Amazon Bedrock Agents | Amazon Web Services

Summary

In this video, Mark Roy, a Principal ML Specialist Solution Architect at AWS, introduces 'Agents for Amazon Bedrock'. This service enables the creation of applications that can execute complex business tasks using natural language interactions. It leverages Amazon Bedrock's large language models to decompose tasks into actionable steps, utilizing existing APIs, applications, and databases. The demo showcases a CRM agent that uses historical data to identify customer concerns and suggest meeting topics, demonstrating the powerful integration of AI with business processes.

Takeaways

• 🧑‍💻 The speaker is Mark Roy, a principal ML specialist solution architect at AWS.
• 🌟 Agents for Amazon Bedrock is a service that allows building applications which can execute multi-step business tasks.
• 🤖 Generative AI is powerful but lacks the ability to take actions; Agents for Amazon Bedrock fills this gap.
• 🔧 Agents can interact in natural language, utilize large language models, and take actions using existing APIs and databases.
• 🛡️ It's a fully managed service that ensures security and provides control over access to agents and actions.
• 📈 Agents can break down requests into steps, execute them, and provide transparent insights into the execution plan.
• 🔄 The process involves constructing an LLM prompt, invoking Amazon Bedrock, and using Chain of Thought reasoning to execute tasks.
• 💾 The demo showcases a CRM agent that leverages CRM data to help salespeople plan their work.
• 📅 The agent can suggest meeting topics, agendas, and preferred times based on customer interactions and preferences.
• 🔗 Agents can be integrated into custom applications or used standalone through the AWS console or API.
• 📚 Resources are available for getting started with agents, including a blog post and an online workshop.

Q & A

What is the main purpose of Agents for Amazon Bedrock?

Agents for Amazon Bedrock allows customers to build applications that can execute multi-step business tasks using natural language, leveraging the power of large language models (LLMs) and integrating with existing APIs, applications, databases, and knowledge stores.

How does Agents for Amazon Bedrock differ from traditional applications?

Agents for Amazon Bedrock differs from traditional applications by not being hard-coded. Instead, it dynamically comes up with a plan on its own, given a set of tools, actions, and knowledge bases, and then executes that plan for the user.

What is the 'secret sauce' feature of Agents for Amazon Bedrock mentioned in the script?

The 'secret sauce' feature of Agents for Amazon Bedrock is its ability to take a user's request, break it down into a set of steps, and execute those steps on behalf of the user, all while being fully secure and managed.

How does the orchestration process work in Agents for Amazon Bedrock?

The orchestration process in Agents for Amazon Bedrock involves decomposing a task into steps, constructing a proper LLM prompt, invoking Bedrock, and then using Chain of Thought reasoning to execute each step using actions or knowledge bases until the task is complete.

What is Chain of Thought reasoning as mentioned in the script?

Chain of Thought reasoning is a method used by Agents for Amazon Bedrock to figure out a plan and execute each step using either an action or a knowledge base to complete a task.

How does transparency work in Agents for Amazon Bedrock?

Transparency in Agents for Amazon Bedrock is achieved by providing access to the plan that was created and how it was executed, allowing users to understand the decision-making process behind the scenes.

What is an example of an agent built in the demo in the script?

In the demo, a CRM (Customer Relationship Management) agent was built to leverage data in a CRM system, helping salespeople plan their work by interacting with customer data, opportunities, and historical activities.

How can users interact with Agents for Amazon Bedrock?

Users can interact with Agents for Amazon Bedrock through the test console while developing the agent, or by using the API to integrate the agent into their own applications or scripts.

What actions are available to the CRM agent demonstrated in the script?

The CRM agent has actions available such as getting a company overview, retrieving recent interactions or activities with details like dates, times, and meeting notes, and accessing customer preferences.

How does the CRM agent use the information from recent interactions to suggest a meeting topic and agenda?

The CRM agent uses the information from recent interactions by analyzing the customer's concerns and interests, then suggesting a meeting topic and agenda that aligns with those interests, such as an overview of smartship capabilities, pricing, and packages.

What resources are available for users to get started with Agents for Amazon Bedrock?

Users can refer to the announcement blog post for more information and participate in an online workshop that covers Amazon Bedrock overall with a specific module on Agents for Amazon Bedrock.

Outlines

00:00

🧠 Introduction to Amazon Bedrock Agents

Mark Roy, a principal ML specialist solution architect at AWS, introduces agents for Amazon Bedrock, a service that allows customers to build applications using large language models that can execute multi-step business tasks. These agents can interact in natural language and leverage existing APIs, applications, databases, and knowledge stores. The service is fully managed, secure, and provides control over access to agents, actions, and knowledge bases. It also offers transparency into the decision-making process behind task execution.

