Venture Clienting with a Leading Swiss Home Solutions Provider 

Focus Area 2: AI-Powered Service Desk Automation 

What is the company looking for? 

WhatsApp has become a critical communication channel for the companies over 30 stores, enabling fast and direct interaction with professional customers — primarily plumbers. 

However, the current process — using shared phones or WhatsApp Web — is manual, time-consuming, and inefficient, with typical response times around half an hour per inquiry, and only a small fraction of requests resolved within a few minutes. This inefficiency translates into a significant annual workload, tying up considerable staff resources and impacting both productivity and customer satisfaction. 

The company is therefore looking for an assistant solution that enhances human-to-human WhatsApp communication. The goal: to enable staff to handle inquiries faster and more reliably by providing live context analysis, product suggestions, and seamless integration — all while preserving the personal, professional interaction customers expect. 

Core Question & Criteria 

How can we develop a scalable and user-friendly assistant solution that supports staff in handling WhatsApp customer inquiries faster and more efficiently while preserving human-to-human interaction? 

Planned Solution Development Process 

The development of the WhatsApp Support Assistant is envisioned in three phases: 

• Phase 1 (current scope): 
• The assistant reads live WhatsApp chats. 
• Searches for requested articles in the background. 
• Suggests the correct products to staff via an intuitive UI. 
• Enables staff to place the product in the customer’s cart or create an SAP order directly. 
• Phase 2 (planned for 2026): 
• The system is further enhanced and fully integrated with Salesforce’s WhatsApp functionality. 
• Allows for richer CRM context and seamless data flow between WhatsApp and Salesforce. 
• Phase 3 (future vision): 
• The assistant autonomously responds to inquiries when it is ≥95% confident in its answer. 
• Escalates chats to human staff only when confidence falls below this threshold. 

The solution (Phase 1) should thus be able to: 

• Read and extract content and context from live WhatsApp chats. 
• Analyze chat intent and identify requested products or information. 
• Search and match relevant product data in real time from internal systems (e.g., SAP, PIM). 
• IMPORTANT: Understand and interpret images submitted via WhatsApp, such as customer photos of devices or spare parts — even when these are incomplete, low-quality, or taken in real-world settings. This functionality is critical for enabling accurate product matching and recommendations. 
• Suggest accurate, context-aware replies and product recommendations to staff. 
• Enable staff to place products in the customer’s cart or create SAP orders directly. 
• Integrate seamlessly into existing workflows and systems, supporting SAP, PIM, Salesforce, and Snowflake where relevant. 
• Maintain human-to-human communication, acting only as a background assistant. 
• Be scalable and adaptable to future phases, including Salesforce integration and autonomous response capabilities. 

In Scope: 

• Software solution, no hardware development. 
• Prefered is a ready-to-use (Plug & Play) solution, but adaptable or customizable existing solutions are also in scope. 
• Leverage existing data sources; no extensive new data collection required. 
• Focus strictly on assisting staff with WhatsApp-based customer communication. 
• Compliance with Swiss data privacy and security standards (e.g., revDSG). 

Out of scope: 

• Fully autonomous chatbot for direct customer interaction (Phase 3, future). 
• Broader CRM, logistics, or manufacturing process control outside WhatsApp support.