From complaint to solution in minutes with AI, Mendix and Lobster Data
Leveraging AI within your organization has so much potential, but where to start? Let’s talk about a practical use case that we’ve seen at one of our customers and have built a demo for.
Every company knows the scenario. A customer sends an angry email: a product arrived damaged, an order is late or something doesn't work. And then the clock starts ticking. Not just on the customer's patience, but on your service team. What is the issue? Which order was it? Is it still under warranty? What does the ERP system say? Before anyone has even thought about an answer, an employee has opened four systems and lost twenty minutes.
But what if the answer was already waiting for them?
That's exactly what we built in a recent demo project: an AI-supported complaint handling process where artificial intelligence, the Lobster data platform and a Mendix complaint handling application work seamlessly together. What we automated is a complaint that arrives by email and is automatically read, understood, enriched with order data from the core system and presented to an employee. Complete with a suggested resolution and a draft reply. All the employee has to do is what humans do best: judge and decide.
Let’s dive into this
The setup is surprisingly simple, and that's the point.
A customer emails a complaint about a product. Lobster, the central data platform, monitors the mailbox and picks up the message. An AI model, initiated by Lobster, reads the email and determines what it's actually about: what kind of complaint is it, how urgent is it and which order does it concern?
With that order number in hand, Lobster retrieves the relevant data from the core ERP system: the customer, the delivery date, the warranty status, the order value. AI combines the complaint with this context and formulates a concrete proposal, for example: "Within warranty, urgent situation at a key customer: send a replacement immediately and offer a return label." It even drafts the reply email.
All of this lands in a Mendix application built specifically for the service team. The employee opens one overview and sees everything: the original email, the complete order data, AI's assessment and a ready-to-send response. The employee can analyze and approve the AI suggested solution or adjust and take other appropriate actions.
Een klant stuurt een e-mail omdat een product beschadigd is aangekomen, een levering vertraagd is of een product niet goed werkt. Vanaf dat moment begint de zoektocht.
Om welke bestelling gaat het? Valt het product nog binnen de garantie? Wat staat er in het ERP systeem? Heeft de klant eerder contact gehad?
Voordat een servicemedewerker überhaupt kan reageren, zijn vaak meerdere systemen geopend en zijn er al snel 20 minuten verstreken.
The benefits
1. Reaction time to customers. What used to take twenty minutes of searching, switching and typing now takes less than one minute. The complaint is pre-processed before an employee even looks at it. For the customer, it means a faster reply; for the organization, more complaints handled with the same team.
2. No more system-hopping. The employee no longer needs access to the mailbox, the ERP system and a CRM. Everything relevant is brought together in one overview. Besides saving time, it prevents errors and lowers the threshold for new employees to get productive.
3. Better tracking and control. Every complaint becomes a structured case with a status, a history and a clear outcome. Management gets insight: how many complaints come in, about which products, how fast are they resolved? What used to live scattered across inboxes is now measurable.
4. The human stays in charge. The whole setup is AI supported, not AI decided. Every suggestion is reviewed by an employee before any action is taken. This keeps the process trustworthy and compliant, while still capturing nearly all the efficiency gains.
Can we help you?
Contact Martijn!
The power of Mendix: a business application in days, not months
Data and AI suggestions are worthless if employees can't work with them comfortably. With Mendix, we built the complaint handling application: a clear dashboard, a case screen with everything in one view and buttons that match exactly how the service team wants to work.
Lobster and AI are integrated in the app.
That's the strength of Mendix: the application follows the business process, not the other way around. Does the team want an extra status, a different prioritization, a link to yet another system? It's adjusted in hours. And because Mendix integrates effortlessly with the surrounding landscape, in this case Lobster, the application is never an island, but a full participant in the chain.
Klanten krijgen sneller antwoord
Dezelfde service afdeling kan aanzienlijk meer klachten verwerken.
The power of Lobster: one platform for all integrations
The power of Lobster: one platform for all integrations
A scenario like this stands or falls with integration. Email, AI models, ERP systems, business applications: they all speak different languages. This is where Lobster shows its strength: one platform through which all data flows run.
The mailbox connection, the calls to the AI model, the retrieval of order data, the handover to the Mendix application: every integration is configured in Lobster, without custom code. With support for modern standards like MCP, even AI agents connect to the existing landscape in a controlled, manageable way. No tangle of point-to-point connections, but one central place where you see, manage and monitor all data traffic.
Small demo, big potential
Our setup was deliberately kept simple: a demo to show what's possible. But the pattern, AI that reads and understands, Lobster that connects everything, Mendix that empowers the custom process, applies to far more processes than complaints alone. Think of order intake, invoice processing, supplier questions or internal service requests.
The technology is ready. The question is no longer whether AI can strengthen your processes, but where you start.
Curious what this could look like in your organization? We'd be happy to show you the demo live.