05:01

🔍 Demonstrating Amazon Bedrock Agents

The video script walks through a demo of a CRM agent built with Amazon Bedrock that helps salespeople manage customer interactions. The agent uses Amazon Bedrock's large language model, CLA 2.1, and is equipped with actions to retrieve a company overview, recent interactions, and customer preferences. The agent is shown to take a natural language request, break it down into steps, execute them by interacting with the CRM, and provide a comprehensive response. The demo also illustrates how to trace the agent's execution steps and how to integrate the agent into a custom application using the provided API.

10:01

🛠 Building Applications with Amazon Bedrock Agents

The script explains how to build applications using Amazon Bedrock agents by leveraging existing APIs and databases. It demonstrates a simple Lambda function in Python that handles events from the agent, retrieves necessary information, and returns it as an action response. The agent then uses this information to complete tasks. The video concludes with resources for getting started with agents, including an announcement blog post and an online workshop, and encourages viewers to explore building applications with Amazon Bedrock agents.

Keywords

💡Amazon Bedrock

Amazon Bedrock is a service that provides large language models (LLMs) to customers. It is a key component of the video's theme, as it is the platform on which agents operate. The script mentions that Amazon Bedrock is used across various industries and use cases, highlighting its versatility and power. The video discusses how agents for Amazon Bedrock can take advantage of these LLMs to execute multi-step business tasks.

💡Agents

Agents, in the context of the video, refers to a feature of Amazon Bedrock that allows for the creation of applications capable of executing complex, multi-step business tasks. They are integral to the video's narrative as they represent an advancement in generative AI, moving beyond mere data processing to taking actionable steps. The script provides a demo of how agents can interact with CRM data to provide insights and plan actions.

💡Generative AI

Generative AI is a type of artificial intelligence that can generate new content, such as text or images, based on existing data. In the video, generative AI is noted as being widely used, but the script emphasizes the need to move beyond generation to actionable tasks, which is where agents for Amazon Bedrock come into play.

💡Multi-step business tasks

This phrase refers to the ability of agents to perform a series of actions in sequence to accomplish a business goal. The video explains that agents can break down complex requests into a set of steps and execute them, which is a significant advancement in AI capabilities. An example from the script is the CRM agent that plans and executes steps to identify customer concerns based on recent interactions.

💡Natural Language

Natural language is the conversational language that people speak in their daily lives, as opposed to formal or computer languages. The video emphasizes that agents can be interacted with in natural language, making them more accessible and user-friendly. An example is the request for a meeting topic and agenda based on customer interests and preferences.

💡APIs and Applications

APIs (Application Programming Interfaces) and applications are tools that agents can interact with to perform tasks. The script mentions that agents can leverage existing APIs and applications, which means they can be integrated into a company's current tech stack to automate processes. This is showcased in the demo where the agent uses actions to interact with a CRM system.

💡Knowledge Stores

Knowledge stores refer to databases or repositories of information that agents can access to perform their tasks. In the video, knowledge stores are depicted as a resource that agents can use to gather information and make decisions, similar to how humans would use knowledge to solve problems.

💡Secure and Managed Service

This term highlights that Amazon Bedrock's agents are not only powerful but also secure and easy to manage. The video script assures viewers that the service is fully managed, meaning AWS takes care of the infrastructure and maintenance, and that it is secure, with controls over who can access the agents and their actions.

💡Transparent Execution

Transparent execution refers to the ability to see how an agent has processed a request and carried out actions. The script explains that agents provide a trace of their decision-making process, which is valuable for understanding and auditing how tasks are completed. This transparency is a key feature that differentiates agents from traditional black-box AI systems.

💡Orchestration

Orchestration, in the context of the video, refers to the process by which agents decompose tasks into steps and then execute those steps in an organized manner. It is a core concept in the video's demonstration of agent functionality. The script describes how agents use a 'Chain of Thought' reasoning to figure out a plan and execute it, which is a form of orchestration.

💡Lambda Functions

Lambda Functions are serverless computing functions provided by AWS that can be used to run code without provisioning or managing servers. The video script mentions using Lambda functions to interact with databases and APIs, which simplifies the process of building applications that leverage agents. An example given is a Python Lambda function that retrieves meeting results to fulfill an agent's request.

Highlights

Introduction to Agents for Amazon Bedrock by Mark Roy, Principal ML Specialist Solution Architect at AWS.

Generative AI is widely used across industries but lacks the ability to take action.

