TorqAgent Walkthrough 03.27.24
Summary
Hi, I'm John Metz, an analyst with IntelliBridge Corporate Economic Threat Intelligence. In this video, I'll be giving you a brief walkthrough of the TorqAgent Module, a key component of IntelliBridge's Torque Cloud Platform. I'll show you how to create a new agent, customize its nodes, and use the available tools to complete tasks. Pay attention to the important features on the page, such as saving the agent, sharing it with others, and embedding it in other workflows. Stay tuned for a demo of the agent in action, where I'll ask it a question about a specific company tied to the Chinese government. Enjoy!
TorqAgent Walkthrough 03.27.2024
Transcript
0:00 Hello, my name is John Metz, and I'm an analyst with IntelliBridge Corporate Economic Threat Intelligence or CETI. Today I'm going to give you a brief walk through of the Torque Agent Module, which is one of the modules available as part of IntelliBridge Torque Cloud Platform.
0:15 This is something that I use extensively on a day-to-day basis, so I'm going to be speaking to you in this and in other videos as someone who's really seen firsthand the benefits of this module.
0:25 So So from the Torque Cloud Home page, I'm going to click on the Torque Agent Module here and the hand side, which will take me to this page.
0:32 To make a new agent, I'm going to click on add new up in the right hand side, and that'll take me here.
0:38 The basic building blocks for agents are called nodes, and to add a node, I'll click the plus button right here.
0:43 It pulls up this drop down menu, where I can choose a type of agent and a variety of other nodes, including chat models, document loaders, and vector stores for accessing datasets, and chains to multiple agents together.
0:55 But for now, I'm going to keep this very simple. First, I'll select an agent, and I'm going to go with Torque AGI, which is our most standard versatile agent.
1:04 As you can see here, this node can be customized by connecting with other nodes through a simple drag and drop interface.
1:10 Some of these nodes are required. This particular agent requires an open AI chat model, a memory node, and at least one tool which can use to respond to user queries.
1:18 So I'll go ahead and add those nodes, then connect them to these dots to give the agent access to them.
1:24 Starting with the chat model, can see that the chat model itself is customizable, a memory, and I'll go with buffer memory, which is our most basic standard option.
1:49 And then at least one tool, and so I'll go with the web search tool, which will give this agent access to the web, to gather information and respond to my queries.
2:01 We also have the option of adding a prompt template, although in this case that isn't required. As I've connected all those nodes, I've got a working basic AI agent that's going to use the available tools to complete the tasks I assign it.
2:14 There are a few more important features on this page. First, only to save the agent before I use it and after making any changes.
2:21 Only to give it a name and in this case I'll just call it demo agent, and then today's date. I can change that name at any time.
2:33 This gear icon here allows me to duplicate this chat flow, upload another, export or download this chat flow, and delete it.
2:43 The arrow button right here allows me to share this with another user. And they do need to be a torque cloud user.
2:50 So let's type in the email and hit share. And now there are two users on this. On the left hand side here there's an embed functionality so that I can embed this in another workflow or website.
3:04 To do so without an API key I can just click on no authorization here. Hit the copy button which looks like a clipboard.
3:14 And voila I've now copied this. Last but not least, I can click on this button here to save any chat flow as a template to help.
3:25 So now that we've done all of that, I'll click the chat icon here to go to chat with my agent.
3:31 The bar on the side here lists information about the agent like the agent type, what tools it is access to and what language model we're using as well as any ongoing or completed tasks.
3:41 So I'm going to ask this agent a question and since the CETI team is focused on threats connected to state sponsored actors, I'm going to ask about a specific company tied to the Chinese government.
3:51 So is SenseTime sanctioned by the US government? I'll click this button here to ask that question. As you can see, the agent is using the search tool which I gave it access to earlier to complete this task and here's its final answer.
4:10 One last feature I wanted to show you in this overview is this which is the agent evaluator. That's when I click on this, it'll show me what the input was, how the agent responded and various different measures about quality which you can use to evaluate the quality of the response.
4:25 You can also use this tool to measure sentiment and potential risks. Like the risk of PII leaks. Finally I'll click on this pencil icon to go back to the chat flow view and that concludes this overview of what I've found to be a really valuable and easy to use, low or no code option that's really immeasurably
4:46 enhanced the productivity of my team.
