Customers call and write about stock, their prices and order status, and sales reps answer the same questions dozens of times a day. In the PrestaShop B2B shop, by email or in Teams, the assistant answers within seconds from SAP Business One or Business Central data, sees only that customer's contract, and hands new orders and discount requests to the sales rep.
Built for ISO 27001NIS2 DORA
A wholesale sales rep's day is made of short questions: "is it in stock", "what is my price for this item", "when does the order ship", "send me an invoice copy". Each answer means opening SAP Business One or Microsoft Dynamics 365 Business Central, finding the customer, checking the contract, looking up the shipment status in the DPD or DHL system and writing an email. While the rep is busy, the customer waits or calls again, and over lunch or after hours requests simply pile up. At the same time the risk of quoting a wrong price is real, because different customer groups have different contract terms.
The assistant works in your B2B web shop on PrestaShop or Magento, by email or in a Teams channel. When it receives a question, it first identifies who is asking, from the web shop login or a verified email address, and connects to SAP Business One or Business Central data on behalf of that customer only. A Claude model via API understands the question, turns it into a data query and answers specifically: stock level, the contract price from the customer record, order status with the number, a shipment tracking link from DPD or DHL, or an invoice PDF. It technically cannot see other customers' prices or stock reservations, because access is restricted by the integration layer, not by a model instruction. If a customer wants to place an order, asks for a discount or raises a complaint, the assistant hands the conversation to the sales rep with a summary. We connect to SAP Business One through its integration interface or report export, and to Business Central through an API, file export or email flow, depending on the version. The model runs on the Anthropic API or Amazon Bedrock in an EU region, conversations are recorded in a log, and the data is not used to train the model. Integration is especially simple when your ERP already runs on managed ERP hosting.
Four steps from input to a record in your system, with human approval wherever there is risk.
The customer writes in free text: "do you have 40 units of code X and what is my price". The assistant identifies the customer from their PrestaShop B2B shop login or verified email address and understands which products, orders or invoices the question is about.
The integration layer gives the model only that customer's record in SAP Business One or Business Central: contract prices, stock, orders and invoices. From these the model builds a clear answer with figures and dates and, when needed, adds an invoice PDF from the system and a shipment tracking link from DPD or DHL.
Informational questions are answered immediately. A new order, a discount request, a complaint or a question that is not in the ERP data is handed to the sales rep by email or in Teams with a conversation summary, and the customer is told who will get in touch and when.
Every conversation and every query to SAP Business One or Business Central is written to a log. Every month we review what the assistant did not answer or answered inaccurately and extend the instructions, and when contracts or the product structure change, the integration.
A chat window in your PrestaShop or Magento B2B shop, an email mailbox and a Teams channel: the customer chooses where it is most convenient to ask.
Access to SAP Business One or Business Central data is restricted per customer at the integration level, through an API or report export, with a separate technical user and a permissions log.
A list of situations the assistant does not answer but hands to a person, and a summary template so the sales rep immediately sees the context and the customer record.
A monthly overview of what customers ask most often, where the assistant did not answer, shipment statuses from DPD or DHL in the answers, and our support with a response from 30 minutes.
Claude assistant via API + SAP Business One or Business Central data through an integration interface, an API or a report export, depending on the version, and a chat window in the PrestaShop B2B shop. We choose the platform by the task and by where your data may be: read more about this service and its limits on the service page.
No. The assistant receives data through an integration layer that returns only the logged-in customer's contract, stock and documents from SAP Business One or Business Central. The model simply does not have other customers' data, so it cannot show it, either by accident or when asked.
Yes. We connect to SAP Business One through its interface or a report export, and to Business Central through an API, file export or email flow, depending on the version. In a PrestaShop or Magento B2B shop the assistant is embedded as a chat window and uses the customer's login to establish identity. We choose the integration method after a check.
Access to SAP Business One or Business Central through an API or reports, a list of customers and their contacts to establish identity, and a few dozen typical questions with the correct answers. The first working channel is usually launched within 2–4 weeks.
A new price list or quote for every customer group means a day of work in Excel and Word.
In wholesale, orders arrive as emails, Excel sheets, customer PDFs and EDI messages from retail chains, and the sales rep retypes them into SAP…
Every supplier sends a catalogue in its own format: different columns, units, language and codes.
At month end accounting matches statements to invoices by hand and writes reminders to debtors one by one.
This service is delivered by specialists with valid international certifications. Their expertise is confirmed by independent organisations, not by us.
AI systems are deployed with protection against data leakage and model abuse.
The AI security programme and its controls are managed following the ISACA methodology.
AI risks are identified and managed like any other IT risk.
We reply within 1 business day
