Google on Tuesday unveiled a series of integrations between Duet AI and its data management and analytics platforms, including infusing the generative AI assistant into BigQuery and Looker.
In addition -- among a spate of other new data management, analytics and developer-oriented tools -- the tech giant introduced a database migration service; an interface for engineering and analytics within BigQuery Studio; and new features for AlloyDB, a fully managed SQL database first launched in May 2022.
The new capabilities -- most of which are in preview -- were revealed during Google Cloud Next, a user conference held in-person in San Francisco with a virtual component.
Beyond adding functionality to the Google Cloud Platform, with the new features Google is positioning itself as a generative AI innovator compared to AWS and Microsoft, its main competitors, according to Doug Henschen, an analyst at Constellation Research.
He noted that Microsoft partnered with OpenAI to develop generative AI capabilities while Google has built its own generative AI tools as the two tech titans strive to outdo each other.
"Google initially had to react to Microsoft's announcements earlier this year," Henschen said. "But the company had a deep well of AI assets and expertise to draw on. And I still see it as the leader among all three clouds in the depth and breadth of its AI capabilities, now including generative AI."
Before Tuesday, Google most recently made significant updates to its data management and analytics portfolio in March 2023.
At the time, the tech giant unveiled the preview of Looker Modeler, a standalone service that lets customers of any analytics vendor's platform use Looker's semantic modeling capabilities to define and store metrics. It also revealed a tool within BigQuery called data clean rooms that enables customers to combine their own data with external datasets to better contextualize their operations.
In addition, Google in March launched new pricing options for BigQuery.
Duet AI for data, analytics
Duet AI, first revealed in preview last spring, is Google's version of generative AI for workspace applications and is now generally available.
Microsoft has similar capabilities, which it calls Copilots.
Duet AI in Google Cloud was first introduced at the Google I/O conference in May, with features for developers such as code help and chat assistance.
Doug HenschenAnalyst, Constellation Research
Now Google is expanding the breadth of Duet AI within Google Cloud to include Duet AI in BigQuery, Duet AI in Looker and Duet AI in Database Migration Service. All are now in preview, and all are aimed at easing burdens on data workers with generative AI, according to Gerrit Kazmaier, vice president and general manager of data and analytics at Google Cloud.
"They're all about bringing our generative AI to all the data workers out there," he said during a press conference before Google Cloud Next. "It's about understanding that every AI project is a data project in disguise. You [normally] have to bring the data to the enterprise AI system. We're changing that [so users] can unlock AI with their business data."
Looker is Google's primary analytics platform. The tech giant acquired the previously independent vendor for $2.6 billion in June 2019 and, in October 2022, consolidated its previously disparate BI tools -- including Data Studio and Connected Sheets -- under the Looker name.
Duet AI in Looker is a natural language processing capability that enables customers to query their data using freeform language rather than write code and generate code with LookML with an understanding of users' intent.
In addition, the feature can create presentations on its own -- including dashboards and reports -- with natural language summaries so that users don't have to go through the time-consuming process of creating their own summaries and presentations.
BigQuery, meanwhile, is Google's fully managed data warehouse. Duet AI in BigQuery is integrated directly into the BigQuery interface and is designed to help users write SQL queries and Python Code to make users more efficient as well as enable them to spend less time on querying and coding and more on analysis.
The tool uses generative AI to provide suggested code, generate code blocks on its own and recommend fixes.
Given that both Duet AI in BigQuery and Duet AI in Looker demonstrably reduce the need to write code and enable the use of freeform natural language to engage with data, the tools -- and those like them in the works from other vendors -- are significant additions, according to Henschen.
Beyond improving efficiency, the tools also have the potential to make analytics accessible to a broad audience of self-service users. Analytics use within organizations has been stuck around one quarter of all employees for most of the 21st century, and generative AI tools could be the means of breaking though that barrier.
"Duet AI for both BigQuery and Looker is a potential game changer as it promises to make things easier for analysts and business users alike," Henschen said. "There's a palpable sense that the gen AI capabilities promised by Google, Microsoft and others may finally make analytics and BI broadly accessible and understandable to business users."
