronik with the three items from the last order.
✦Quote #4711 created. 3 items taken over from order #3982. Ready for review.
Approve Reject
Quote #4711
Muster Elektronik GmbH
Approved ✓
  • 3x Cat6 network cable, 5 m €89,00
  • Onboarding package Basic €1.200,00
  • Setup & training (2 days) €980,00
Total €2.269,00

Illustrative depiction of the process. Every action requires your explicit approval before it is written to weclapp.

What customers say

The game changer for Weclapp!

Kurt Liebenberg, Managing Director, BQM GmbH

I'm impressed by how neatly and cleanly quotes, orders and purchase orders can be created and managed with the AI agent. It's an enormous relief for my employees.

Marcel Sohn, Head of Sales D/A/CH at SEIKOM-Electronic

I bow deeply, this is outrageously good.

Karl Heinz Hermann, Managing Director, Das Gesunde Tier

It was insane how fast that went.

Martin Spieker, Founder & Director at Potluck

Your team already uses AI, but not where your money lives.

An AI assistant without a connection only sees what you copy into it. With wals.pro AI 4 weclapp, it works with your real data: reading instantly, writing only with your approval. AI alongside the work or AI inside your weclapp: the difference in five lines.

  • Before: You export lists and paste them into the chat window. Status: unknown.
    With wals.pro AI: Your AI asks weclapp directly. Status: now, source: your tenant.
  • Before: Reordering by gut feeling or after the last walk through the warehouse.
    With wals.pro AI: Purchase proposal based on stock, coverage and open orders. The purchase order comes as a preview.
  • Before: Matching payments to open items by hand, invoice by invoice.
    With wals.pro AI: Review the matching proposal, approve, matched, with an entry in the audit log.
  • Before: Drafting a quote in the chat, then creating every item in weclapp by hand.
    With wals.pro AI: Quote from the meeting note as a preview. Customer, items, prices checked.
  • Before: Afterwards, nobody knows who asked or changed what.
    With wals.pro AI: Roles per person, approval per action, audit log per call. Everything is traceable.

Ask your weclapp: start for free

Set up wals.pro AI 4 weclapp via MCP: 0.5 to 3 hours

No project, no new tool: you need your weclapp access, your AI assistant and, depending on scope, 0.5 to 3 hours. Our recommendation: start read-only and enable write actions only once your team knows the previews — via roles with templates per area.

Step by step in the help center →

Dashboard of wals.pro AI 4 weclapp with tenant, plan and a Getting started checklist

1 · Register and connect weclapp

Register at ai.wals.pro, wait for activation, enter your weclapp URL and API key. Hosted in the EU, separated per tenant. Nothing needs to be installed in your weclapp.

Connections in wals.pro AI 4 weclapp: connect AI assistants such as Claude, Cursor and VS Code via OAuth

2 · Connect your AI assistant via OAuth

Claude, ChatGPT, Microsoft Copilot Studio, Cursor and other AI clients are supported. Connect once via OAuth 2.1. Your AI subscription stays your contract.

User management in wals.pro AI 4 weclapp with the roles admin and viewer per team member

3 · Start read-only, extend permissions later

Start read-only: revenue per month, open items, stock per warehouse. You grant write rights later via roles — every write action comes as a preview first, and 19 skills, task-based guides from quotation to purchasing, guide your assistant.

View skills and house rules →

What your team starts with today

Five decisions, from problem to result. Reading instantly, writing only after your approval: ask, review the preview, approve. The result is in weclapp, the path to it in the log. All free to try from today →

Icon: Inventory planning and stock

Inventory planning & stock

Twelve articles will run out in three weeks, but nobody has time for the coverage list today. The AI shows stock, coverage and open orders per article — and the order proposal per supplier as a preview. After your approval, the purchase order is in weclapp: 12 items, one supplier, logged. If goods arrive without a purchase order, the AI books the goods receipt with supplier and delivery note number — after your approval.

