After introducing a number of generative AI-powered data management and analytics capabilities in preview and development last year, Google Cloud is now making some of those tools publicly available.
Additionally, the tech giant is announcing a new wave of generative AI capabilities.
In August 2023, Google Cloud introduced features such as Duet AI, a generative AI platform that has since been built into Gemini, and integration with Looker, the tech giant's leading analytics platform.
Additionally, the vendor unveiled integrations with both BigQuery, Google Cloud's fully managed data warehouse, and Gemini and Vertex AI, Google Cloud's machine learning platforms that include generative AI technologies.
Beyond Looker and BigQuery integration, Google Cloud announced support for AlloyDB AI, a database with vector search and storage capabilities, and open source platforms like Apache Hudi and Delta Lake in its data lakehouse BigLake .
Google Cloud on Thursday made AlloyDB AI, a data management and analytics feature that was in preview, generally available through an integration with Vertex AI, along with Gemini models for BigQuery.
Meanwhile, Google Cloud on Thursday announced new unstructured data analytics capabilities for BigQuery and support for vector searches across all databases, among other new features.
Progress towards GA
According to David Menninger, an analyst at ISG's Ventana Research, general availability of tools that have been in preview and development may seem relatively unimportant because it is a natural progression of product development. , says the general availability of generative AI capabilities is an important step. .
In the 15 months since OpenAI announced ChatGPT, which marks a significant improvement in generative AI technology, generative AI has expanded its product development efforts not only in Google Cloud's analytics and data management portfolio, but also in the portfolios of fellow tech giants AWS and Microsoft. has become a top priority issue.
Additionally, data platform vendors such as Databricks and Snowflake are prioritizing generative AI, as are more specialized analytics and data management vendors such as Alteryx, Informatica, Tableau, and ThoughtSpot.
Generative AI has the potential to not only expand the use of analytics within enterprises, but also improve the efficiency of employees already working with data.
Generative AI Large Language Models (LLMs) include an extensive vocabulary that enables true natural language processing (NLP) that is not possible with NLP tools due to their limited vocabularies. True NLP, on the other hand, virtually eliminates the need to write code to query and analyze data, broadening the potential audience for complex analytics platforms.
Additionally, LLMs can be trained to generate code and automate processes, increasing the productivity of existing data professionals by eliminating time-consuming and repetitive tasks.
But despite what the addition of generative AI capabilities to data management and analytics platforms means, most of the tools announced by data management and analytics vendors are at some stage of development.
Tableau just made its first generative AI tool publicly available on February 22nd, and a handful of other vendors have also made some features publicly available. But so far they are the exception.
So it makes sense for Google Cloud to make some of its AI-powered generative data management and analytics capabilities available now.
“We are in the race to bring generative AI capabilities into enterprise production environments,” Menninger said. “We're still early in that race, so it's important that we're making progress. But we can't declare any individual vendor a winner just yet.”
With AlloyDB AI now generally available, developers can use vectors to feed search augmentation generation (RAG) pipelines that continuously update and train AI, including generative AI models. You will be able to build AI applications.
BigQuery's Gemini model, on the other hand, brings generative AI to BigQuery, essentially allowing customers to connect BI to AI.
Following the general availability of both AlloyDB AI and generative AI capabilities in BigQuery, Google Cloud is rolling out even more data management features and capabilities in April, according to Gerrit Kazmaier, vice president and general manager of data analytics at Google Cloud. The company plans to make the analysis function generally available. “We're thrilled to be partnering with Google,” said Andi Gutmans, general manager and vice president of databases at Google Cloud.
Both executives spoke at a press conference on Tuesday.
Gutmans said the tech giant could have delayed the general availability of AlloyDB AI and BigQuery's generative AI integration until Google Cloud Next '24, a user conference scheduled for April 9-11. He said no. However, the feature is already ready and many other product developments will be announced during the conference.
“Customers have shown tremendous interest and excitement about many of these features and have been asking for them since yesterday,” Gutmans said. “We didn't want to artificially suppress features that were ready, so we decided to free them up. [before the conference]Please rest assured. There are many announcements planned for April. ”
next wave
In addition to making AlloyDB AI's integration with BigQuery's Gemini models generally available, Google Cloud has introduced a new wave of generative AI capabilities planned for its data management and analytics portfolio.
Prominent among these is its emphasis on vector retrieval and storage.
Business intelligence has traditionally been based on structured data such as financial records and POS transactions. However, it is estimated that only about 20% of the world's data is structured data, and 80% of the world's data is unstructured data such as text, audio files, videos, and photos.
Vectors are numerical representations of data that, when assigned to text, audio files, or other unstructured data, give structure to that data. Providing structure with vectors allows you to search, discover, and manipulate previously unstructured data.
In particular, vector search and storage is emerging as a way to feed and train generative AI models with enterprise-specific data, enabling them to inform business-specific decisions.
AlloyDB AI automatically generates vector embeddings using SQL, turning AlloyDB into a vector database.
Next, Google Cloud will add support for vector search in open source Redis databases and now in preview in Google Databases Cloud SQL, Spanner, Firestore, and Bigtable to its entire database suite.
“Our belief is that any database that stores operational data that might be used for generative AI use cases should have vector capabilities,” Gutmans said. “Good vectors should be a fundamental feature of a database.”
Similar to vector search capabilities for Google Cloud's databases, the tech giant announced new features for BigQuery aimed at enabling customers to operationalize unstructured data.
Google announced vector search in BigQuery in preview on February 14th. On Thursday, Google Cloud added integration with Vertex AI (also in preview), aimed at allowing customers to analyze text and audio for potential insights.
Until now, such data could not be used in any meaningful way, Kazmaier said. However, when used in combination with structured data, customers can develop new applications for data such as customer sentiment from audio files.
“This data is not typically used for corporate data analysis,” Kazmayer said. “We are entering an era of entirely new data and analytical capabilities.”
Menninger similarly noted Google Cloud's integration of Gemini, BigQuery, and Vertex AI to extract analytical value from unstructured data files.
“Gemini's multimodal capabilities that enable text, image, and video data analysis are new and at the forefront of what's happening with GenAI,” he said. “Many of a company's data processes are unstructured, so it's important to include this information in your analysis to get the full picture.”
In addition to its focus on making unstructured data more accessible, Google Cloud also offers RAG support in BigQuery and a developer framework for building generative AI models across its database suite. We announced support for LangChain.
next step
Gutmans and Kazmaier said Google Cloud will make more generative data management and analytics capabilities powered by AI generally available in the near future.
And as the tech giant introduces other new tools, one of its guiding principles will be to provide access to an ever-expanding swath of data signals, Kazmaier said.
He pointed out that big data is essentially a collection of large numbers of similar data. However, releasing unstructured data for analysis actually provides a broader range of data, thereby enhancing decision-making.
“We tend to think of GenAI primarily as a support function,” he said. “But if you think about data and analytics and the challenges that we have left, it's really about moving from so-called big data to so-called wide data.”
Meanwhile, Menninger said that with a proper focus on generative AI, it is important not only for Google Cloud but for all data management and analytics vendors to continue adding traditional AI capabilities.
He noted that traditional AI is an important aspect of applications such as fraud detection and predictive maintenance. Therefore, it should not be ignored.
“Investing in GenAI is sucking all the air out of the room,” Menninger said. “It’s important to recognize that traditional AI remains a critical element in certain use cases, which is why Google and other vendors are combining GenAI and traditional AI to meet all the requirements of today’s enterprises. I am interested in how to integrate the world of…
Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with over 25 years of experience. He is responsible for analysis and data management.