Developer Keynote Highlights
Words:
1,366
Duration:
00:07:28
Recorded On:
Uploaded On:
8 months ago
[Fri, 20 Sep 2024 00:59 UTC]
Transcribed On:
8 months ago
[Fri, 20 Sep 2024 01:00 UTC]
Last updated:
Not yet updated
Location:
Using:
Uploaded
[Developer Keynote Highlights.mp4]
Download
Language:
English (US)
Channels:
2
Sample Rate:
44100 Hz
Shareable Link:
Speakers:
Speaker A -
28.7%
Speaker B -
71.3%
| Duration | 00:07:28 |
|---|---|
| Word Count | 1,366 |
| Source | Uploaded [Developer Keynote Highlights.mp4] |
| Recorded | |
| Transcribed | Sep 20, 2024 at 01:00 AM UTC |
| Updated | |
| Location | |
| Language | English (US) |
| Sent to |
 Google Drive
 Add to Headliner
 Notion
|
| Speakers |
Speaker A -
28.7% |
|
Speaker B -
71.3% |
|
| Notes |
- The way we build applications is changing and the Salesforce platform provides developers with so much opportunity. In uh, meet the developers and the developer keynote, you'll get a view into the new paradigms for developers and how you can build the future.
|
| Speaker A | ||
| [00:00:00] | Exciting time to be a Salesforce developer. With huge advances in data, AI and agents, the way we build applications is changing and the Salesforce platform provides developers with so much opportunity. In uh, meet the developers and the developer keynote, you'll get a view into the new paradigms for developers and how you can build the future. Now I am joined right now by the amazing Mo mo, Mohit Srivasra. Hello, Mohit. | |
| Speaker B | ||
| [00:00:23] | Thank you, Jillian. Thank you for having me here. | |
| Speaker A | ||
| [00:00:25] | So you are our principal developer advocate at Salesforce. Yes, longtime friend, yes. Good to have you here. | |
| Speaker B | ||
| [00:00:31] | Thank you so much for having me here. | |
| Speaker A | ||
| [00:00:33] | Let's talk about what's going on at Dreamforce for developers. | |
| Speaker B | ||
| [00:00:36] | Yeah, you can see at uh, Dreamforce we have all of these tools in display here and I've seen developers actually learning them, going to this camp trailhead ground, trying all these things and building stuff. And that is amazing because it's not just about you're talking through all these things, you're actually able to get hands on and build things. | |
| Speaker A | ||
| [00:00:59] | I saw there was some stat like almost 2000 people have already built agents using agent forest so far. | |
| Speaker B | ||
| [00:01:05] | That's right. Because it's so simple, we make it so easy for people to get started. It's, yeah, you know, you can easily build and it's just enabling them to be more creative. Like you just have to understand your business problem and then uh, you know, use these tools that we have here at uh, Dreamforce. | |
| Speaker A | ||
| [00:01:26] | So let's talk a little bit about some of those tools. What exactly are some of these new capabilities and tools specifically for developers when it comes to agent force? | |
| Speaker B | ||
| [00:01:33] | Yeah, absolutely. Now before I get into these agents, let's talk about what are some of the ingredients for them. Ingredients for building these agents. It all starts with large language model. You need a large language model. I think at Salesforce we have done very good job by not only just building our own models, also being open, they can bring in all these models and build these applications. Um, but more than model, these agents are going to need knowledge, which is the data and that's what I'm going to talk about. One of the tools that we have for that is data cloud. We had a theater here, intro to data cloud and uh, every time it was packed. So you can see everybody wants to build that data foundation. And if you are a developer, that is where you need to invest your energy on. Understand what are the features of this data cloud, how you can connect Salesforce and other systems with it, how you can customize it, customize to your need so that's about data foundation right there. Uh, and then we have all the interesting stuff, which is how can you bring AI capabilities into your application? And that's where we have promptBuilder. Um, a lot of developers have experienced promptbuilder now because we launched at TDX, we are adding a lot of new capabilities. Uh, and with prompt builders, you can build these prompt, bring that generative AI capabilities that you want into your application using both the low code and the pro code tools that we have. Um, and here, creativity is only your, uh, limit with prompt builder, actually, because I've seen some interesting lightning components that's just going to make your, uh, users more productive. And then finally agent force, build all these agents. | |
| Speaker A | ||
| [00:03:17] | All right, so let's talk about agents. Uh, let's talk a little bit about how developers are going to really harness the power of agent force and start building. What do you envision? What are some of the things that they're going to start doing? | |
| Speaker B | ||