Icon: Revenue and margin

Revenue & margin

The monthly report needs revenue per customer and a year-over-year comparison. Until now that meant export, Excel, an evening gone. The AI totals invoices per customer and period directly in weclapp and shows the top 10 with change. The figure is in the chat, no file. Read-only, no approval step, but in the audit log.

Icon: Payments and open items

Payments & open items

€3,480 has arrived in the account — no invoice number in the payment reference, only the customer name. The AI finds the matching invoice by amount and customer and shows the allocation as a preview: due date, remaining amount, open item. After your approval: payment allocated, open item closed, entry in the log.

New invoices from a quotation or sales order come as a draft; they are only booked as an open item after your review and approval.

Icon: Quotes and orders

Quotes & orders

After the customer meeting, three items and a discount sit in your notes — the quotation is still missing. The AI resolves customer and articles and shows the quotation field by field: items, prices, discount, validity. You approve — the quotation is in weclapp with a number; without approval, nothing happens. Once entry is complete, you get the link to the quotation page, which your customer opens without an attachment and without weclapp access.

Icon: Master data and data quality

Master data & data quality

48 articles without a price, 30 customers without an email address — each of them will slow down an invoice or delivery at some point. The AI lists the gaps and shows every correction you specify as a preview — field by field. You close email gaps one at a time; everyday fields such as the responsible person for customers or task dates you change in bulk: before/after per record, one approval, logged.

Drei Beispiele für deinen Einstieg mit ERP und KI

Die folgenden Abläufe sind Beispiele, keine Kundenberichte oder Erfolgszusagen. Du brauchst einen freigeschalteten Zugang zu wals.pro AI 4 weclapp, deinen verbundenen KI-Assistenten und einen weclapp-API-Zugang mit passenden Rechten. Welche Daten die KI sieht, hängt von diesen Rechten und deiner Rolle ab.

Vertrieb: aus Gesprächsnotizen eine Angebotsvorschau

Deine Eingabe: Kund:in, gewünschte Artikel, Mengen und vereinbarte Konditionen. Die KI liest die passenden Kunden- und Artikeldaten in weclapp und bereitet ein Angebot vor. Deine Prüfung: Stimmen Kund:in, Positionen, Preise und Rabatt? Erst nach deiner Freigabe wird der Beleg angelegt. Das Ergebnis und der Aufruf lassen sich nachvollziehen.

Grenze: Mehrdeutige Kunden- oder Artikelangaben musst du klären. Eine angelegte Angebotsnummer beweist weder den Versand noch die Annahme durch deine Kund:innen.

Geschäftsführung: offene Posten als Arbeitsliste

Deine Eingabe: Stichtag, gewünschte Kund:innen und die Frage nach überfälligen offenen Posten. Die KI liest die verfügbaren Forderungs- und Kundendaten und stellt eine Liste mit Restbetrag und Fälligkeit zusammen. Du prüfst Zeitraum, Filter und die zugrunde liegenden Belege.

Grenze: Das ist eine lesende Auswertung. Sie bucht keine Zahlung und versendet keine Mahnung. Fehlende oder verspätet zugeordnete Zahlungen können das Ergebnis verändern.

Service: Ticketstand prüfen und Kommentar vorbereiten

Deine Eingabe: Ticketnummer und der bestätigte neue Bearbeitungsstand. Die KI liest das Ticket und bereitet deinen Kommentar als Vorschau vor. Du kontrollierst Inhalt und Ziel; erst nach deiner Freigabe wird geschrieben.

Grenze: Ein interner Kommentar und eine Antwort an Kund:innen sind verschiedene Aktionen. Empfänger:innen und mögliche Benachrichtigungen müssen zum gewünschten Schritt passen. Unklare Ausführungsergebnisse zuerst am Ticket prüfen, bevor du erneut schreibst.

