This year’s Dreamforce conference in San Francisco is all about the new Agentforce platform. In his keynote speech on Tuesday, Salesforce founder Marc Benioff presented the technology behind it and concrete use cases that are intended to usher in the “third wave of AI”. Agentforce enables companies to automate customer interaction through autonomous digital agents. They are intended to enable almost human interaction. (brutkasten reported). But what specific advantages do the digital agents offer, and how will they change customer interaction and the world of work? We spoke to David Hable about this at Dreamforce. He is Co-Country Leader of Salesforce in Austria.
brutkasten: Agentforce was introduced at the Dreamforce conference. The platform is intended to usher in a new era in customer interaction. What advantages do agents offer?
David Hable: We are definitely entering a new era. Agents will be able to do a lot because they have direct access to the information they need in the background via our data cloud and the platform. They don’t have to read this data first or search through it like a human does; they have it available immediately. In many cases, agents can even provide information faster and more precisely than a human could. And now they do it in a way that seems almost human-like.
This development combines the best of both worlds. Chatbots have always been good at asking for digital information, but the interaction often seemed a bit clumsy. They were ready-made answers that never really felt natural. Now we are able to combine the best of both worlds – the efficiency and speed of digital systems with natural, human interaction.
What use cases for agents will there be in the future?
We presented various areas at Dreamforce: sales, service, commerce and marketing. Our platform also makes it very easy to build new agents. In sales, for example, it’s about training sales people. Instead of putting new sales employees in contact with customers straight away, they can first interact with our digital agents a few times. Another example in the sales environment is cold calling. Here, too, we presented solutions where agents can take on this task in the future. There are various use cases in customer contact.
Think of it this way: If you have information about customers in your systems, we can automate interactions with agents. And it doesn’t just have to be Salesforce systems. With Datacloud, we can connect external data into our system. When a customer asks for information such as invoices, delivery times, prices or other data-based requests, we can automate these interactions and efficiently deliver the requested information to the customer via Agentsforce.
With 45,000 participants, Dreamforce is one of the largest events in the tech world | (c) martin pacher / brutkasten
What impact will this have on the world of work? Will call center employees become redundant in the future?
I don’t think they will disappear completely. There will always be use cases where human interaction is essential. Our focus is on providing the best possible experience for the customer when AI and humans interact. The technology enables employees to focus on the really important tasks instead of having to deal with administrative tasks.
There was a lot of talk about Trustworthy AI at the Dreamforce conference. How can we ensure that data is handled securely?
We started in 1999 and were the first true public cloud provider. Salesforce basically invented SaaS solutions. It hadn’t existed before us. But even then, and even five to ten years ago, there was often a certain skepticism about cloud applications. There was a fear of handing over control of your own data to other providers. From the beginning, we met this challenge by establishing trust as a core value in Salesforce and in the interaction with our customers.
The same principle applies today in the field of AI. We have been working with AI for over ten years and all our developments are based on this principle of trust. In AI, we speak of the so-called trust layer. This ensures that the data processed in the AI is not used without authorization. Personal data is encrypted before it even reaches the Large Language Model (LLM).
David Hable at Dreamforce | (c) martin pacehr | brutkasten
These security measures mean that our customers can be confident that they can work with our AI and still use their company-specific data. Because without this data, it won’t work. An LLM alone that only contains general information from sources like Wikipedia is not enough to enable valuable customer interactions. The AI needs to know the relevant customer data in order to be able to have meaningful and valuable interactions with customers.
Before Dreamforce, Marc Benioff spoke about a “hard pivot” that Salesforce is making with the move to agents. How will this change of course affect Salesforce?
I see this as a clear development, not necessarily a change of course. If you look at Marc’s keynotes a year ago, this path was already apparent – it was already on the slides. First we started with predictive AI, then we moved on to generative AI, and the next step was autonomous AI. Now we are at exactly that point.
How does Salesforce differ from competitors like Microsoft?
If you want to work effectively with AI, you need continuous company information and processes. That’s exactly what we can ensure with our platform. Our platform provides access to information from areas such as sales, service, marketing, trade and many others. This enables us to manage these processes efficiently and balance them optimally. That’s what sets us apart from the competition – we are the only ones who can offer this comprehensive integration of information and processes.
What is the pricing of the new Agentforce model?
Pricing is basically based on usage, i.e. how much you consume. A conversation with an agent will cost at least two US dollars.
What is the availability of Agentforce in Austria?
Agentforce will be rolled out globally from October – including in Austria. We will also be presenting the technology at our Innovation Day on November 12th in Vienna. There you can try out for yourself how easy it is to build an agent. You don’t have to be a programmer or technician to do this.
Link tip to the Salesforce page
*Disclaimer: Travel expenses were covered by Salesforce.