| [00:03:27] | Yes, so, uh, let's understand agent force. So agent force, or agents are next generation systems which can take, uh, basically do all these tasks autonomously, of course, with human intervention and human, uh, in the loop as well. Um, but if you look at agent force, we need to understand what it takes for developers to build. If you're doing your own diY, as Mark said in his keynote, it's a lot of work, a lot of work, and we make it really super easy for you with agent force. With agents, you're bridging that humans with software. So you as developers, you have already built apex flows. And again with promptbuilder, you can build these prompt templates too. So all you're doing is writing natural language instructions on top of it so you can invoke these right things. So you're already building that software stuff and all that is going to be reusable. But with large language models, we'll give you that capability. And now with Agentforce, you write instructions, um, just in natural language and make your thing happen. I mean, it's just the large language model doing the reasoning for you, but the actual software. Again, you as a developer are, ah, building your software system. So it's exciting. | |
| Speaker A | ||
| [00:04:48] | Yeah, I mean, it's like everyone has already been building the building blocks with which then you will assemble to create an agent. | |
| Speaker B | ||
| [00:04:55] | Yes. | |
| Speaker A | ||
| [00:04:56] | Right. I think that's a really helpful story. | |
| Speaker B | ||
| [00:04:58] | Yes. And previously we had to assemble it as a developer. Like I had to go as a developer and say, if this, then do this, then, and then add these branches and logics, right? Uh, now, large language model can do it. If we can describe it in large, you know, describe it in natural language. | |
| Speaker A | ||
| [00:05:15] | Yeah. You can type a sentence. | |
| Speaker B | ||
| [00:05:17] | Yes. | |
| Speaker A | ||
| [00:05:17] | You can build an agent. | |
| Speaker B | ||
| [00:05:18] | Yes. That's how easy it is. | |
| Speaker A | ||
| [00:05:21] | I love that. Okay, so let's talk a little bit about some use cases. Do you have specific use cases that you've seen people already think about when they're using agent force? | |
| Speaker B | ||
| [00:05:28] | Yeah, absolutely. So one of the use cases, very common. And we have this in our trailhead module as well as service agents. Uh, absolutely. Uh, and we showed this at keynote also, like, if you're reaching out to your customer support today, you don't want to be waiting. Uh, and even as a human, the service agent is actually looking at the knowledge base, looking at the company policies, and responding to you. And these agents are working in the same way. So that's a very good use case. Again, we give out of box template and you as a developer can understand your business use case, your processes that you need, use flows and apex to build them. And then just use our agent builder. So, you know, write natural language instructions so all, you know, and then expose it in various channels, like experience cloud site that we are showing in that trailhead module. But it could be, you know, in your Apple messaging, it could be any channel that you can imagine. Slack. Yes, I know you love slack, and. | |
| Speaker A | ||
| [00:06:29] | Yes, I sure do. Yeah. | |
| Speaker B | ||
| [00:06:30] | Yes, on slack too. Expose all these agents and deploy them. | |
| Speaker A | ||
| [00:06:34] | That's great. Okay, so we are, um, wrapping up here, but just one quick note. What is one thing that developers should really start thinking about if they want to skill up for building with agents? | |
| Speaker B | ||
| [00:06:43] | Yes. So only thing that they need to understand in my, uh, experience is learning, uh, how to write natural language instructions, because that's a new skill. Prompt engineering is what we call them. So understanding that and little bit of understanding your domain, like you're working for these customers, understand their business needs because that's what you'll be focusing on. | |
| Speaker A | ||
| [00:07:05] | I love that. All right. Learn how to write sentences instead of code. I love it. Thank you so much, mo, for joining us. Really appreciate you. There are so many resources for developers to get started. You can scan the QR code on your screen to learn more about how to build the future. And have some fun while you're doing it. Take the developer Dreamforce Quest. That link is Sforce Co. DF 24 DevQuest. Thank you so much for watching. |
| Speaker A | ||
| [00:00:00] | Exciting time to be a Salesforce developer. With huge advances in data, AI and agents, the way we build applications is changing and the Salesforce platform provides developers with so much opportunity. In uh, meet the developers and the developer keynote, you'll get a view into the new paradigms for developers and how you can build the future. Now I am joined right now by the amazing Mo mo, Mohit Srivasra. Hello, Mohit. | |
| Speaker B | ||
| [00:00:23] | Thank you, Jillian. Thank you for having me here. | |
| Speaker A | ||
| [00:00:25] | So you are our principal developer advocate at Salesforce. Yes, longtime friend, yes. Good to have you here. | |
| Speaker B | ||
| [00:00:31] | Thank you so much for having me here. | |
| Speaker A | ||
| [00:00:33] | Let's talk about what's going on at Dreamforce for developers. | |
| Speaker B | ||
| [00:00:36] | Yeah, you can see at uh, Dreamforce we have all of these tools in display here and I've seen developers actually learning them, going to this camp trailhead ground, trying all these things and building stuff. And that is amazing because it's not just about you're talking through all these things, you're actually able to get hands on and build things. | |