Was wir für dein Team einrichten

Du kannst selbst starten. Wenn du Begleitung brauchst, klären wir mit dir den ersten Anwendungsfall, Zugänge, Rollen und Prüfkriterien. Gemeinsam testen wir den Ablauf mit geeigneten Beispieldaten und vereinbaren, wie dein Team Vorschauen prüft und Fehler meldet. Einrichtung, Schulung und laufende Betreuung sind getrennte Projektleistungen; der Plattformzugang allein umfasst sie nicht.

ERP- und KI-Einstieg besprechen · Weitere Systeme anbinden · Regelmäßige Abläufe umsetzen lassen

Your first five questions

Sorted by role, ready to copy: phrase them in your own words. Almost all are read-only; where a preview is created, it says so. The answers come from your tenant, and nothing is executed until you approve.

Icon: Management

Management

  • Show me revenue per month for the last twelve months.
  • Who are our top 10 customers by revenue this year?
  • Which open items are overdue: total per customer, sorted by age?
  • What is the stock value per warehouse as of today?
  • How many open tickets does each person in charge have right now?
Icon: Sales

Sales

  • List all open quotes per customer with amount and date.
  • Which quotes are older than 14 days with no response?
  • Compare revenue per customer with the previous year: who has grown, who hasn't?
  • Which active items have no sales price?
  • Create a quote from this meeting note and show me the preview first.
Icon: Purchasing and warehouse

Purchasing & warehouse

  • Which items are below the reorder level, with coverage in days?
  • Give me a purchase proposal per supplier for the next four weeks.
  • Which goods receipts were booked this week?
  • Which deliveries have been shipped but have no tracking number?
  • Show me the stock value per warehouse and the ten most valuable items.
Icon: Accounting

Accounting

  • Show me overdue receivables by age: 1–30, 31–60, over 60 days.
  • Which incoming payments have not yet been matched to an invoice?
  • Which supplier invoices are still in status “new” and waiting for review?
  • Which outgoing invoices are open and due in the next seven days?
  • What is the total of open supplier invoices per supplier?
Icon: Service

Service

  • Show me open tickets per customer and priority.
  • Which tickets have had no reply for more than 48 hours?
  • Which recorded hours have not yet been billed, per customer?
  • Which contracts expire in the next 90 days?
  • Write a comment on ticket 4711 with today's status, as a preview first.

Over 125 companies have registered for wals.pro AI 4 weclapp.

No demo magic, but controlled work in live operation — with the AI assistants your team already uses. See it in your own weclapp — start for free →

125+
Companies registered for wals.pro AI 4 weclapp

Registrations: our own, rounded figure. Every write action goes through preview and approval, every call is recorded in the audit log.

Who can do what? Five questions, five answers.

For us, control is not a setting you discover later — it is the path every action takes. EU hosting (europe-west1) · strict tenant isolation · OAuth 2.1 · deletion only with preview and approval · your AI provider stays your contract.

All security details on the trust page →

Questions about security or compliance? Book a meeting →

01

Who can read?

Viewer, Editor and Admin roles plus area templates such as Sales or Accounting. Every person has a role that determines which data can be read and which actions are allowed.

02

Who can write?

Only those whose role includes the action. Every person has exactly one role. Invited members start as Viewers (read-only); whoever sets up the tenant is Admin. Admins can also grant just the maintenance of custom fields without releasing all fields of a record. Every change runs only after a preview and your approval.

03

What exactly gets changed?

Every write action comes as a preview first — field by field, item by item. The preview also names consequences in weclapp, such as an invoice email your rules would trigger when an invoice is finalized. Nothing runs without asking.

04

Who approves?

You do, with a one-time approval bound to the payload via SHA-256. It applies to exactly this action with exactly these values, and never again afterwards.

05

What is in the log?

Every call in the audit log: who, when, which action, which result. If each person stores their own weclapp access, the change also appears in weclapp under their name and their weclapp rights apply. Export from Pro.

