Reimagine data analytics for the era of AI

The emergence of generative AI is poised to become one of the most significant technological shifts in modern memory, opening up endless transformative possibilities for enterprises. Our customers are already seeing incredible benefits with AI. Organizations like TIME are exploring new possibilities with gen AI to engage with customers and build a more robust community. Wendy’s is innovating fast-food order management, Orange is exploring next-generation contact centers, and Priceline uses BigQuery’s AI capabilities with its own proprietary algorithms to offer customers personalized services and product recommendations. The new functionality we are announcing today unlocks even more ways for customers to innovate.

Data is at the center of gen AI, which is why we are bringing new innovations to our Data and AI Cloud to help companies activate their data with AI. 

• First, we are helping to interconnect your data and workloads by announcing BigQuery Studio, a single interface for data engineering, analytics, and predictive analysis to simplify end-to-end data workflows. Additional features provide data teams with a simplified data foundation and include enhanced support for unstructured data, cross-cloud analytics, secure data sharing, and governance. 
• Second, we are bringing AI to your data in BigQuery with integration to Vertex AI foundation models. New innovations for real-time model inference and vector embeddings help allow you to securely run generative AI at scale on your business data.
• Lastly, we are boosting productivity of your data teams with a preview of Duet AI in Google Cloud to reimagine data work with products such as Looker, BigQuery, and Dataplex. 

These innovations will help organizations harness the potential of data and AI to realize business value — from personalizing customer experiences, improving supply chain efficiency, and helping reduce operating costs, to helping drive incremental revenue. 

Data teams work with different tools for managing data warehouses, data lakes, governance, and machine learning, which can slow productivity. To interconnect the ways teams work with data we are announcing BigQuery Studio. Now available in preview, BigQuery Studio provides customers with a single interface for data analytics in Google Cloud. Now you can bring your data engineering, analytics, and predictive analysis together, simplifying how data teams work across end-to-end workflows without having to switch between tools. 

"Shopify has invested in employing a team with a diverse array of skill sets to remain ahead of trends for data science and engineering. In early testing with BigQuery Studio, we liked Google's ability to connect different tools for different users within a simplified experience. We see this as an opportunity to reduce friction across our team without sacrificing the scale we expect from BigQuery." - Zac Roberts, Data Engineering Manager, Shopify

BigQuery Studio also allows data teams to edit SQL, Python, Spark and other languages to easily run analytics at petabyte scale without additional infrastructure management overhead. Notebooks are the preferred environment for writing and editing Python, so we’ve integrated BigQuery Studio with Colab Enterprise, a new offering that brings Google Cloud enterprise security and compliance support to the popular Colab data science notebook developed by Google Research.

To provide flexibility of choice in your data science notebook, we are extending partnerships with Hex, Deepnote, and Jupyter. In addition, DataFrame in BigQuery provides the data structure for teams to access their favorite notebook with large datasets that would typically surpass memory limits. 

Disconnected and unstructured data present additional challenges for data teams. Large amounts of valuable data is contained in videos, documents, log files, and audio recordings which can be used with generative AI. Today, we are adding new capabilities to unify your structured business data with unstructured data, as well as help to provide secure access, without the need to move it. These innovations include:  

• Enhanced support for open source formats like Hudi and Delta Lake within BigLake, which unifies data lakes and warehouses to breakdown data silos. Customer use of BigLake to combine data lake and warehouse workloads across clouds has grown 27x to hundreds of petabytes in the past 6 months. Innovations also include performance acceleration for Apache Iceberg which provides continuous data optimization for large-scale ingestion. 
• Enhancements to analyze and train your data without moving it are available with cross-cloud materialized views and cross-cloud joins in BigQuery Omni. Now companies can bring together data across multiple clouds in a single lakehouse. In addition, Spark integration on Google Distributed Cloud extends the power of fast analytical query processings to on-premises to support data residency requirements. 
• New governance capabilities in Dataplex for data lineage, quality, and metadata management help users understand what data to analyze, and train ML models on trusted data sources to help improve accuracy. 
• New privacy-centric connections, including BigQuery data clean rooms and Ads Data Hub for Marketers, which can help you understand your Google and YouTube campaign performance.

