Jump to Content

BigQuery and AlloyDB hit major milestone with AI-enabled updates

February 29, 2024
Brad Calder

Vice President and GM, Google Cloud

Try Gemini 1.5 models

Google's most advanced multimodal models in Vertex AI

Try it

While generative AI is a revolutionary technology that’s disrupting industries and transforming business operations, its success depends on data. Data is crucial for AI, as it is the foundation upon which machine learning algorithms learn, are grounded, make predictions, and improve performance over time. As a result, many CIOs are asking the question: are we bringing our data together and applying the latest AI and best data tools to innovate? A key challenge for customers harnessing generative AI is the ability to access, manage, and activate all types of data, regardless of format.

The rise of generative AI is widely welcomed by data teams. In our recent study, Google Data and AI Trends Report, 2024, 84% of data decision makers believe generative AI will help their organization access insights faster. But there are challenges too, with over 80% of respondents agreeing that the lines between data and AI roles are starting to blur, where many data analysts are now seeking to leverage capabilities that were traditionally reserved for data scientists and vice versa. For example, data analysts need ML for improved insights and data scientists require broader sets of enterprise data to train their models.

84% of data decision makers believe generative AI will help their organization access insights faster.

The opportunity for data to impact enterprise AI strategies extends to developers as well. When developers can access and leverage their organization’s data directly with their foundation models, they can deliver innovation much more quickly.

As we look to help organizations meet their evolving business needs, we're focused on empowering data teams to unify their data and combine it with groundbreaking AI for transformative experiences. To that end, today we are announcing a major milestone with new AI-enabled product innovations for both BigQuery and AlloyDB for PostgreSQL.

Access powerful generative AI models in BigQuery

Today, we’re making Gemini 1.0 Pro accessible for BigQuery customers via Vertex AI. With these new integrations to Vertex AI, data engineers and data analysts can now use Gemini models for multimodal and advanced reasoning capabilities for their BigQuery data. This can help healthcare providers improve patient care, make supply chains more efficient, and increase customer engagement in telco, retail and financial services.

Today, we are also announcing BigQuery integration to Vertex AI for text and speech. Available in preview, these new capabilities help companies extract insights from unstructured data like documents and audio files to unlock new analytics scenarios that combine unstructured data with structured business data. For example, a data analyst can derive insights from call-center audio recordings to help improve future experiences. Additionally, the recently announced BigQuery vector search allows for vector similarity search and recommendation queries on BigQuery data. This functionality, also commonly referred to as approximate nearest-neighbor search, is key to empowering numerous new data and AI use cases such as semantic search, similarity detection, and retrieval-augmented generation (RAG) with large language models (LLMs). For example, vector search can help retailers improve product recommendations, summarize fixes for common customer care support tickets, or even help discover trends across large sets of documents.

These new generative capabilities not only empower data teams — they also break new ground from traditional approaches in data warehousing by accessing data and delivering insights in fresh new ways. But what about operational databases?

AlloyDB: a modern database built for the generative AI era

Operational databases, with their wealth of application data, play a critical role in how developers build new, AI-assisted user experiences. Seventy-one percent of respondents in the Data and AI Trends Report plan to use databases integrated with gen AI capabilities. That’s why we continue to invest in key AI and ML enabling technologies for all our database offerings, including adding support for in-database embeddings generation, vector search, and support for tools and frameworks that help developers build gen AI apps faster.

At Next ‘23, we announced AlloyDB AI, an integrated set of capabilities in AlloyDB for PostgreSQL for building enterprise gen AI applications. Today, we are pleased to announce that AlloyDB AI is generally available. We’re also pleased to announce the public preview of vector search capabilities across more of our databases including Spanner, MySQL, and Redis, and we are adding integrations with LangChain, a popular framework for developing applications powered by language models.

All these database capabilities join our existing integrations with Vertex AI to provide an integrated platform for developers. Spanner and AlloyDB integrate natively with Vertex AI for model serving and inferencing with the familiarity of SQL, while Firestore and Bigtable integrate with Vertex AI Vector Search to deliver semantic search capabilities for gen AI apps.

71% of respondents in the Data and AI Trends Report plan to use databases integrated with gen AI capabilities.

Data is the fuel for AI, and what powers its effectiveness. We’ve long shared robust privacy commitments that outline how we protect user data and prioritize privacy, and generative AI doesn’t change these commitments — it reaffirms their importance. To truly take advantage of generative AI, you need the ability to access, manage, and activate all your data across your analytical and operational systems. And it all starts with a unified data cloud with built-in AI technologies.

To learn more about these new innovations in data and AI, join our live Data Cloud Innovation Live webcast on March 7, 2024, 9:00 AM - 10:00 AM PST.

Posted in