| Speaker A | ||
| [00:00:59] | I saw there was some stat like almost 2000 people have already built agents using agent forest so far. | |
| Speaker B | ||
| [00:01:05] | That's right. Because it's so simple, we make it so easy for people to get started. It's, yeah, you know, you can easily build and it's just enabling them to be more creative. Like you just have to understand your business problem and then uh, you know, use these tools that we have here at uh, Dreamforce. | |
| Speaker A | ||
| [00:01:26] | So let's talk a little bit about some of those tools. What exactly are some of these new capabilities and tools specifically for developers when it comes to agent force? | |
| Speaker B | ||
| [00:01:33] | Yeah, absolutely. Now before I get into these agents, let's talk about what are some of the ingredients for them. Ingredients for building these agents. It all starts with large language model. You need a large language model. I think at Salesforce we have done very good job by not only just building our own models, also being open, they can bring in all these models and build these applications. Um, but more than model, these agents are going to need knowledge, which is the data and that's what I'm going to talk about. One of the tools that we have for that is data cloud. We had a theater here, intro to data cloud and uh, every time it was packed. So you can see everybody wants to build that data foundation. And if you are a developer, that is where you need to invest your energy on. Understand what are the features of this data cloud, how you can connect Salesforce and other systems with it, how you can customize it, customize to your need so that's about data foundation right there. Uh, and then we have all the interesting stuff, which is how can you bring AI capabilities into your application? And that's where we have promptBuilder. Um, a lot of developers have experienced promptbuilder now because we launched at TDX, we are adding a lot of new capabilities. Uh, and with prompt builders, you can build these prompt, bring that generative AI capabilities that you want into your application using both the low code and the pro code tools that we have. Um, and here, creativity is only your, uh, limit with prompt builder, actually, because I've seen some interesting lightning components that's just going to make your, uh, users more productive. And then finally agent force, build all these agents. | |
| Speaker A | ||
| [00:03:17] | All right, so let's talk about agents. Uh, let's talk a little bit about how developers are going to really harness the power of agent force and start building. What do you envision? What are some of the things that they're going to start doing? | |
| Speaker B | ||
| [00:03:27] | Yes, so, uh, let's understand agent force. So agent force, or agents are next generation systems which can take, uh, basically do all these tasks autonomously, of course, with human intervention and human, uh, in the loop as well. Um, but if you look at agent force, we need to understand what it takes for developers to build. If you're doing your own diY, as Mark said in his keynote, it's a lot of work, a lot of work, and we make it really super easy for you with agent force. With agents, you're bridging that humans with software. So you as developers, you have already built apex flows. And again with promptbuilder, you can build these prompt templates too. So all you're doing is writing natural language instructions on top of it so you can invoke these right things. So you're already building that software stuff and all that is going to be reusable. But with large language models, we'll give you that capability. And now with Agentforce, you write instructions, um, just in natural language and make your thing happen. I mean, it's just the large language model doing the reasoning for you, but the actual software. Again, you as a developer are, ah, building your software system. So it's exciting. | |
| Speaker A | ||
| [00:04:48] | Yeah, I mean, it's like everyone has already been building the building blocks with which then you will assemble to create an agent. | |
| Speaker B | ||
| [00:04:55] | Yes. | |
| Speaker A | ||
| [00:04:56] | Right. I think that's a really helpful story. | |
| Speaker B | ||
| [00:04:58] | Yes. And previously we had to assemble it as a developer. Like I had to go as a developer and say, if this, then do this, then, and then add these branches and logics, right? Uh, now, large language model can do it. If we can describe it in large, you know, describe it in natural language. | |
| Speaker A | ||
| [00:05:15] | Yeah. You can type a sentence. | |
| Speaker B | ||
| [00:05:17] | Yes. | |
| Speaker A | ||
| [00:05:17] | You can build an agent. | |
| Speaker B | ||
| [00:05:18] | Yes. That's how easy it is. | |
| Speaker A | ||
| [00:05:21] | I love that. Okay, so let's talk a little bit about some use cases. Do you have specific use cases that you've seen people already think about when they're using agent force? | |
| Speaker B | ||