Agents for Amazon Bedrock allows building applications that can execute multi-step business tasks.

Agents can interact in natural language and utilize existing APIs, applications, databases, and knowledge stores.

Agents uses the power of Bedrock's large language models.

Agents can break down requests into steps and execute them autonomously.

The service is fully secure, managed, and provides control over access to agents and actions.

Agents provide transparency on the execution plan and how tasks are completed.

Orchestration of tasks by Agents involves decomposing tasks into executable steps.

Agents use Chain of Thought reasoning to figure out a plan and execute steps.

Demo of Amazon Bedrock console showing how to create and manage agents.

Example of a CRM agent that leverages CRM data to assist salespeople.

Agents can be invoked through the test console or API for development.

Detailed trace of agent's multi-step plan execution shown in demo.

Agents can have natural language conversations and make decisions based on previous interactions.

Agents can suggest meeting topics and agendas based on customer interests and preferences.

Agents can be integrated into custom applications or scripts.

APIs are available for invoking agents and creating agents.

Lambda functions can be used to connect Agents with databases and APIs.

Resources provided for getting started with Agents for Amazon Bedrock.

Encouragement for viewers to explore and build with Agents for Amazon Bedrock.

Transcripts

welcome my name is Mark Roy and I'm a

principal ml specialist solution

architect at

AWS and in this brief video I'm going to

take you through an overview of agents

for Amazon Bedrock as well as give you a

quick demo of how it works so let's Jump

Right In uh first of all let me cover

what is agents for Amazon Bedrock uh and

as you probably know uh generative AI

has been being used by

pretty much every customer and across

all use cases and and industries so it's

uh and it's been shown to be extremely

powerful one of the things it's not

doing however is taking action so

customers are looking to build Solutions

using large language models services

like Amazon Bedrock but they want to

take it to the next step uh so what

agents allows you to do is build these

applications that can execute multi-step

business tasks you can interact with it

in natural language take advantage of

the full power of llms large language

models uh under the hood but also take

actions by taking advantage of your

existing apis and applications databases

and and uh knowledge stores so a quick

look rundown of the features uh that

agents provides one it uses the power of

bedrocks large language models uh

secondly and this is really where the

secret sauce comes in uh it's able to

take your request break that down into a

set of steps and execute that on your

behalf so it's not a hard-coded

application uh it's actually coming up

with a plan on its own given a set of

tools or actions and knowledge bases and

then executing that for you uh it's also

done in a fully secure fashion this is a

fully managed service and you get full

control over for who can use which uh

agents and actions and knowledge bases

and then lastly uh because this is a a a

uh it's making decisions uh under the

hood about how to execute it also gives

you access to uh what that plan was and

how it executed that uh

transparently so because the

orchestration is something kind of new

uh and interesting let me dive into that

a little bit further so when you use an

agent you hand it a task and agents for

Amazon Bedrock decomposes that it

figures out a plan what are the steps

that I need to do to get that job done

on your behalf uh and based on that plan

it will execute uh various actions or do

some uh searches across uh multiple

knowledge bases it will look at the

results that it gets and and proceed to

The Next Step until it's got uh

uh the U task complete until it has a

final answer let's double click on that

a little bit more uh deeply and then

I'll jump into a demo so here we've got

a Bedrock agent been that's been handed

a task uh and under the hood it does all

of the heavy lifting of constructing a

proper llm prompt it takes into account

all of the conversation history you've

been having with the agent it gives the

metadata about which action

what tools are available and what

knowledge bases are available it gives

it the set of instructions that the

agent Builder provided uh and then it

will invoke uh Bedrock under the hood

and and then uh the agent will figure

out a plan using what's called Chain of

Thought uh reasoning here and it will

execute each of those steps using either

an action or a knowledge base and when

it gets the full set of results it will

then use Bedrock again uh to prepare a

final response using natural language uh

back to the actual

user very powerful new way to go about

taking full advantage of uh large

language models so let's see a demo of

how that

works uh so what you're looking at here

is the Amazon Bedrock console I've

clicked on agents and I've brought up uh

an agent that I've already built this

agent is is a CRM agent or customer

relationship management agent uh that

makes it easy to uh leverage um the data

that's available uh in your CRM so the

uh let's say I'm an account manager and

I work with multiple different

customers uh the CRM lets me keep track

of who those customers are the

opportunities uh that I'm working on uh

with them as well as the history of all

the different uh activ activities let's

say meetings and discussions and events

and so forth uh so all of that's

available so uh when you define an agent

you give it a

name uh and uh here I've got a set of

versions that I've been building let's

dive into the current working draft so

you've got the name of your agent and

then a set of instructions uh so here

