Michael Maoz

And it also violated privacy and a whole bunch of rules that would get you shut down in Europe for GDPR or in California or any place where you really have to know the provenance of the data.

And it also violated privacy and a whole bunch of rules that would get you shut down in Europe for GDPR or in California or any place where you really have to know the provenance of the data.

And it also violated privacy and a whole bunch of rules that would get you shut down in Europe for GDPR or in California or any place where you really have to know the provenance of the data.

Yeah. And you're asking a question that could be phrased a little bit differently. And how do I do something right away with the data I have so I don't have to do that big, just like people spent years building a big snowflake repository of information or Databricks or whatever. And it's like, well, we'll get around to fixing our customer process when we get that project done.

Yeah. And you're asking a question that could be phrased a little bit differently. And how do I do something right away with the data I have so I don't have to do that big, just like people spent years building a big snowflake repository of information or Databricks or whatever. And it's like, well, we'll get around to fixing our customer process when we get that project done.

Yeah. And you're asking a question that could be phrased a little bit differently. And how do I do something right away with the data I have so I don't have to do that big, just like people spent years building a big snowflake repository of information or Databricks or whatever. And it's like, well, we'll get around to fixing our customer process when we get that project done.

We don't want you to do that because the tools are there now with things like Agent Force and other tools, but... to do something now. So the first thing to say is, why don't we take information that you can trust? You probably have a knowledge base around simple things. So we have a client in the UK in the health care industry. And they said, we really want to jumpstart our generative AI program.

We don't want you to do that because the tools are there now with things like Agent Force and other tools, but... to do something now. So the first thing to say is, why don't we take information that you can trust? You probably have a knowledge base around simple things. So we have a client in the UK in the health care industry. And they said, we really want to jumpstart our generative AI program.

We don't want you to do that because the tools are there now with things like Agent Force and other tools, but... to do something now. So the first thing to say is, why don't we take information that you can trust? You probably have a knowledge base around simple things. So we have a client in the UK in the health care industry. And they said, we really want to jumpstart our generative AI program.

Let's just take all the emails that we have answered over the last year and analyze them, ask us some very basic things. It turns out there are only three or four things that people ask repeatedly on this site. So I said, that's a great thing. We know this data is clean. We know that the search doesn't go outside of this canonical database that we created or knowledge base we created.

Let's just take all the emails that we have answered over the last year and analyze them, ask us some very basic things. It turns out there are only three or four things that people ask repeatedly on this site. So I said, that's a great thing. We know this data is clean. We know that the search doesn't go outside of this canonical database that we created or knowledge base we created.

Let's just take all the emails that we have answered over the last year and analyze them, ask us some very basic things. It turns out there are only three or four things that people ask repeatedly on this site. So I said, that's a great thing. We know this data is clean. We know that the search doesn't go outside of this canonical database that we created or knowledge base we created.

So let's point AgentForce at that. And what happened immediately out of the box was that 80% of the emails disappeared. Now, not 100%, because if agent force detected that it wasn't really sure that the answer it was giving was accurate enough to send directly to the customer, they put it to the agent. But the 80% went away, the 20% stayed. And of course it's getting better.

So let's point AgentForce at that. And what happened immediately out of the box was that 80% of the emails disappeared. Now, not 100%, because if agent force detected that it wasn't really sure that the answer it was giving was accurate enough to send directly to the customer, they put it to the agent. But the 80% went away, the 20% stayed. And of course it's getting better.

So let's point AgentForce at that. And what happened immediately out of the box was that 80% of the emails disappeared. Now, not 100%, because if agent force detected that it wasn't really sure that the answer it was giving was accurate enough to send directly to the customer, they put it to the agent. But the 80% went away, the 20% stayed. And of course it's getting better.

It's an iterative process. And now they're, you know, they're, they're improving, but having that human in the loop so that the technology can detect, I'm not really that confident with this. I'm going to send it back. And we'll talk later about our Atlas reasoning engine, which is also doing an amazing thing on helping automate that process as well. But in parallel, the,

It's an iterative process. And now they're, you know, they're, they're improving, but having that human in the loop so that the technology can detect, I'm not really that confident with this. I'm going to send it back. And we'll talk later about our Atlas reasoning engine, which is also doing an amazing thing on helping automate that process as well. But in parallel, the,

It's an iterative process. And now they're, you know, they're, they're improving, but having that human in the loop so that the technology can detect, I'm not really that confident with this. I'm going to send it back. And we'll talk later about our Atlas reasoning engine, which is also doing an amazing thing on helping automate that process as well. But in parallel, the,

Then you can go a little further and say, we have six projects going, eight projects going, but every one of them points back to different places in different systems. It's in Jira. The data's in Confluence. The data's in SharePoint. The data's in a Salesforce CRM. Some of it's in SAP or wherever it might be.

Then you can go a little further and say, we have six projects going, eight projects going, but every one of them points back to different places in different systems. It's in Jira. The data's in Confluence. The data's in SharePoint. The data's in a Salesforce CRM. Some of it's in SAP or wherever it might be.