Hi, I'm John Metz, an analyst with IntelliBridge Corporate Economic Threat Intelligence. In this video, I'll be giving you a brief walkthrough of the TorqAgent Module, a key component of IntelliBridge's Torque Cloud Platform. I'll show you how to create a new agent, customize its nodes, and use the available tools to complete tasks. Pay attention to the important features on the page, such as saving the agent, sharing it with others, and embedding it in other workflows. Stay tuned for a demo of the agent in action, where I'll ask it a question about a specific company tied to the Chinese government. Enjoy!
TorqAgent Walkthrough 03.27.2024
Transcript
0:00 Hello, my name is John Metz, and I'm an analyst with IntelliBridge Corporate Economic Threat Intelligence or CETI. Today I'm going to give you a brief walk through of the Torque Agent Module, which is one of the modules available as part of IntelliBridge Torque Cloud Platform.
0:15 This is something that I use extensively on a day-to-day basis, so I'm going to be speaking to you in this and in other videos as someone who's really seen firsthand the benefits of this module.
0:25 So So from the Torque Cloud Home page, I'm going to click on the Torque Agent Module here and the hand side, which will take me to this page.
0:32 To make a new agent, I'm going to click on add new up in the right hand side, and that'll take me here.
0:38 The basic building blocks for agents are called nodes, and to add a node, I'll click the plus button right here.
0:43 It pulls up this drop down menu, where I can choose a type of agent and a variety of other nodes, including chat models, document loaders, and vector stores for accessing datasets, and chains to multiple agents together.
0:55 But for now, I'm going to keep this very simple. First, I'll select an agent, and I'm going to go with Torque AGI, which is our most standard versatile agent.
1:04 As you can see here, this node can be customized by connecting with other nodes through a simple drag and drop interface.
1:10 Some of these nodes are required. This particular agent requires an open AI chat model, a memory node, and at least one tool which can use to respond to user queries.
1:18 So I'll go ahead and add those nodes, then connect them to these dots to give the agent access to them.
1:24 Starting with the chat model, can see that the chat model itself is customizable, a memory, and I'll go with buffer memory, which is our most basic standard option.
1:49 And then at least one tool, and so I'll go with the web search tool, which will give this agent access to the web, to gather information and respond to my queries.
2:01 We also have the option of adding a prompt template, although in this case that isn't required. As I've connected all those nodes, I've got a working basic AI agent that's going to use the available tools to complete the tasks I assign it.
2:14 There are a few more important features on this page. First, only to save the agent before I use it and after making any changes.
2:21 Only to give it a name and in this case I'll just call it demo agent, and then today's date. I can change that name at any time.
2:33 This gear icon here allows me to duplicate this chat flow, upload another, export or download this chat flow, and delete it.
2:43 The arrow button right here allows me to share this with another user. And they do need to be a torque cloud user.
2:50 So let's type in the email and hit share. And now there are two users on this. On the left hand side here there's an embed functionality so that I can embed this in another workflow or website.
3:04 To do so without an API key I can just click on no authorization here. Hit the copy button which looks like a clipboard.
3:14 And voila I've now copied this. Last but not least, I can click on this button here to save any chat flow as a template to help.
3:25 So now that we've done all of that, I'll click the chat icon here to go to chat with my agent.
3:31 The bar on the side here lists information about the agent like the agent type, what tools it is access to and what language model we're using as well as any ongoing or completed tasks.
3:41 So I'm going to ask this agent a question and since the CETI team is focused on threats connected to state sponsored actors, I'm going to ask about a specific company tied to the Chinese government.
3:51 So is SenseTime sanctioned by the US government? I'll click this button here to ask that question. As you can see, the agent is using the search tool which I gave it access to earlier to complete this task and here's its final answer.
4:10 One last feature I wanted to show you in this overview is this which is the agent evaluator. That's when I click on this, it'll show me what the input was, how the agent responded and various different measures about quality which you can use to evaluate the quality of the response.
4:25 You can also use this tool to measure sentiment and potential risks. Like the risk of PII leaks. Finally I'll click on this pencil icon to go back to the chat flow view and that concludes this overview of what I've found to be a really valuable and easy to use, low or no code option that's really immeasurably
4:46 enhanced the productivity of my team.
Updated on: 20/02/2025
Thank you!