| [00:05:28] | Yeah, absolutely. So one of the use cases, very common. And we have this in our trailhead module as well as service agents. Uh, absolutely. Uh, and we showed this at keynote also, like, if you're reaching out to your customer support today, you don't want to be waiting. Uh, and even as a human, the service agent is actually looking at the knowledge base, looking at the company policies, and responding to you. And these agents are working in the same way. So that's a very good use case. Again, we give out of box template and you as a developer can understand your business use case, your processes that you need, use flows and apex to build them. And then just use our agent builder. So, you know, write natural language instructions so all, you know, and then expose it in various channels, like experience cloud site that we are showing in that trailhead module. But it could be, you know, in your Apple messaging, it could be any channel that you can imagine. Slack. Yes, I know you love slack, and. | |
| Speaker A | ||
| [00:06:29] | Yes, I sure do. Yeah. | |
| Speaker B | ||
| [00:06:30] | Yes, on slack too. Expose all these agents and deploy them. | |
| Speaker A | ||
| [00:06:34] | That's great. Okay, so we are, um, wrapping up here, but just one quick note. What is one thing that developers should really start thinking about if they want to skill up for building with agents? | |
| Speaker B | ||
| [00:06:43] | Yes. So only thing that they need to understand in my, uh, experience is learning, uh, how to write natural language instructions, because that's a new skill. Prompt engineering is what we call them. So understanding that and little bit of understanding your domain, like you're working for these customers, understand their business needs because that's what you'll be focusing on. | |
| Speaker A | ||
| [00:07:05] | I love that. All right. Learn how to write sentences instead of code. I love it. Thank you so much, mo, for joining us. Really appreciate you. There are so many resources for developers to get started. You can scan the QR code on your screen to learn more about how to build the future. And have some fun while you're doing it. Take the developer Dreamforce Quest. That link is Sforce Co. DF 24 DevQuest. Thank you so much for watching. |
| Speaker A | ||
| [00:00:00] | Exciting time to be a Salesforce developer. With huge advances in data, AI and agents, the way we build applications is changing and the Salesforce platform provides developers with so much opportunity. In uh, meet the developers and the developer keynote, you'll get a view into the new paradigms for developers and how you can build the future. Now I am joined right now by the amazing Mo mo, Mohit Srivasra. Hello, Mohit. | |
| Speaker B | ||
| [00:00:23] | Thank you, Jillian. Thank you for having me here. | |
| Speaker A | ||
| [00:00:25] | So you are our principal developer advocate at Salesforce. Yes, longtime friend, yes. Good to have you here. | |
| Speaker B | ||
| [00:00:31] | Thank you so much for having me here. | |
| Speaker A | ||
| [00:00:33] | Let's talk about what's going on at Dreamforce for developers. | |
| Speaker B | ||
| [00:00:36] | Yeah, you can see at uh, Dreamforce we have all of these tools in display here and I've seen developers actually learning them, going to this camp trailhead ground, trying all these things and building stuff. And that is amazing because it's not just about you're talking through all these things, you're actually able to get hands on and build things. | |
| Speaker A | ||
| [00:00:59] | I saw there was some stat like almost 2000 people have already built agents using agent forest so far. | |
| Speaker B | ||
| [00:01:05] | That's right. Because it's so simple, we make it so easy for people to get started. It's, yeah, you know, you can easily build and it's just enabling them to be more creative. Like you just have to understand your business problem and then uh, you know, use these tools that we have here at uh, Dreamforce. | |
| Speaker A | ||
| [00:01:26] | So let's talk a little bit about some of those tools. What exactly are some of these new capabilities and tools specifically for developers when it comes to agent force? | |
| Speaker B | ||
| [00:01:33] | Yeah, absolutely. Now before I get into these agents, let's talk about what are some of the ingredients for them. Ingredients for building these agents. It all starts with large language model. You need a large language model. I think at Salesforce we have done very good job by not only just building our own models, also being open, they can bring in all these models and build these applications. Um, but more than model, these agents are going to need knowledge, which is the data and that's what I'm going to talk about. One of the tools that we have for that is data cloud. We had a theater here, intro to data cloud and uh, every time it was packed. So you can see everybody wants to build that data foundation. And if you are a developer, that is where you need to invest your energy on. Understand what are the features of this data cloud, how you can connect Salesforce and other systems with it, how you can customize it, customize to your need so that's about data foundation right there. Uh, and then we have all the interesting stuff, which is how can you bring AI capabilities into your application? And that's where we have promptBuilder. Um, a lot of developers have experienced promptbuilder now because we launched at TDX, we are adding a lot of new capabilities. Uh, and with prompt builders, you can build these prompt, bring that generative AI capabilities that you want into your application using both the low code and the pro code tools that we have. Um, and here, creativity is only your, uh, limit with prompt builder, actually, because I've seen some interesting lightning components that's just going to make your, uh, users more productive. And then finally agent force, build all these agents. | |
