Google Cloud on Tuesday unveiled plans to inject Gemini into its analytics and data management platforms with the goal of providing customers with a foundation for developing AI models and applications that aid analysis.
Gemini in BigQuery, Gemini in Databases and Gemini in Looker are now in public preview. All were revealed during Google Cloud Next ’24, the tech giant’s user conference in Las Vegas.
First introduced in December 2023, Gemini is a large language model (LLM) from Google Cloud that enables users to understand and generate text, images, audio, videos, code and other types of information.
When integrated with data management and analytics tools, Gemini and other LLMs such as ChatGPT from OpenAI and Claude from Anthropic can help nontechnical workers interact with data using natural language processing (NLP). They can also make data experts more efficient by reducing time-consuming tasks.
Gemini for Google Cloud includes the integrations with the tech giant’s data management and analytics tools. It aims to provide a unified environment in which customers can develop models and applications using their own proprietary data so that end users can make business-specific decisions aided by AI.
Unification, meanwhile, is key, according to Doug Henschen, an analyst at Constellation Research.
After the launch of ChatGPT in November 2022, many vendors integrated their products with LLMs to add generative AI (GenAI) capabilities. All but some of those integrations, however, were additions. The generative AI capabilities weren’t injected into existing tools in a unified way.
It’s not just about adding GenAI features and capabilities. Google’s data and AI strategy is about delivering a comprehensive platform with well-integrated capabilities so you can do it all without having to move data around or cobble together disparate services. Doug HenschenAnalyst, Constellation Research
“It’s not just about adding GenAI features and capabilities,” Henschen said. “Google’s data and AI strategy is about delivering a comprehensive platform with well-integrated capabilities so you can do it all without having to move data around or cobble together disparate services.”
Google Cloud is not alone in attempting to create a connected environment for AI, he continued. For example, Databricks has aggressively acquired and developed capabilities that enable users to build AI models and applications. Similarly, AWS and Microsoft have developed tools designed to foster AI and analytics in concert with one another.
“The big push across all the hyperscale clouds is to offer a single platform for data that seamlessly supports all your AI, GenAI, analytics and wider application development and operational needs,” Henschen said.
Google Cloud previously introduced integrations between Duet AI, a generative AI platform, and its data management and analytics tools in August 2023. However, the tech giant subsequently folded Duet AI into Gemini in February. The new fusion of Gemini and Google Cloud’s data management and analytics tools therefore represents an evolution of the integrations with Duet AI.
Gemini, data management and analytics
Gemini for BigQuery, Gemini for Looker and Gemini for Databases all aim to help data workers be more productive by enabling them to use AI in concert with their own proprietary data, according to Thomas Kurian, Google Cloud’s CEO.
BigQuery is the tech giant’s fully managed data warehouse, Looker is its primary analytics platform, and Google Cloud offers a series of database options including AlloyDB and Cloud Spanner.
“Our goal is to provide organizations with a digital platform — AI powered — to help them accelerate their digital transformation for their business and for their industry,” Kurian said on April 4 during a virtual press conference. “Our data cloud allows people to manage their data, understand insights from their data, and use AI along with data to analyze, predict and summarize information.”
Using Gemini in BigQuery, customers can engage with their data in a notebook-like environment using natural language. Without writing code, users can integrate data and develop pipelines that inform models and applications.
In addition, Gemini in BigQuery provides users with embedded visualizations, AI-augmented data preparation capabilities designed to help cleanse and discover the most relevant data for an application, query recommendations, and the ability to translate text to SQL or Python code.
A connection between BigQuery and Vertex AI also comes with Gemini in BigQuery. Vertex AI is Google Cloud’s machine learning and generative AI platform that features access to a host of proprietary and open source AI models to provide users choice when deciding which model to use with their data.
Gemini in Looker enables customers to essentially chat with their business data, providing an intelligent assistant for self-service and collaborative data analysis.
Conversational Analytics — which is in private preview — is a tool that will enable users to ask questions using natural language. In addition, features in public preview include capabilities that enable users to create visualizations using natural language, connect those visualizations with Workspace and share them to co-workers also without having to write code.
Finally, Gemini in Databases extends NLP capabilities to Google Cloud’s cadre of databases, enabling developers and administrators to build applications and govern data without writing code.
The integration includes SQL generation and summarization in Database Studio, management capabilities in Database Center that enable administrators to oversee all their databases in a single pane, and smart recommendations that provide users with database management advice.
Separately, Gemini in BigQuery, Looker and Databases provide capabilities that will benefit Google Cloud’s data management and analytics users, according to Henschen. The significance of each, however, lies not in the individual components, but in their comprehensiveness.
“It’s significant both in terms of the depth and breadth of GenAI capabilities promised both within each product and across the entire portfolio of services,” Henschen said.
For example, Gemini’s presence and assistance in BigQuery from data ingestion through preparation, cleansing and pipeline development before providing query recommendations provides range, he continued.
Similarly, while numerous vendors are adding generative AI capabilities to their databases, Google Cloud is doing it comprehensively across its database portfolio and with sophisticated features.
“The breadth of Gemini assistance is a differentiator,” Henschen said.
In addition to the integrations between Gemini and Google Cloud’s data management and analytics tools, the tech giant unveiled the following capabilities aimed at helping customers derive insights from AI and data analysis:
Improved support for vector search and storage in BigQuery and AlloyDB to help customers discover relevant data to feed the retrieval-augmented generation pipelines that are used to train AI models and applications.
Connections between Vertex AI and both BigQuery and AlloyDB to provide access to AI models in Vertex AI.
Natural language search capabilities in Google Distributed Cloud (GDC) built using Gemma, an open model from Google DeepMind, that also enable customers to use third-party models.
Workload performance improvements in GDC fueled by GPU and Tensor Processing Unit optimization capabilities.
Integrations with vendors including Denodo and MongoDB that expand Google Cloud’s data management and analytics ecosystem.
Just as Gemini for BigQuery, Looker and Databases are designed to foster AI development and analysis, so are the many other new features unveiled on Tuesday, according to Kurian.
“All of this is allowing our customers to build their own AI-powered applications,” he said.
Seven benefits of generative AI for the enterprise.
Next steps
Although generative AI has been the primary focus for many data management and analytics vendors for more than a year, few tools integrating generative AI with data and analytics are generally available.
There are some exceptions.
For example, data observability specialist Monte Carlo and data lakehouse specialist Dremio each released generative AI tools in June 2023. More recently, Tableau launched Pulse, a tool that generates insights and delivers them to users, in February, and MicroStrategy, after initially releasing generally available capabilities in October 2023, launched an embeddable AI chat tool in March.
Even Google Cloud has made some generative AI capabilities generally available, including AlloyDB AI.
The majority, however, remain in some stage of preview, such as those unveiled by Google Cloud on Tuesday. As a result, even if one vendor’s generative AI product development is slightly ahead or behind competitors, it is far too soon for any to be declared significantly more or less innovative than another.
Henschen noted that Google Cloud was slightly behind AWS in adding generative AI to its transactional databases. BigQuery and Vertex AI, meanwhile, compare favorably to platforms being built by others, as does Google Cloud’s emphasis on openness in terms of enabling customers to use AI models developed by third parties.
Therefore, regarding generative AI with data management and analytics, Google Cloud would be wise to continue building on what it has done already, according to Henschen.
“I’d say Google just needs to stick to its strategy, which has been very consistent,” he said. “Google leads with its strengths. Openness to third-party vendors and model providers has been another consistent and important part of the Google Cloud strategy.”
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.
