040-22 86 65 28 frage@knowhere.to

Case Study  VELUX

Enhancing the customer journey while converting 38% of users

How VELUX Group is going against the stream with a chatbot focused on marketing and sales to support users in the information seeking rather than the customer service phase.

Alexander Dick, Digital Business Development Manager

+ 38%

converting users to a qualified leads

72 %

of user request are covered by the chatbot

50 %

degree of automation

Challenge: How to reduce complexity?

Our main challenge is trying to increase relevance for the individual user and looking to improve and optimise the customer journey, limiting frictions and making it easier to buy and install the right roof window.

When a customer wants to buy a window, many things matter: Roof pitch, roof material, purpose of window (light, view), room type… – adding into choices of number of windows, window sizes, best model, combinations, electric/manual, Iot, type of blinds, other accessories.

This makes it almost impossible for an end user to make a decision without professional assistance. So we need dealers and installers to understand how to advise end users for buying the optimal products.

In general, our main frictions in the customer journey are:

  • Complex product program
  • Complex processes
  • Complex price perception

This information is based on an extensive frictions analysis conducted with customers and potential customers in 9 markets.

Our research shows that proper support and guidance reduce the user’s perception of complexity, offers a better buying experience and increase conversion.

And this is where we saw the potential of putting a chatbot into play – in order to help reduce some of the frictions that are part of causing the (experienced) complexity.

Solution: Digital assistant with an atypical focus

After a comprehensive market-analysis during which we screened many chatbot providers, we decided for knowhere with their Full-Service AI Chatbot Solution as the perfect fit for our needs.

What we decided to do is to go against the stream and launch a chatbot with an atypical focus on marketing- and sales to support users in the information seeking rather than the customer service phase.

We divided the goals of the chatbot into a strategic and an operational focus.

Strategic focus:

  • Information seeking phase
  • Guides & assists users
  • Reduces frictions
  • Reduces complexity

Operational focus:

  • Acts as playmaker / navigator
  • Increases conversion
  • 24/7 availability
  • Human handover

Based on natural language processing the chatbots understand the intent behind each user question. Here are some examples for the intents of the VELUX chatbot:

  • Quote
  • Inspiration
  • Pricing
  • Accessories
  • Purchase
  • Delivery
  • Window Type
  • Repair
  • Replacement
  • Projects
  • Installer


Before the chatbot implementation, our expectations were difficult to manage due to the unpredictability of what users might ask. We hoped that 1/20 conversations would lead to a conversion point.

What we then saw after going live was that even though the bot is still new and still learning 42% of the users were converting as they were very engaged and motivated to use the chatbot to communicate with us. Conversion points are:

  1. Marketing qualified leads e.g. referrals to Velux shops, referrals to buy online from dealer​, referrals to map searches (dealer + installer), newsletter sign ups​, phone calls, download content​​, roof Window Price Calculator.
  2. Sales qualified leads​ e.g. contact Pro contacts (get a quote)​, phone call (let’s call you)​, emails (send us more info)​, live chat request​, book showroom​.

Directly from the start 72% of the user requests were inside of the chatbot scope. It’s early days and we are still learning what the users are asking for and what they expect to get back from the chatbot. This is part of improving the customer satisfaction.

What we then saw is that the AI chatbot is automating more and more user requests over time through two things:


  1. By asking intelligent feedback questions the chatbot can learn from each user dialogue and improve its understanding of each user intent.
  2. The AI systems proposes new intents based on questions that were not understood in the first place, which then are implemented by knowhere so the chatbot knows more and more “topics”.

See the Chatbot in action


VELUX Group is a multinational company present in more than 40 markets with a very wide and complex product catalogue. Velux has different product program variations, different local organisational setups, different market sizes – and huge market differences i.e. in the form of architecture, amount of daylight, heat, insects etc – as well as customer buying power and different degrees of DIY and route to purchase.


Building Materials, Home Decor


B2C Chatbot




Digital assistent


+ 38%

converting users to qualified leads

72 %

of user request are covered by the chatbot


degree of automation

Keys to Success

  1. Clear strategic and operational focus from the beginning for a well designed Conversation Design.
  2. Broad variety of intents and answers to attract many users to the chatbot and make it more likely to convert them to marketing or even sales qualified leads.

Start with a free consulting call:

You can expect to learn:

  • how to automate your customer communications in different messaging channel
  • how to qualify and route customer request to the right place with artificial intelligence
  • how to lower you response time to customer requests to under one minute