Kevin Petrie, an analyst at Eckerson Group, also highlighted the importance of generative AI assistants such as Duet AI.
He noted that numerous other data management and analytics vendors are developing such tools but that Google has an advantage over those developing search and AI tools with its history and experience developing large language models, such as Bard and PaLM.
"Generative AI assistants are becoming must-have features of cloud data platforms and BI tools," Petrie said. "Google is well-positioned to offer such capabilities given its heritage in search and its expertise with large language models. Because this is Google and not a startup, users have justifiably high expectations."
More new capabilities
Beyond integrations between Duet AI and Looker and BigQuery, some of the most significant data-related features the tech giant unveiled at Google Cloud Next promote openness and unification, according to Henschen.
Toward that end, Google unveiled improved support for open source formats such as Apache Hudi and Delta Lake within BigLake, which is the tech giant's storage engine for unifying data warehouses and lakes.
In addition, Google added performance acceleration for Apache Iceberg to better enable users to ingest large amounts of data as well as a tool called AlloyDB Omni, now in preview, that lets users run AlloyDB on any cloud or on premises.
Meanwhile, to better enable unification within Google's own ecosystem, the tech giant introduced a new BigQuery Studio interface in preview with which users have a single interface for data engineering, machine learning and analytics.
"Openness and generative AI advances are the two big themes on the analytics front," Henschen said.
Other data management and analytics capabilities Google unveiled at Google Cloud Next include the following:
- An integration between BigQuery and Vertex AI -- Google's machine learning platform, including generative AI capabilities -- designed to enable users to easily build their own language models and generative AI applications with customer data.
- AlloyDB AI, which includes features such as vector search and vector embedding that help fuel generative AI applications as well as an integration with Vertex AI to help customers develop those generative AI applications using SQL.
- Cloud Spanner Data Boost to improve the processing of operational data.
- New developer tools designed to enable users to more quickly get started building generative AI tools.
- An integration between Duet AI and Google's Data Migration Service that enables users to easily convert code so they can migrate data from Oracle to AlloyDB.
Cloud Spanner Data Boost is generally available, while each of the others are in preview.
In sum, the new capabilities represent Google's attempt to enable customers to develop generative AI applications irrespective of which Google tools they use and which clouds they deploy on besides Google, according to Andi Gutmans, vice president and general manager of Google databases.
"We're giving customers the ability to unlock generative AI innovation no matter where they are," he said.
Petrie, meanwhile, noted that many of the new features go beyond just enabling users to develop generative AI applications in an open, unified way.
They also represent Google's recognition that there's an ongoing convergence between generative AI and data management and analytics.
"It's interesting to see how Google is tackling the convergence of generative AI and traditional analytics," Petrie said. "For example, [Google is] connecting BigQuery tables with language models in Vertex AI and using BigQuery Studio to build prompts. By applying generative and BI to both text and tabular data in an integrated way, Google delivers more detailed and accurate outputs to users."
With nearly all the new capabilities introduced at Google Cloud Next in preview as well as many of the generative AI features unveiled by the tech giant at previous events this year, Henschen said he's curious to finally see what the tools look like beyond the demonstration stage.
Rather find out what more Google has planned for Looker, BigQuery and the rest of its data management and analytics capabilities, he said he'd like the tech giant make the many generative AI features now in preview generally available.
With few exceptions, the generative AI tools introduced by Google, Microsoft, AWS and a host of independent vendors are in preview.
Some vendors, like Dremio and Monte Carlo, have released text-to-code translation capabilities into full production. But most of the more ambitious generative tools -- like most of those being unveiled at Google Cloud Next -- are in preview.
Only once the tools are generally available will there be a true sense of generative AI's potential.
"I'm eager to see the demos and previews move into early trials and general availability," Henschen said. "Early adopters will find out what works and what doesn't and make some unexpected discoveries about generative AI that vendors don't foresee. Generative AI has the potential to change analytics and BI as we know it, but it's time for reality to catch up with the promises."
Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with more than 25 years of experience. He covers analytics and data management.