What is Amazon Q? The AI chatbot for companies

Since the release of ChatGPT, the chatbot has been quickly embraced by everyday users, but treated with caution by businesses and CIOs. While AI chatbots have massive potential for driving efficiencies and customer value, they also present a security risk. Even with tools with ChatGPT Enterprise that attempt to bridge this gap, you need to manually input business data piece by piece to get any insights. 

Announced at re:Invent 2023, Amazon Q changes all that. According to AWS, it’s the first AI chatbot that is truly built with business in mind from the ground up. The breadth of the features Q offers is quite exciting for both businesses and technologists.

• What is Amazon Q?
• How does Amazon Q help out developers and IT experts?
  • Q&A for learning and using AWS
  • Instance selection for EC2
  • Troubleshooting errors and networking issues
  • Integration with your IDEs
  • Automating the app development life cycle, writing tests
• How do I get access to Amazon Q?
• Does Amazon Q take user permissions into account?
• How much does Amazon Q cost?
• What data sources can I connect to Amazon Q?
• Can I add my own custom connectors to Amazon Q?
• How do I tailor Amazon Q to my business? A step-by-step guide
• Can I connect Amazon Q to enterprise systems?

Amazon Q is an generative AI chatbot for workplaces that can be tailored to your business. You can plug it into data sources like Amazon S3, Microsoft 365, GitHub, Dropbox, OneDrive, Salesforce, Slack, and more. Prioritizing security and privacy, customer content from Amazon Q is never used for model training.

Just like ChatGPT, you can use Amazon Q to have conversations, solve problems, generate content, and gain insights. However, with Amazon Q it knows your business in a way ChatGPT can’t, because it’s plugged into all that data. Just like Q from Star Trek, it acts as an all knowing being who can answer any question (or at least, when it comes to your organization).

For example, you can ask Amazon Q where you had the highest sales, and get a response customized to your business data. Another example is you could give it access to your blog and get it to craft social media wrap-up posts from your latest content, while referencing your corporate style guide.

When it asks a question, Amazon Q always provides the source it pulled it from (unlike ChatGPT which can be a bit hit and miss). This means you’re always aware of what business reports it’s using, so you can fact check it.

Amazon Q is built into services like AWS Quicksight, so you can use it to build visuals and dashboards. It’s also now in Amazon Connect for helping customer service agents by automatically detecting customer intent during calls, providing agents with immediate, real-time responses and actions, and links to reference documents. 

Pro tip: Check out this Amazon Q Slack Gateway git repository that makes it so you can use Amazon Q as a Slack Bot. Yes, that’s right — you can use this business-trained chatbot in Slack!

If you’re building applications or maintaining infrastructure on AWS, Q is a game changer. Trained on 17 years of AWS expertise, it can teach you how to get started with AWS, what services you should use, how you should set up architecture, resolve errors, and help you code new features.

NOTE: These features are part of the Amazon Q Builder plan (See the Pricing section in this article), and not included in the Amazon Q Business plan.

First, you’ve got Amazon Q’s Q&A capability, located on the right hand side of the management console. If you’re familiar with any sort of chatbot, this is fairly intuitive (A colleague of mine called it “An AWS version of Clippy”).

You can ask it stuff like “I’m planning to create Serverless APIs with 100/k requests a day. What services should I use?” and Q will be able to recommend what will fit best. Given there’s over 200 AWS services that change all the time, this is a real lifesaver. This Q&A feature is integrated into the AWS management console, AWS console mobile application, AWS documentation, AWS websites, and slack and teams through AWS Chatbot.

You can also get instance selection advice for EC2 for your workload, instead of scratching your head about what’s going to work for you. To access this, type “What instance families should I use to deploy a Web App Server for hosting an application?” (Yes, it’s a secret password) or just launch the Amazon EC2 console, go into Instance type, and select Get advice on instance type. 

You can also ask something more niche, like “What EC2 instance types provide the highest performance for video encoding and transcoding workloads supporting gaming use cases?” All of this takes out some of the headache of architecting.

With Amazon Q, you can troubleshoot errors as well in the console. You may notice these buttons with “Troubleshoot with Amazon Q” popping up in your AWS Lambda function tests now when they fail (depending on your region, since Q’s in preview) which is pretty neat.

By clicking on this, and then Research the error, Amazon Q suggests a fix for your particular issue, so you’re not spending time looking it up yourself.

Another “secret password” you can use with Amazon Q is to ask it about network connectivity issues. Just ask something like “Why can’t I SSH to my EC2 instance?” or “Why can’t I reach my web server from the internet”, and it will analyze your end-to-end connection to see what’s wrong (and tell it Adam sent you. Ipsen, not Selipsky, though you’ll probably get faster results with the latter.)