 

The importance of AI and data continues to be a major focus for our customers. Many data teams are using their analytical data warehouses and lakes to build ML models using BigQuery ML as their starting point. In fact, customer use of BigQuery ML in the past two years has seen over 250% query growth. This year, customers have run hundreds of millions of prediction and training queries in BigQuery ML. 

To get improved insights from your data with generative AI, we are announcing access for Vertex AI foundation models, including PaLM 2, directly from BigQuery. This can remove complexity and allow data teams to scale simple SQL statements in secure ways against large language models, opening up endless possibilities for insights. 

Using new model inference in BigQuery, customers can run model inferences across formats like TensorFlow, ONNX, and XGBoost. In addition, new capabilities for real-time inference can identify patterns and automatically generate alerts. 

Faraday, a leading customer prediction platform, previously had to build data pipelines and join multiple datasets. Now, not only can they simplify sentiment analysis but they can also take the customer sentiment, join it with additional customer first-party data, and feed it back into the LLMs to generate hyper personalized content — all within BigQuery.

For model tuning, we are adding vector and semantic search in BigQuery. Vector and text embeddings allow for efficient search and retrieval of unstructured data, such as text or images. This capability powers gen AI applications to more efficiently retrieve unstructured data and provide context to LLMs. In addition, customers can automatically synchronize vector embeddings in BigQuery with Vertex AI Feature Store for model grounding. 

Access to trusted data is critical to building and training new AI models — particularly specialized models for industries like financial services, retail, and manufacturing. We offer data sets in BigQuery from leading data providers including CoreLogic, Dun & Bradstreet, and TransUnion. Now, customers can use thousands of datasets from hundreds of providers including Acxiom, Bloomberg, Equifax, Nielsen, and Zoominfo. The availability of these datasets helps Google Cloud to be the best place for enterprises to build and train new AI models.

To help data teams of all skill levels solve their everyday work challenges and boost productivity, we’re announcing our always-on generative AI-powered collaborator, Duet AI, is in preview across a variety of products in our portfolio such as Looker, BigQuery, and Dataplex. Powered by Google's state-of-the-art foundation models, these innovations can help data teams clean data, prepare it for analysis, answer questions, and predict trends.

To provide insights to non-technical users via natural language, we are announcing Duet AI in Looker, which enables fast and simple conversational queries that empower you to get answers and refine results into visuals and reports. In addition, Duet AI provides automatic presentation creation with intelligent summaries, formula and visual assist to quickly create calculations, and the ability to rapidly create code using LookML with an understanding of intent. 

By bringing generative AI and natural language search together, our vision for Duet AI in Looker is to allow you to “talk” with your business data, much in the same way you “ask Google” a question. It’s like having a brilliant data analyst available for every employee.

Duet AI in BigQuery is a collaborative experience integrated directly into the BigQuery interface. It provides contextual assistance for writing SQL queries and Python code, which can allow data teams to focus more on analyses and outcomes. It can auto-suggest code in real-time and generate full functions and code blocks, as well as recommend fixes. And, for improved access to trusted data, Duet AI in Dataplex provides metadata search using natural language for a view of your ML assets and datasets. Sign up now to try Duet AI.

“Duet AI in BigQuery provides contextual awareness and extends our investment in Google Cloud's integrated data platform. We see this as an architectural advantage, eliminating the need to train, host, and manage custom models.” - VP of Data Engineering, Aritzia

Google has changed the way the world accesses information. Now, with Google’s Data and AI Cloud, we can bring new levels of simplicity, scale, security, and intelligence to your business data. To learn more about product innovations, hear customer stories, and gain hands-on knowledge from our developer experts, watch our data analytics spotlights and breakout sessions at Google Cloud Next on-demand.