How we helped Frydenbø Bil improve the user experience and increase the efficiency of the customer dialogue

Screenshot 2025-01-31 at 09.02.37

Many customers found it difficult to find the right car, and the existing website required a lot of navigation and filtering to find relevant options

With a large number of customers and complex customer journeys across different car brands and importers, there was a desire to improve the digital user experience and take more control of the customer journeys. Together with us at Markedspartner, they have developed an innovative AI agent. It's called FrydenBot and helps customers navigate complex needs in the car buying process.

About Frydenbø Bil

Frydenbø Bil was established in 1994 and is one of the leading car dealers in Norway. They represent the car brands Volvo, Polestar, Audi, Škoda, Zeekr and Volkswagen, and offer a wide range of services from sales of new and used cars to service and workshop services.

Read more

The collaboration with Markedspartner

Through close collaboration and innovative use of generative AI, Frydenbø Bil and Markedspartner have created a groundbreaking solution that improves the car buying experience for customers. This project shows how technology can be combined with real needs to create value for both the business and customers.

"When we started working with Markedspartner, we quickly realized that we had found a good partner. We have worked very closely together right from the start. They bring a lot of expertise and we complement each other well. We also find them very innovative when it comes to finding and creating new solutions in smart ways."

- Heidrun Reigstad Tveit, CMO/Head of Marketing, Frydenbø Bil

What were the challenges?

When Frydenbø Bil was creating a new website, they identified several challenges in their customer journeys:

  • Many customers found it difficult to find the right car, especially used cars, due to complex needs such as number of seats, fuel type or budget. In addition, there are so many car brands and models to choose from.
  • Both Finn.no and existing websites required a lot of navigation and filtering to find relevant used car options.
  • Customers are often reluctant to ask sellers "stupid" questions.
  • A desire for a solution that could personalize the customer journey and make it more efficient.

So what did we do?

We proposed to use generative AI to create an interactive AI agent that could, among other things, provide customers with tailored car suggestions based on their unique needs. The project included:

Image_AI_Frydenbot_website_hthnmx4o01

Development of an MVP (Minimum Viable Product)

A chatbot using GPT-4 to analyze the customer's needs and match them with available vehicles from Frydenbø's warehouse.
Implementing a simple interface where customers could type in questions like: "I need a car with three Isofix seats and my budget is under 500.000 NOK."

 

 

Used cars

Data integration:

Use of Frydenbø's product data to ensure relevant and correct suggestions. Continuous improvement of the datasets based on user interactions.
Feedback

Testing and internal feedback

A hidden launch for employees and selected users to gather feedback. Iterative development to optimize the user experience and solve technical challenges.

What were the results?

Frydenbot has already shown promising results:

  • Engagement: 75% of users who open the chatbot send a message. On average, they send five messages, indicating an active dialog.
  • Conversion: 20% of users click through to specific car suggestions.
  • Better user experience: Customers quickly get relevant options without having to navigate through complex filters.

"With FrydenBot, we've made car shopping easier and more intuitive for our customers - they get recommendations in seconds without having to navigate through complex filters and websites. Working with Markedspartner has given us new perspectives and valuable expertise, where together we have developed a solution that actually solves our customers' everyday challenges. And this is just the beginning - we see great opportunities to further develop FrydenBot into a full-fledged customer service agent that can handle everything from car selection and questions about car ownership, to form submissions."

- Heidrun Reigstad Tveit, CMO/Head of Marketing, Frydenbø Bil

Lessons learned

  1. Real needs drive success: By focusing on a concrete problem, such as finding the right car for specific needs, the AI solution became relevant and valuable.
  2. Iterative approach: Small steps with continuous testing and user feedback have ensured that the product meets expectations.
  3. Ease of use: Simple improvements to the user interface, such as mobile optimization and sample questions, have been key to increasing engagement.
  4. Dataset quality is crucial: To ensure accurate answers, the data must be correct and up-to-date.
  5. Continuous updates: When ChatGPT rolls out updates, it gives us improvements Frydenbø can roll out too. such as better load times and voice control.

The road ahead

Frydenbø Bil and Markedspartner plan to further develop the purchase helper:

  • Expand the functionality to include more product categories and services so that FrydenBot becomes a full-fledged customer service agent.
  • Explore how AI can improve other parts of the customer journey, such as customer service and upselling.
  • That FrydenBot can perform actions such as submitting forms and booking test drives.
  • Increase the number of users of FrydenBot to reduce email and phone inquiries.

"We hope this inspires others to explore generative AI as part of their customer journeys. It's about testing, learning and constantly improving. This is just the beginning"

- Ken Leren, CTO at Markedspartner