| Speaker A | ||
| [00:03:17] | All right, so let's talk about agents. Uh, let's talk a little bit about how developers are going to really harness the power of agent force and start building. What do you envision? What are some of the things that they're going to start doing? | |
| Speaker B | ||
| [00:03:27] | Yes, so, uh, let's understand agent force. So agent force, or agents are next generation systems which can take, uh, basically do all these tasks autonomously, of course, with human intervention and human, uh, in the loop as well. Um, but if you look at agent force, we need to understand what it takes for developers to build. If you're doing your own diY, as Mark said in his keynote, it's a lot of work, a lot of work, and we make it really super easy for you with agent force. With agents, you're bridging that humans with software. So you as developers, you have already built apex flows. And again with promptbuilder, you can build these prompt templates too. So all you're doing is writing natural language instructions on top of it so you can invoke these right things. So you're already building that software stuff and all that is going to be reusable. But with large language models, we'll give you that capability. And now with Agentforce, you write instructions, um, just in natural language and make your thing happen. I mean, it's just the large language model doing the reasoning for you, but the actual software. Again, you as a developer are, ah, building your software system. So it's exciting. | |
| Speaker A | ||
| [00:04:48] | Yeah, I mean, it's like everyone has already been building the building blocks with which then you will assemble to create an agent. | |
| Speaker B | ||
| [00:04:55] | Yes. | |
| Speaker A | ||
| [00:04:56] | Right. I think that's a really helpful story. | |
| Speaker B | ||
| [00:04:58] | Yes. And previously we had to assemble it as a developer. Like I had to go as a developer and say, if this, then do this, then, and then add these branches and logics, right? Uh, now, large language model can do it. If we can describe it in large, you know, describe it in natural language. | |
| Speaker A | ||
| [00:05:15] | Yeah. You can type a sentence. | |
| Speaker B | ||
| [00:05:17] | Yes. | |
| Speaker A | ||
| [00:05:17] | You can build an agent. | |
| Speaker B | ||
| [00:05:18] | Yes. That's how easy it is. | |
| Speaker A | ||
| [00:05:21] | I love that. Okay, so let's talk a little bit about some use cases. Do you have specific use cases that you've seen people already think about when they're using agent force? | |
| Speaker B | ||
| [00:05:28] | Yeah, absolutely. So one of the use cases, very common. And we have this in our trailhead module as well as service agents. Uh, absolutely. Uh, and we showed this at keynote also, like, if you're reaching out to your customer support today, you don't want to be waiting. Uh, and even as a human, the service agent is actually looking at the knowledge base, looking at the company policies, and responding to you. And these agents are working in the same way. So that's a very good use case. Again, we give out of box template and you as a developer can understand your business use case, your processes that you need, use flows and apex to build them. And then just use our agent builder. So, you know, write natural language instructions so all, you know, and then expose it in various channels, like experience cloud site that we are showing in that trailhead module. But it could be, you know, in your Apple messaging, it could be any channel that you can imagine. Slack. Yes, I know you love slack, and. | |
| Speaker A | ||
| [00:06:29] | Yes, I sure do. Yeah. | |
| Speaker B | ||
| [00:06:30] | Yes, on slack too. Expose all these agents and deploy them. | |
| Speaker A | ||
| [00:06:34] | That's great. Okay, so we are, um, wrapping up here, but just one quick note. What is one thing that developers should really start thinking about if they want to skill up for building with agents? | |
| Speaker B | ||
| [00:06:43] | Yes. So only thing that they need to understand in my, uh, experience is learning, uh, how to write natural language instructions, because that's a new skill. Prompt engineering is what we call them. So understanding that and little bit of understanding your domain, like you're working for these customers, understand their business needs because that's what you'll be focusing on. | |
| Speaker A | ||
| [00:07:05] | I love that. All right. Learn how to write sentences instead of code. I love it. Thank you so much, mo, for joining us. Really appreciate you. There are so many resources for developers to get started. You can scan the QR code on your screen to learn more about how to build the future. And have some fun while you're doing it. Take the developer Dreamforce Quest. That link is Sforce Co. DF 24 DevQuest. Thank you so much for watching. |