ifm: From Product Search to Digital Product Advice
With a portfolio of more than 14,000 products, ifm offers customers a wide range of solutions. Finding the right product through traditional keyword search can be challenging, particularly when customers know what they want to achieve but do not know the relevant product name or technical term.
The Solution: The diva-e Conclusion Product Advisor provides a new way to approach product search. Instead of starting with a product name, customers can describe their application and what they want to achieve. The AI Product Advisor uses this context to help identify suitable products.
Use Case
From Article Numbers to Application-Based Search. Customers searching for article numbers such as IA0032, IFT200 or Pi2795 already know exactly what they need.
The more challenging scenario is when a customer knows the application but not the right product or technical term. The AI Product Advisor provides an additional entry point in these situations.
Customers can continue to search by article number as before. Alternatively, they can describe their application or problem in natural language.
A typical customer query:
“How can I reliably detect and count shipping boxes on a roller conveyor in the dusty environment of a logistics center?”
There is no product name or specific technical specification in this query. Instead, it describes the environment, the application and the desired outcome.
The AI Product Advisor understands this context and recommends suitable products – in this case, 3D sensors from the O3D series. It compares different variants and explains why they are suitable for the application.
In this way, product search becomes a form of technical guidance.
From AI Demo to a Reliable Digital Service
Building for Production, not Just for the demo: Turning an AI prototype into a production-ready application requires more than connecting a language model to a chat interface. For the Product Advisor, stable and reliable operation was therefore a key consideration from the beginning.
Quality Assurance
Responses need to be based on approved sources and remain technically traceable.
Clear Boundaries
The Product Advisor needs to recognize when it can provide useful guidance and when technical specifications or personal advice are required.
Monitoring and Automated Testing
Both the underlying models and the product data can change. Response quality therefore needs to be monitored continuously.
User Feedback
Customers can rate responses and provide feedback through a free-text field. This feedback can be used to further improve the service.Clear Responsibilities
Maintaining data, evaluating technical quality and deciding on changes require clearly defined processes and responsibilities.
This approach turned the initial pilot into a digital service that can be operated reliably in production.
The Result
More than a Chatbot: The AI Product Advisor is a new layer of advice within the customer journey – and another step towards a more intelligent, customer-focused digital sales experience.
Since June 2026, the AI Product Advisor has been live in the first three countries and covers a portfolio of more than 14,000 products. Further use cases and international rollouts are planned.
First Success Metrics
Initial evaluations show that customers are actively using this new way to search for and learn about products. The assistant is also being used internally by teams in Service, Product Management and Sales. Other countries have also expressed interest in offering the service.