Before an AI writes to weclapp: 6 checkpoints

Check these six points for any AI you give write access to weclapp, whoever the provider is. This is how wals.pro AI meets them:

  1. Permissions: The AI works only with the permissions of the weclapp user whose API key you store, and roles in wals.pro AI can only narrow them, never widen them.
  2. Preview: Every write action, whether create, update, delete or status change, comes as a preview first, and the preview itself changes nothing in weclapp.
  3. Approval: The server executes only the exact preview you saw, the approval works only once and only briefly, and the server rejects it if any value changes.
  4. Audit log: Every call is logged (who, when, which action, which result), without the contents of your records; how much of it you can view and export depends on your wals.pro plan.
  5. Data validation: Before every preview, validation runs against a fixed contract per data type that rejects unknown fields and, on create, missing required fields.
  6. Test first: Start read-only and try changes in a separate weclapp test system with its own credentials before you grant write access in your live system.

What a writing AI in weclapp does not do alone

Particularly sensitive changes, such as to supplier tax fields, are never written on your approval in the chat alone. An admin decides them in the dashboard. The AI can never grant admin rights or change plan and billing. It creates invoices as drafts and does not finalize them in the same step.

All six checkpoints in detail are on the page MCP server for weclapp.

The guide AI that writes in weclapp explains why AI that writes to weclapp needs a preview, approval and an audit log, and which approaches exist.

New in wals.pro AI

What has been added since launch. Is something missing, or do you have an idea? Write to us at support@wals.pro or report it directly in the AI chat. Every piece of feedback is reviewed and answered, and handling and implementation are free. We decide which requests we implement.

Illustration: Record card with a custom attribute maintained via a chat bubble (AI-generated)

Read and maintain custom fields

Your custom fields map your process. The AI now reads them with field names and option labels instead of internal IDs and maintains them on customers, articles, documents, tickets and tasks, including document line items such as sales order items.

Every value is validated against its field type, and if a required field is empty, the AI asks before writing. As with every change, there is a preview first; nothing is written until you approve.

Try it for free now →

Illustration: chat speech bubble and PDF quote next to the wals.pro brain

Revise quotes instead of rebuilding them

Until now, a changed quote had to be created from scratch, and the old one was left unused. Now you can adjust the items of an open quote directly in the chat: change a quantity, add or remove an item. The AI shows exactly what will change beforehand and only writes after your approval.

For documented revisions, the AI uses weclapp's own versioning: same quote number, new version, and the previous one stays traceable.

Try it for free now →

Illustration: item list in a browser window with a chat speech bubble for replenishment in chat

Inventory planning & purchase proposals: now live

Inventory planning directly in the chat. The AI knows stock, incoming goods and open demand per item, warehouse by warehouse, pre-order warehouses included. It suggests order quantities based on your own rules (minimum stock, pack size, minimum order quantities) and creates draft purchase orders grouped by supplier from them, of course only after your approval.

This means you no longer need a separate planning tool alongside the ERP for purchase proposals.

Try it for free now →

Illustration: Three record cards changed together via a chat bubble and confirmed with a check mark (AI-generated)

Change several customers and tasks in one step

Reassigning the responsible person for several customers or moving a series of tasks to a new date used to mean reviewing and approving each change individually. Now such everyday changes to existing customers, suppliers and tasks go through one preview with one approval, as they already do for articles. The preview shows, per record, what was in it before and what will be in it afterwards.

Immediately before the first write, every record is checked against the preview once more: if one has changed in the meantime or is missing, nothing is written, and you learn which records are affected. The result is shown per record. Nothing is rolled back: if a record only fails during writing, the records already written stay changed. Nothing is created or deleted in the process, and you keep changing fields with consequences for documents, email sending, tax or payment one at a time. Meant for everyday changes, not for import or migration.