Amazon Q integrates with all supported IDEs, which is great news for developers. If you’re unfamiliar with the code in the IDE, or need a description of what an application does and how it works, you can ask Q and CodeWhisperer to explain it. Because Q has a deep knowledge of AWS and your code bases, these descriptions can be quite detailed. E.g. Explaining the functionality of your support ticket system using Python Flask and AWS Lambda.

When it comes to the daunting task of creating an app — planning, coding, testing, and documentation — Q can help. You can use a combination of Amazon CodeWhisperer to generate chunks of new code, and leverage Q to do the rest — drafting, review, and publishing features to your application. 

Now, you just write a prompt, and Q does the heavy lifting for you. Q writes a draft plan, which you can collaborate and iteratively improve until it’s polished. Then, Q can implement the plan across the files in your app, while you remain in control reviewing any changes. You can also use Q to generate tests for your functions by simply asking it, which speeds up your QA process.

Amazon Q is currently in preview in AWS regions US East (N. Virginia) and US West (Oregon). However, the Amazon Q conversational Q&A capability is in preview across all commercial regions. To check out the business offering, visit this product page and click “Get started with Amazon Q today.”

Yes. Amazon has taken the rule of least privilege into account with Amazon Q, and it is aware of what systems they can access. If a staff member can’t access data without Q, they can’t access it with Q either.

To set this up, you can integrate Amazon Q with your external SAML 2.0-supported identity provider, like Okta, Azure AD, and Ping identity. Amazon Q lets admins tweak how it works with configurable controls. Q accesses access control list (ACL) permissions to inform what it should show people or not.

The big question! Amazon Q has two plans, Amazon Q Business at $20/month per user, and Amazon Q Builder at $25/month per user. The former is the base plan that lets you plug into all your business data, while the latter also has all the AWS knowledge and developer features. To learn more about this, check out the Q pricing page.

Amazon Q offers over 40 built-in connectors for popular applications and document repositories. These include Amazon S3, Salesforce, Google Drive, Microsoft 365, ServiceNow, Gmail, Slack, Atlassian, and Zendesk.

Below is a full list of supported connectors:

• AEM (Cloud)
• AEM (Server)
• Alfresco (Cloud)
• Alfresco (Server)
• Aurora (MySQL)
• Aurora (PostgreSQL)
• Amazon FSx Windows
• Amazon RDS (Microsoft SQL Server)
• Amazon RDS (MySQL)
• Amazon RDS (Oracle)
• Amazon RDS (PostgreSQL)
• Amazon S3
• Amazon Q custom connector
• Amazon Q Web Crawler
• Amazon WorkDocs
• Box
• Confluence (Cloud)
• Confluence (Server)
• Dropbox
• Drupal
• GitHub (Cloud)
• GitHub (Server)
• Gmail
• Google Drive
• IBM DB2
• Jira
• Microsoft Exchange
• Microsoft OneDrive
• Micrsoft SharePoint (Cloud)
• Microsoft SharePoint (Server)
• Microsoft SQL Server
• Microsoft Teams
• Microsoft Yammer
• MySQL
• Oracle Database
• PostgreSQL
• Quip
• Salesforce Online
• ServiceNow Online
• Slack
• Zendesk

Yes! You can add custom connectors and then use Amazon Q SDK to action them.

• Navigate to Amazon Q in the console, and select Applications in the left menu.
• Choose Create application

• Provide an application name

• Create a new AWS Identity or Access Management (IAM) service role for Amazon Q, or select an existing one.

• Click Create.

• You’ll now need to select a retriever. This is an index that Q pulls data from in real time. You’ve got two options: the Amazon Q native retriever, or using an existing Amazon Kentra retriever. Native gives you the options of all the Amazon Q supported data sources listed above, while Kentra goes with your existing indexes. Select what works for you, and click Next.

• Connect your data sources, then click Finish. The data source sync will start, and may take a few minutes to a few hours.

Yes, you can connect plugins that allow it to manage access to ServiceNow, Jira, Salesforce, and Zendesk. This can be great for allowing Amazon Q to make those pesky tickets or analyze sales forecasts.

Conclusion: Amazon Q is the “next generation” of AI chatbots

… And yes, the Star Trek reference is completely intentional. With all that Amazon Q offers, it will be interesting to see how AWS’s competitors — namely, Microsoft and OpenAI — step up to the challenge. Will ChatGPT Enterprise incorporate more data points like AWS has with Q? Will we see a souped-up version of Azure Advisor? It’s an exciting time, and both businesses and developers are the real winners from this innovation race.