VENDORiQ: Google Infuses its Data and Analytics Platforms with AI

Why It’s Important

BigQuery consolidates essential Google Cloud analytics features into a single platform, streamlining the management of data workloads for organisations. This streamlined approach encompasses structured data in BigQuery tables, unstructured data such as images and documents, and streaming workloads, all within one unified product experience.

At Google Next, it was announced that BigQuery now supports searching through huge amounts of data, as large as 10 petabytes per project per day. The data stream connectors enable data to move, in real-time, from databases like Oracle and MySQL right into BigQuery and Cloud Storage.

With the direct linkage of BigQuery and Vertex AI, Google is introducing the capability to associate Vertex AI models with your enterprise data, eliminating the need to duplicate or transfer your data from BigQuery. This facilitates multi-modal analytics utilising unstructured data, precise adjustments of LLMs, and the application of vector embeddings in BigQuery.

The direct linkage between BigQuery and Vertex AI now facilitates the smooth preparation and AI-based analysis of multimodal data, including documents, audio, and video files, as well as simplifying the inclusion of such data in subsequent AI-based processes.

While BigQuery provides a serverless model for managing workloads, IBRS notes that is it crucial for organisations to effectively oversee their computing resources to keep expenses in check. Fine-tuning Cloud usage for optimal cost-performance can be intricate, and the cost of running artificial intelligence (AI) processes can quickly add-up if not well managed.

The integration of Gemini in BigQuery and its contextual understanding of an enterprise’s business gained through access to metadata, usage data, and semantics is a powerful new feature that is yet to be fully appreciated. Gemini in BigQuery extends beyond chat assistance, introducing new visual experiences such as data canvas — a novel, natural language-based interface for data exploration, curation, wrangling, analysis, and visualisation workflows.

Who’s Impacted

• Data teams
• Cloud architects
• AI project leads
• Innovation leads

What’s Next

Organisations must evaluate their existing infrastructure and data systems to determine how the BigQuery innovations align and integrate with current capabilities and needs. Before full-scale adoption, they should run proof-of-concept projects to test the new BigQuery features in real-world scenarios.

Additional Reading

• Modernising Analytics In The Cloud – Governance
• Modernising Data Warehouse Solutions in the Cloud – Data Discovery & Cataloguing
• IBRS Framework For End-To-End Data Analytics Platform Architecture Review