Try it for free now →

Illustration: Parcels in front of a warehouse shelf with a delivery note and a check mark, controlled via a chat bubble (AI-generated)

Goods receipt without a purchase order

When goods arrived without a purchase order, the only options so far were manual entry or a stock correction without supplier and delivery note. Now you give your AI assistant the supplier, warehouse and items, including serial numbers, and the AI books a regular goods receipt with supplier and delivery note number — only after the preview and your approval.

Try it now →

Where you notice it first

Four items where you notice it first: run the numbers with your own figures →

Icon: Time per quote

Time per quote

The conversation takes care of gathering context, building items and checking. You review the preview and approve.

Icon: Data preparation for import

Data preparation for import

CSV import files are prepared in the conversation so that weclapp imports them directly. During a weclapp onboarding, this saves the pure data preparation.

Icon: BI budget

BI budget

Reporting is included with wals.pro AI 4 weclapp. Many standard reports are simply answered by the chat, without you building your own BI setup for them.

Icon: Whole team, tiered permissions

Whole team, tiered permissions

Connect weclapp once, centrally — every team member gets a personal access with their own role (read-only or also write). This way, colleagues with ERP knowledge who don't sit in the ERP client every day can work with it too.

What does wals.pro AI 4 weclapp cost?

wals.pro AI is in productive use. Until 31.12.2026, all features are free for everyone. After that, the entry tier stays permanently free, with no credits: Free with 1 user, 500 read and 10 write actions per month. If you need more capacity, you choose a paid tier: Starter, Pro, Business and Enterprise on request. In the wals.pro support packages, access is permanently included from the Basic package. Not included: your AI subscription. You bring Claude, ChatGPT or Copilot Studio yourself.

Feedback and feature requests are not part of the pricing tiers: we handle them free of charge in every plan, and every request is reviewed and answered. Which requests we implement, we decide based on the impact for all customers.

Start for freeBook a call →

Is wals.pro AI 4 weclapp right for us?

The best fit: SMEs with weclapp who want to use AI where the numbers live, reading instantly, writing with approval. Recognise yourself on the left? Start for free →

Good fit

  • SMEs with weclapp: retail and e-commerce, manufacturing, service providers, logistics
  • Teams that already use an AI assistant (Claude, ChatGPT, Copilot) or plan to, and want to let it work with their real numbers
  • Organisations that want to introduce AI in a controlled way: one role per person, preview, approval, log

Not a good fit

  • Companies without weclapp
  • Anyone who wants an AI without an approval step that writes into the ERP on its own
  • Anyone looking for a finished dashboard instead of an assistant

Ask your weclapp the first question.

Register, connect, ask: in just a few minutes. Start read-only, enable writing later. Nothing needs to be installed in your weclapp. Until 31.12.2026, all features are free, and the entry tier stays permanently free, with no credits.

Start for free Book a call →
  • Start for free: all features free until 31.12.2026, no credits
  • Preview, approval, log for every write action
  • Works with Claude, ChatGPT and Copilot Studio
  • Direct exchange instead of a generic mass solution
  • Feedback & feature requests: handled free of charge

Request a use case

Briefly describe your weclapp context and the first question your team wants to ask. We'll get back to you with the right next step.

FAQ

The most important questions about wals.pro AI 4 weclapp. Yours is missing? Ask us directly →

Basics: Is there an MCP server for weclapp? · Guides: Help & setup

wals.pro AI 4 weclapp

What is MCP (Model Context Protocol)?

MCP stands for Model Context Protocol, the open standard an AI assistant uses to access another system. Claude, ChatGPT, Microsoft Copilot Studio and many other clients support it. wals.pro AI 4 weclapp is a hosted MCP server for your weclapp: you don't install or operate anything yourself, you simply connect your assistant to your weclapp via OAuth. Reading works instantly, and every write action comes as a preview first and only runs after your approval.

For the different ways to connect weclapp via MCP, see our overview Is there an MCP server for weclapp?

