My First Impressions of Amazon Q

Every year in late November, Amazon Web Services (AWS) hosts its re:Invent conference. This conference introduces many new services and features. This year has been no exception. This year is heavily focused on Generative AI (GenAI). Yesterday, they announced a new service called Amazon Q, a generative AI assistant. After this announcement, when you log into the web console of your AWS account, you are greeted with a screen similar to the following:

I'm excited about this idea; I see its potential to help many users, myself included, get stuff done. Let's try it out. Hmm, what to try. How about?

"Please provide me a list of my S3 buckets."

That's not bad. It would be cool if it gave me my list of buckets or a link to that part of the console. It also could have suggested running the provided command in CloudShell. But still, not bad.

Next, let's try "Create a website." intentionally keeping this generic to see where it goes.

That's disappointing. It could have done a lot with that. Let's give it more to work with, "How can I create a website?"

I'm happy with that answer.

Let's see if it will write CloudFormation for us, "Please write me CloudFormation to create a CloudFront distribution."

That's disappointing. Well, this new service is in preview. Hopefully, this functionality will be there eventually. Last, let's try asking a question unrelated to AWS, "What are the latest headlines in world news?"

That answer makes sense to me.

I like this new service introduction. I see a lot of potential for it, and it's in preview, so it is understood that it will not be perfect. I look forward to Amazon continued development of this service. Some hopes I have to see in the future:

1. Help users find what they're looking for, for example links to specific places in the AWS web console.
2. Educate users and encourage best practices
3. Generate IaC (Infrastructure-as-Code)
4. Provide answers related to the current AWS account. For example, "Please list my S3 buckets." or "Please list all of my running EC2 instances."
5. Identify cost optimization opportunties across your AWS account.
6. Define settings for the type of solutions you create based on regulations, corporate policies, and/or preferences. For example, "We are an EKS shop." or "All external traffic must be routed through our corporate data center via our direct connect." Amazon Q should then make recommendations that adhere to those preferences.

What are your thoughts? Let me know in the comments below.

Thanks for reading.