How I built an AI study assistant with Amazon Q
Turn your docs into a chat buddy - my easy guide to making a Q app!
Brooke Jamieson
Amazon Employee
Published Mar 28, 2024
We’ve all been there - wasting precious time looking through scattered docs, databases, and websites trying to piece together the specific information we need to answer a question or complete a task. Hunting for information and manually piecing it all together is frustrating and a productivity drain. Imagine having an AI assistant that could instantly synthesize all that distributed data and give you clear, relevant answers through natural conversation.
This post will show you just how easy it is to build your own Amazon Q app to unlock the valuable information you already have. Keep scrolling to see how I used this to make my own AI study assistant with Amazon Q for my New York Drivers License test too!
While you may have seen Amazon Q in the AWS Management Console or your IDE, it’s actually a broad AI service with applications for both technical and business users. At it’s core, Amazon Q is a generative AI assistant powered by Large Language Models (LLMs) from Amazon Bedrock’s foundation model service. It can understand natural language queries and give you accurate responses by analyzing and synthesizing information across the data sources you connect.
For businesses, Amazon Q offers a way to make distributed organizational knowledge more accessible. It integrates a wide range of data like docs, databases, websites and other services with pre-built connectors. This means your users can ask questions conversationally and get synthesized answers from all those connected sources. Admins have full control, and can define topic areas and specify what information is used to generate responses.
For developers, Amazon Q provides relevant technical guidance and capabilities in the AWS Management Console, your IDE and via an API. You can get customized advice and embed this experience into applications. There’s also Code Transformation capabilities to help you maintain, upgrade and migrate your applications and an Amazon Q data integration in AWS Glue to help you integrate data with natural language.
Building your own Amazon Q application to unify and query your organization’s distributed data is a simple process in the AWS Management Console.
First, name your app, then select a retriever and connect relevant data sources like docs, databases, external services and websites. Amazon Q indexes and integrates all of these sources for you.
Next, you can optionally integrate other services with pre-built plugins, set admin controls over things like security and topic areas, and define content guardrails.
The third step lets you preview and customize the look and interactions of the end-user chat experience for your new AI assistant.
Finally, you can deploy the completed web experience by integrating with your identity provider for secure user access, then share the URL.
That’s it - in just 4 steps, you’ve created and customized an interactive AI assistant capable of answering contextualized questions by pulling from all of your connected data sources through natural language queries.
I relocated from Australia to New York City for my job with AWS in 2022. While I have an Australian and an International drivers license, I still need to take a written and practical test to get a New York state license. Studying for this is a challenge - the NY DMV driver’s manual with all the relevant information is a massive PDF, but it’s difficult to navigate and has some regional language that I’m not super familiar with, even through I’m a competent driver.
To help me study for my test, I built my own Drivers License Study Buddy app with Amazon Q. I simply uploaded the official manual PDF as the data source, did some basic setup like naming the app, and a quick customization to change the text on the user interface to make it more relevant. Now, I can ask natural language questions like “How do I safely drive through snow?” (which is something I’ve never really had to think about before when driving in sunny Queensland) and my AI study tool consults the DMV manual, synthesizes the relevant sections, and gives me clear, cited answers pulled rom the official guide.
It’s super easy for me to ask clarifying questions and get quick answers when I need them, which gives me more time to focus on learning to parallel park on the opposite side of the road than I’m used to.
While my Drivers License Study Buddy app is just a personal proof of concept, it really highlights the broad potential for organizations (with much more data than a single pdf) to turn their own documents and data sources into interactive AI assistants with Amazon Q. Imagine connecting your company’s sales data repositories, research libraries, customer interaction logs, product documentation, and more.
There’s a full rundown on the Getting started with Amazon Q (Preview) page, but the important thing to note here is that there’s 3 key ways to begin with Amazon Q.
- Business: You can make a generative AI application for business (like my little drivers license study tool) by following the 4-step process outlined above.
- Management Console: If you’d like to ask a question about AWS, you can click the Amazon Q icon in the right sidebar of the AWS Management Console.
- IDE: You can access Amazon Q by adding the AWS Toolkit to your IDE and authenticating through IAM Identity Center or AWS Builder ID.
There you have it - with just a few clicks in the AWS console, you can turn scattered information into an interactive AI assistant with Amazon Q. Whether it’s technical docs, research papers, operational data or something as simple as a PDF study manual, Amazon Q makes it easy to connect the data and start asking questions in natural language.
While my little drivers license study tool was just a personal toy project (and I promise this is not the only study I’m doing!) it really highlights the broader potential here.
Amazon Q is still in preview, so now is the perfect time to start experimenting! Build your own AI app, and let me know in the comments what you have created, or what wild questions you’ve asked it!
What will you dream up? This early feedback helps shapes the service for everyone. I can’t wait to see how you all leverage this technology to be less frustrated and more productive.
Any opinions in this post are those of the individual author and may not reflect the opinions of AWS.