What does wals.pro AI 4 weclapp cost?

Until 31.12.2026, all features are free for everyone. After that, the entry tier stays permanently free, with no credits: Free with 1 user, 500 read and 10 write actions per month. If you need more capacity, you choose a paid tier: Starter, Pro, Business and Enterprise on request. In the wals.pro support packages, access is permanently included from the Basic package. Your AI subscription runs separately: you keep paying it directly to your provider. Start for free now →

Which AI tools does it work with?

Among others with Claude (claude.ai, Desktop, Claude Code), ChatGPT, Microsoft Copilot Studio, Cursor, VS Code, Codex, Notion, Langdock and GitHub Copilot CLI; Windsurf, GitHub Copilot, Mistral Vibe, Google Gemini and Gemini CLI are currently under review. You connect the assistant your team already uses via OAuth. wals.pro AI does not bring its own language model — it is the secured execution layer to weclapp. There is no separate chat front end to learn.

Can I find wals.pro AI 4 weclapp directly in Claude?

Yes, wals.pro AI 4 weclapp is listed in Claude's connector directory: open the entry, connect and sign in with your wals.pro AI account in which your weclapp is connected — you don't copy an address for this. This route connects your live system; for the demo system or another workspace, you add the system address from the Connections page as a separate connector, as before. The same rights apply as for any other connection, and every change only runs after a preview and your approval. In Claude Team or Enterprise organizations, the administration may first have to provide the connector.

Can the AI write to my weclapp?

Yes — 40 write actions: quotation, sales order, invoice, payment allocation, comment, document upload, goods receipt, shipping, time booking, master data, prices, purchase order. Never without asking: first a preview field by field, then your one-time approval — payload-bound via SHA-256 —, then execution and an audit entry. This also applies to deletion: always one at a time, with the record in the preview and only after your approval; every deletion is logged. There are no bulk deletions, and production orders can never be deleted.

Is it safe to give an AI write access to weclapp?

It depends on how the access is built. Check six points: permissions, preview, approval, audit log, data validation and test first. With wals.pro AI, the AI works only with the rights of the stored weclapp user, every change comes as a preview first, the server executes only that exact preview with a single-use approval, and every call is in the audit log. Some risk remains, for example from instructions hidden in free text in your data. So only give write access to people who can judge previews.

Ist ein MCP-Server für ein ERP sicher?

MCP selbst ist nur das Protokoll; sicher oder unsicher macht ihn, was der Server zulässt. Für ein ERP zählt: mit wessen Rechten er handelt, ob Änderungen eine Vorschau und eine an den Inhalt gebundene Freigabe brauchen, ob jeder Aufruf protokolliert wird, ob Daten vor dem Schreiben geprüft werden und ob du zuerst testen kannst. Frag jeden Anbieter nach diesen sechs Punkten und danach, wer den Server betreibt.

Can the AI write without my approval?

The server executes only after a preview and a single-use approval: every write action comes as a preview first, and only that exact preview is executed, once and only for a short time. Whether your AI client asks you as well is something you set in the client. Don't allow write actions permanently there, and read every preview. Particularly sensitive changes, such as to supplier tax fields, are decided by an admin in the dashboard, not in the chat.

What rights does the AI have in weclapp?

Exactly the rights of the weclapp user whose API key you store, never more. You can store one shared key or, depending on plan and settings, one per person; then everyone acts as themselves. The role in wals.pro AI can only narrow these rights, for example to specific write actions. An approval in the chat does not widen any rights.

Can I book in goods for which there is no purchase order?

Yes, if the sender is set up as a supplier in weclapp. You give your AI assistant the supplier, warehouse and items, including serial numbers; after the preview and your approval, a regular goods receipt with supplier and delivery note number is in weclapp. Stock corrections without a business partner, returns, credit notes and transfers between warehouses remain separate workflows.

Does the AI send invoices to my customers without asking?

