Salesforce shows how AgentForce 2.0 handles complex financial questions by reasoning across multiple data sources and citing its work. A bank advisor can now ask about portfolio impacts from interest rate changes, not just retrieve basic balances.
So this is a customer record for Emma Reed. Emma is a Cumulus bank wealth Management customer, and this screen helps Cumulus Bank's financial advisors work with Emma to accomplish her investment goals. It also gives them access to AgentForce, where they can ask a quick question like what is the current balance of Emma's portfolio? And because AgentForce is connected to this data, they'll get a quick, helpful answer to this basic question. But not all questions are basic. Imagine asking, how will falling interest rates impact Emma's portfolio? Now this is a complex question, and what you'll see is AgentForce generates an answer that isn't wrong, but it's not as complete as it could be. Because although this answer considered Cumulus Bank's knowledge article on this topic, it did not take the extra time to consider the funds that are in Emma's portfolio or Cumulus Bank's guidance on how to identify which of those funds are sensitive to interest rate changes or how those interest rates might impact EMA's portfolio. The point here is that Agent Force needs to recognize that it should take some extra time with this question to do deep retrieval and reasoning. And that is what Agent Force 2.0 makes possible. So let's switch over to our agentforce 2.0.org and try this question again by asking how will falling interest rates impact Emma's portfolio? And now AgentForce is thinking more deeply about the question it switched from basic retrieval to advanced retrieval, which means this answer is going to take a bit longer to generate. But with Agent Force 2.0's new streaming capabilities, we're able to watch this answer be generated in real time. And it is a much deeper answer, rich with citations, that provide visibility into the reasoning used to generate this answer. Which is why Cumulus Bank's financial Advisors know that AgentForce considered all these underlying data sources to produce this deeply researched answer. And it's an answer they can put to work by requesting a specialist be invited to discuss these considerations. An Agent Force will use all this analysis to find the perfect specialist with the right skills in the right location and coordinate scheduling. Expert answers require expert understanding, and that understanding can't be found in one source. It requires deep retrieval and reasoning across many data sources. And that is what AgentForce 2.0 makes possible. Thanks for watching.