I'm saying that you are a customer

relationship management agent that helps

salespeople plan their work uh so forth

and so so on uh and you also identify

which Amazon Bedrock large language

model you're using in this case we're

using CLA uh

2.1 and then you define a set of

knowledge bases you want to make

available to it as well as uh a set of

actions or apis that are

available in this case I'm just giving

it a couple of actions uh one an ability

to get back a company

overview uh a second one

that lets uh the agent get back the set

of recent interactions or activities uh

and give details back about that like

for

example um uh an array of um dates uh

times uh as well as

um uh the meeting notes uh from those

interactions and then lastly uh it has

an ability to get the customer

preferences so these are the uh actions

that are uh available to this agent and

let's see how the agent uh works so you

can either use the agent from the test

console as you're developing it or you

can use the

API so here's an example call that I

made uh before we started uh this video

uh I asked the agent to help me out with

uh customer

CABC given our recent interactions what

are their top concerns and interests and

it came back with an answer saying hey

they're worried about

reliability and they have a strong

interest in our upcoming Service uh

particularly around pricing and if I

show the detailed trace of what went on

uh behind the scenes uh you can see that

it came up with a multi-step plan four

steps and then it executed uh those

steps the first step was to find the

interactions it passed in the customer

ID which it was able to uh pull from my

request

uh and then it uh invoked uh the action

which interacted with the CRM itself

gave back a set of details here about

all the meetings that we've had when

they happened what the results were uh

and then it uh determines that it's got

enough information and it gives back a

final response basically summarizing

what were the concerns based on uh that

activity uh so pretty powerful and then

you can actually have a conversation as

as well so here I'm going to have a

followup request based on those

interests suggest a next meeting topic

and an agenda and let me know what day

and time is preferred so it's able to

leverage what it is already seen to come

up with a meeting topic which is

basically uh an overview of the

smartship capabilities and it proposed

an agenda uh smartship overview pricing

and packages demo

Q&A uh you notice it came back with a uh

suggested meeting time of Wednesday

afternoon and that was based on its

ability to uh take an action on my

behalf to look up the customer

preferences so I had all of this

powerful interaction with really just a

couple of simple natural language

requests and agents for Amazon Bedrock

uh is able to interpret those uh and

take action on my behalf dynamic Ally

coming up with a plan based on what are

the actions available and what are the

knowledge bases uh

available taking that a little bit

further in addition to using it from the

test console as you're building an

agent you can

also uh write your own uh application or

integrate an agent with your own

application uh you could use it from a

script uh you could use it from an

existing

um tool or website or your uh you know

uh iOS application

and there's a full API to invoke the

agent uh as well as create agents uh as

well so in this case um I built a simple

user interface that takes in a request

it calls invoke agent uh because we

deployed that agent and made Avail made

it available uh to the

API uh and as it it's uh proceeding the

API gives you back a trace of what's

happening so here you see the steps that

it was planning to do uh and then the

first step that it uh took on my behalf

what parameters it passed what results

it got back and then it takes a another

step um comes back with some

results and then it figures out it's got

enough information so it uh returns that

back uh to the invoke agent uh response

and my application would really just

show this uh I could save this Trace

information into a log file if uh if I

wanted to for example uh so you can set

up a a bot of your own uh that will take

natural language input use one or more

agents under the hood and the Agents

will uh take advantage of your existing

apis and databases and document stores

uh to get work done very simple Le to

build um in the way that you hook into

your databases and apis is to write

simple Lambda functions uh in whatever

language uh you'd like in this case I've

got a python Lambda function uh so we

take in an event the agent hands you an

event uh you can look at what action the

agent wanted you to take in this example

uh if I am asked to list recent

interactions call a function uh that

will get that

done in this simple demo I've just uh

retrieved the customer ID parameter and

the comp parameter and then I've got a

hardcoded set of uh meeting results that

I'm handing back in the real world this

is where I would call the CRM API uh to

get that job done and uh then I simply

uh return uh that back as part of the

action response and the agent uh API

takes care of the rest so very

straightforward to use very powerful new

way to look at building applications

using large language

models so with that uh um let me also

point out that we've got some resources

for you to help you get started uh on

the left uh this will get you to our uh

announcement blog post uh to help you

understand uh a little bit more about

agents and on the right we've got a an

online Workshop uh that you can uh learn

about Amazon Bedrock overall with a set

of different modules including a module

specifically on agents for Amazon

bedrock with that thank you for watching

and uh let us know what you go out and

build with agents for Amazon

bedrock