No. The AI creates a new invoice as a draft and then stops until you have reviewed it in weclapp. Finalizing and booking are separate steps with preview and your approval; if a rule in your weclapp would trigger an invoice email in the process, the preview says so, and nothing is finalized without your explicit consent. The AI does not send invoices itself, and it creates invoices from a quotation or sales order from the Pro plan onwards.

What are my employees allowed to do?

You decide. Every person has a role: Viewer, Editor, Admin, an area template such as Sales or Accounting, or a custom role with its own action list. That way, Sales can create quotes but not change prices, and Accounting can match payments but not create orders.

Can it be read-only?

Yes, and that is exactly how we recommend starting. Reading covers searching and reporting across 172 weclapp entities — revenue, stock, open items, tickets. You enable write actions later, selectively via roles, per action.

How do I connect weclapp securely to ChatGPT or Claude?

Through an MCP server that you connect to your AI client via OAuth. With wals.pro AI, you enter your weclapp address and an API key in the dashboard; the key is kept in a secret store. Three decisions make it secure: choose a weclapp user with only the rights needed, start read-only and unlock write actions deliberately through roles, and don't allow write actions permanently in the client. For ChatGPT, check whether your plan allows write access via MCP.

How do I get started?

Register at ai.wals.pro, wait for activation, store your weclapp URL and API key, connect your AI client via OAuth (you find Claude directly in the connector directory), ask your first question. Start read-only, enable writing later. The first five questions per role are at the top of this page — ready to copy; 19 skills, task-based guides from quotation to purchasing, guide your assistant through the workflows. Start for free now →

How quickly are we ready, and does anything need to be installed?

Connected in a few minutes; setup takes 0.5 to 3 hours depending on scope — no project, no new tool. All you need is your weclapp access (URL and API key), your AI assistant and a little time. Nothing has to be installed in your weclapp, there is no new front end and no training hurdle.

Where is the data stored, and what about data protection and compliance?

Hosted in the EU (europe-west1), strict tenant separation, OAuth 2.1. Access runs through personal connections with roles and permissions per user, connected AI clients are visible and revocable, and every call is in the audit log. What your assistant reads is processed by your AI provider under your contract. wals.pro only operates the secured execution layer, not its own language model. This makes use possible even in environments with elevated compliance requirements.

How is this different from Zapier, Make or n8n?

Zapier, Make and n8n run flows you define in advance: trigger, steps, result. wals.pro AI is for questions and changes that come up in the conversation, such as an analysis or a single correction; every change goes through preview and approval. The two complement each other: what always runs the same way belongs in a workflow, one-off cases and analyses in the chat.

MCP oder Browser-Agent: Was passt für weclapp?

Ein Browser-Agent bedient die Oberfläche von weclapp wie ein Mensch, mit der Sitzung der angemeldeten Person. Über MCP ruft die KI festgelegte Werkzeuge auf, die über die weclapp-API arbeiten; bei wals.pro AI läuft dabei jede Änderung über Vorschau, Freigabe und Protokoll, und die Daten werden vor dem Schreiben geprüft. Für Änderungen mit Schreibzugriff, die im Gespräch entstehen, ist der festgelegte Weg leichter zu kontrollieren. Ein Browser-Agent kann dort helfen, wo es keine API gibt.

How do I give feedback or request a new feature?

In two ways, and both are explicitly welcome. Either directly in the AI chat: tell your assistant what is missing or not working, and it shows you the report as a preview and only sends it to us after your approval. Or by email to support@wals.pro, ideally with the chat history and your specific request. Handling and implementation cost you nothing, and every request is reviewed and answered. Which requests we actually implement, we decide based on the impact for all customers. What doesn't come, we'll tell you just the same.

We support weclapp customers ourselves as a partner. Can we offer wals.pro AI?

Yes. There is a partner track for weclapp partners, IT service providers and system integrators. The best way is to book a call directly →