Duet AI, a set of generative AI tools developed by Google for text summarising, data organization, and other uses, is being added to additional Google Cloud goods and services.
Google said at its annual Cloud Next conference that Duet AI, which is still in preview but will be generally available at some point this year, can now help with code refactoring, the process of enhancing code by making little changes without changing the system’s overall exterior behavior.
To have Duet AI perform on it (in this case, rewrite the function and convert the database connection to a managed relational database), a developer can open a Duet AI-powered chat window in their preferred software development environment and type a natural language prompt (for example, “Convert this function to Go and use Cloud SQL”) into it. Additionally, operators can talk with Duet AI in the Google Cloud Console, Google Cloud’s dashboard for developing and deploying web apps, websites, and services, to get advice on deployment, cost, and performance optimization as well as “how to” information regarding infrastructure configuration.
Google’s recently released development environment, Duet AI on Cloud Workstations, allows developers to write code while highlighting best practices. Duet may construct flows using existing APIs and assets while Application Integration, the no-code tool to connect software-as-a-service apps in Google Cloud, automatically generates documentation and test cases.
According to Google, some businesses will be able to modify Duet AI to produce context-aware code suggestions using “organization-specific” knowledge from their libraries and codebases. That will enable Duet AI, for instance, to generate code for a function that searches a company’s product catalog for all items priced under $10.
Through new interfaces to Apigee, Google’s API management platform, Duet AI is now able to assist with designing, creating, and publishing APIs from natural language prompts in other places. Additionally, it is better connected with Looker, a business intelligence tool for data exploration and discovery in Google Cloud, and BigQuery, Google’s completely managed serverless data warehouse.
Duet AI in BigQuery, according to Google, is a “collaborative” experience built into the UI of BigQuery that offers “contextual assistance” when crafting SQL queries and Python code. Based on existing metadata and schema, Duet AI in BigQuery can automatically suggest code in real-time, producing complete functions and code blocks while suggesting potential improvements and outlining the code. Additionally, vector embeddings—mathematical representations of data—can be produced using Duet AI to support semantic searches and recommendation inquiries.
Customers can modify the text-to-code model underlying Duet AI using Vertex AI, Google’s platform for developing, training, and deploying machine learning models in the cloud, to make the model’s recommendations align with their coding standards and practices.
New “context-rich insights” and report generation capabilities, as well as a chat feature dubbed Duet AI conversation support, are all powered by Duet AI in Looker. Chat help, which is accessible in Cloud Workstations, Spanner, and Apigee, allows customers to ask questions about their business data and receive responses in natural language, much like AI-powered chatbots like OpenAI’s ChatGPT.
Duet AI in Looker may initiate projects in LookML, the modeling language for expressing data relationships, and automatically generate presentations, summaries, computations, and visualizations based on saved reports via chat support and from other dashboards. Google says it will offer a Duet AI experience to analyze data in a shared notebook later this year.
In addition, Duet AI is coming to Google Cloud’s distributed database management and storage services, including AlloyDB (Google Cloud’s fully managed database service), Cloud SQL, and Cloud Spanner. Duet AI will assist in generating code from Cloud Spanner to structure, change, and query data using natural language. For example, Duet AI will automatically build the necessary code in response to a command like “Write a query to show all data in the messages table”.
Additionally, Duet AI will soon be incorporated into Google’s Database Migration Service (DMS) to speed up the transfer of data from a third party to Google Cloud. Duet AI in DMS will manage the conversion of specific database code, including stored procedures, functions, triggers, packages, and custom query language code, to AlloyDB and Cloud SQL later this year for Oracle clients.
