'Huizensite' Funda works for different types of clients: real estate agents, serious house hunters and consumers who are mainly interested in what is for sale - and what it looks like on the inside. Funda has a direct relationship with the real estate agents in particular. The housing site is a textbook example of 'starting small' with Agentforce.

 

Stephen Fisher (Commercial Operations And Applications Manager) is responsible for the Salesforce platform at Funda, among other things. He is responsible for ensuring that the commercial colleagues within Funda can work as well as possible with Salesforce, the platform that Funda has been working with for almost five years. Fischer was brought in specifically for the implementation at Funda; after that, the platform was implemented step by step. First Marketing Cloud, then Sales Cloud (including for advertising sales) and for the past two years also Service Cloud, including for the provision of services to real estate agents.

AI vision drawn up

Fischer's team drew up a vision for customer support prior to the AI ​​experimentation phase. "It sets out what Funda wants to do with GenAI over the next 24 months. That vision is a living document, it is still being worked on," says Fischer. "But it is clear that we want to say goodbye to repetitive tasks that do not add value. AI agents are very suitable for that. We are also looking for ways to support employees and increase customer satisfaction."

Small scale use cases

Fischer calls the first use case “very small-scale”, aimed at the automated summarization of calls at Funda's customer service department. Every week, this involves 700 to 800 emails from consumers and real estate agents, and around 80 calls every day. “Real estate agents mainly call, consumers mainly send emails. Automated summarization saves time and ensures more uniformity in the registration; so it is mainly an efficiency boost.”

In addition, Funda is preparing a customer-facing use case for the help center. “We see that consumers, as users of Funda, are looking for answers to their questions, but cannot always find them. The solution asks the user to describe the question in more detail. We hope to further improve customer satisfaction with this, because at 74% it is lower than with the brokers where the CSAT “more than 90%,” Fischer explains.

Real-time support

While consumers have relatively simple and few questions, brokers report very specific and often urgent issues. “For example, a photo that appears incorrectly on the screen or a listing that is not correct or does not appear on time. We support brokers in real time, and in the future this could be done via Agentforce – so when they are busy putting an object online, for example. We now mainly want to gain experience based on a limited scope and from there move on to the next use case. Trust must be created within Funda that the application does what is promised,” says Fischer.

More towards the future, you can think of AI agents that will support the brokers in strengthening their possibilities to get leads. Think of the relationship between uploaded image and the composed text: how can you describe an object as attractively as possible? Fischer also mentions an application such as making recommendations regarding additional services that fit well with a certain object that is placed on Funda.

During the Agentforce event, it was often repeated: without good data management, no (well) functioning AI agents. Fischer therefore says that Funda does not have a data problem. “Of course, there is sometimes a problem with data quality, but in general we are well organized: we have a good data warehouse, which is easily accessible.”

Encouraging safe use

Almost as important as the gradual adoption of GenAI and Agentforce is encouraging the use of AI by Funda employees. “We encourage our colleagues to get started with AI themselves and explore what is possible. We do not want to impose too many restrictions, because then you lose momentum. A survey among employees shows that 98% of employees use GenAI, privately and/or professionally. It helps that people are curious and have the freedom to experiment, of course with a few rules that ensure peace of mind for colleagues. For example, in consultation with legal and IT we have agreed which models you may and may not use. In fact, it is an application that you add to your IT landscape.”

Fischer also says that Funda organized an internal AI hackathon with all employees last year. “One of the AI ​​applications that came out of that is generating banners via AI for real estate agents. Not all real estate agents have everything in-house to develop banners themselves.”

Keeping up with developments

“Part of that journey of discovery is also that we try to come up with new ideas together with the teams and the brokers. For example, today (at Agentforce, ed.) we attended a workshop with a number of colleagues. Everyone knows that AI is a theme, but keeping up with all the developments is still quite difficult: it goes fast. We think it is important that everyone in the Funda team is well connected.”

 

Funda is the market leader in making (predominantly) owner-occupied homes visible in the Netherlands. The site attracts 4,6 million unique visitors per month. The website was launched in February 2001 on the initiative of the trade association Nederlandse Vereniging van Makelaars (NVM) and Wegener Media. Since September 2007, in addition to the listings of NVM brokers, the listings of other brokers have also been shown.

(Ziptone/Erik Bouwer)

Featured, Knowledge partners, Technology Salesforce