Amazon Web Services (AWS) is poised for a massive boost in enterprise productivity this week with the full launch of Amazon Q, a new generative AI service.
Doug Seven, the company's general manager and director of AI development experience, said that enterprises will have the ability to leverage Amazon Q to deploy semi-autonomous AI agents that can perform a variety of tasks, which will be a boon for software developers and office workers. For those customers who leverage Q in many different use cases, they can see more than 80% increase in productivity in different roles in different jobs, whether it's in software development or other types of business areas.
Last November, AWS released a preview of Amazon Q at its re:Invent conference, and several customers such as Blackberry, BT Group and Toyota have tested it. Wednesday's general announcement made some small changes, including the launch of two major versions of the product: Amazon Q Developer, which focuses on working with computer code, and Amazon Q Business, which focuses on working with documents and data.
Amazon Q Developer leverages large language models (LLMs) and other generative AI techniques to create AI models that understand computer code. According to AWS, Q Developer can be used to generate, understand, troubleshoot, refactor, and debug code in a variety of languages, including SQL and Java. The software is trained to detect security vulnerabilities and apply fixes, and it is also trained on the AWS environment knowledge base.
As part of this week's launch, AWS is also announcing a new feature called Amazon Q Developer Agents. According to Seven, these agents are not just typical digital assistants, and software developers can dispatch these agents to perform specific tasks on a semi-autonomous basis, allowing human developers to focus on other things.
"Once you and Q agree on a work plan, Q will run independently and create its own development environment. It will branch the code. It will compile the code. It will work on the code, it will make changes, it will do all the things it needs to do. Seven said. And then it will come back to you at some point and say, okay, I'm done, here's the code change I'm suggesting. As a developer, you can review them as you would your peers' code, and accept it if that's what you want. Then move on to the next thing. ”
Amazon Q also allows developers to ask questions about their codebase using natural language. For example, a developer can ask an Amazon Q developer to "explain this code to me" or "show me how it works at a high level", and Q will do just that. It can be used to automatically fix code bugs and even upgrade Java code (the next thing to consider is .Net). Users can ask Q Developer about their AWS environment, including how much money they spent on EC2 instances in the last month. It can even optimize SQL queries and ETL pipelines.
"I'm really excited about the idea of agents in AI and this autonomous or semi-autonomous capability," Seven said. "We still believe that there must be human involvement. It will do the work and present it as a variation. You can review it just like any other developer. So you have complete control over it, but it's really exciting. ”
Customers have reported good results. For example, he said, BT Group, the British telecommunications company, generated about 100,000 lines of code in its first four months of using Q, and their pass rates were slightly higher than the industry average of around 37 percent. Another early tester, National Australia Bank, had a code acceptance rate of 50%. "We've seen great results from customers who have used it," Seven said. ”
According to AWS, Q has achieved good results in both the SWE-Bench and SWE-Bench (Lite) rankings, which are the industry benchmarks for GenAI models. According to Seven, Amazon Q has the highest score among all GenAI products.
Amazon Q Developer is now available. Customers can access it through the AWS console, Slack, or IDEs such as Visual Studio Code and JetBrains. The Pro version of Q Developer costs $19 per developer per month.
At the same time, Amazon Q Business has been trained as a GenAI assistant that can understand users' business documents and data. The product has more than 40 pre-built data connectors to pull data from S3, Gmail, Salesforce, ServiceNow, Slack, Sharepoint, Box, OneDrive, and other systems. It also has built-in analytics with the ability to build reports and dashboards based on the data it finds.
"One of the things that's really remarkable is that digital assistants like this are getting more and more capable. Seven said. "They're able to absorb a lot of information and then summarize it. So being able to ask questions like connecting Q instances to Salesforce data and asking 'what are the top five customer opportunities that I currently have available'. Or connect it with a ticketing system so you can ask: 'What has been the overall customer sentiment over the last 30 days?'"
It also connects with AWS's flagship business intelligence and analytics product, Amazon QuickSight. Seven says QuickSight is powerful, but sometimes it's hard for a layman to make it work exactly the way they want it to. With Amazon Q Business as the natural language layer, novice users can now tell QuickSight (via Q) the type of data they want to see on the dashboard.
"So having Q in QuickSight means I can express what I want in natural language. You can have it "create visualizations of customer adoption over the last 60 days by region," and it can create those visualizations, Seven says. Or "Give me a dashboard where I can see sales by team or salesperson." "It's really amazing because it makes it so easy to create these visualizations.
One early trialist even set up a Q Business endpoint in Slack, allowing employees to get answers to their questions directly in the Slack channel. With the appropriate name and @ symbol, Seven says, it calls Q Business, infers on the data, and generates a response.
QWhat respects data access policies. So, if a user doesn't have permission to access a particular piece of content, they won't be able to access that content.
AWS has also launched Q Apps, which allows users to convert their Q Business queries into reusable, compressed-wrapped applications that can be downloaded from the gallery. Seven said the new service, inspired by AWS's PartyRock experiment, will amplify the potential impact of Q Business.
One potential use for Q Apps is to create an onboarding plan for new hires based on who they work for and in which department, Seven said. Salespeople may also use Q Apps to create sales scripts.
"I think it's really cool. "We've been experimenting for a long time, making tools for non-developers, building applications for business experts, building applications for lines of business." I've developed some tools for this myself. But it's always been a struggle. It's one of the most significant advancements I've ever seen, and one of the best generative AI in supporting non-developer self-service. ”
Both Q Developer and Q Business are based on a variety of underlying technologies, including AWS products such as Trainium and Inferentia chips, official and third-party LLMs on Amazon Bedrock, and the Sagemaker AI development environment, Seven said. He said: "There is a little bit of everything. "We don't have bias of one kind or another. ”
The Pro version of Amazon Q Business costs $20 per user per month.