Google plans to create a ‘family’ of AI agents to optimize data management

In the digital age of 2025, where data is the new gold, the need to manage it efficiently is acute. Google Cloud is demonstrating a bold vision to revolutionize data management. At its Next 25 conference, the company unveiled its ambition to create a “family” of artificial intelligence (AI) agents designed to optimize various business processes related to data management. These advances promise to break down current silos and intelligently integrate AI into applications like BigQuery, while leveraging its powerful Gemini model. But what does such a transformation really mean for companies seeking efficiency and innovation?

Google’s ambitions with Gemini: making generative AI mainstream in BigQuery

Google Cloud is looking to the future with the infusion of generative AI into its BigQuery offering. With thousands of users already adopting Gemini as part of their data management strategy, the company promises tools that don’t just execute commands, but also assist users in their daily lives.

At the conference, Abhishek Kashyap, Head of BigQuery Product Management, reported an acceptance rate of code suggestions generated by the Gemini AI Code Assist tool exceeding 60%. This performance not only illustrates the power of AI in data management, but also reflects a growing demand for dedicated AI solutions, as highlighted by Kevin Petrie, analyst at BARC US.

A tool adapted to the growing needs of data analysis

To meet this growing demand, Google has introduced advanced SQL and Python code generation and completion features in its BigQuery Enterprise Plus editions. These additions will integrate AI into critical tasks such as:

• Data preparation and cleansing
• Designing optimized data models
• Generating recommendations on partitioning and materialized views
• Creating conversational interfaces for easier data access

These new features aim to simplify and streamline the work of data analysts and engineers, allowing them to focus on higher-value tasks instead of spending time on basic data manipulation.

Data preparation assistance: a tangible return on investment

With the preparation assistance tool in BigQuery Studio, users can identify inconsistencies within their data files, resolve schema issues, or even automatically generate SQL code to correct errors. While this doesn’t entirely replace human expertise, it represents a significant step forward in automating repetitive tasks, leaving more time for strategic analysis. Tool

Interconnecting AI Agents: A Future Without Silos

Although Google Cloud has already made significant progress in integrating its AI agents, one of the key challenges lies in creating true interconnection between these agents. As experts indicate, a complete transformation to a silo-free system remains a goal. The goal is to ensure that agents work collaboratively, rather than remaining in separate compartments.

The Vision of a Family of AI Agents

Abhishek Kashyap shared Google Cloud’s vision of a set of interconnected agents, each specialized in different areas such as data engineering, analytics, and data governance. These agents will use the latest Gemini models to automate various tasks:

Upstream data preparation

• Optimized data pipeline design
• Development of business glossaries and metadata
• Data anomaly detection and processing
• The flexibility and responsiveness of these agents will not only inform user decisions but also continuously improve data quality. However, this ambition requires a clear roadmap to establish a harmonious framework within which the agents can interact easily.

Integration Challenges

Despite the enthusiasm surrounding these advances, a crucial question remains: how to ensure effective cooperation between the various AI agents? This challenge is all the more significant given that data scientists, business analysts, and data engineers often possess specific skills and methods that vary greatly.

To continue moving toward this ideal of integration, Google needs to communicate more about how these agents will be able to interact synergistically, thus reducing the risk of duplication of tasks and ensuring consistency in workflows.

Evolving toward data processing that combines SQL and generative AI

Google Cloud’s vision doesn’t stop there. With the introduction of the AI ​​Query Engine, the ambition is to merge the use of SQL with natural language prompts. This development represents a significant shift for data processing, as it will allow users to query both structured data and natural language queries.

AI Query Engine: Toward Richer Analysis

This innovation will enable BigQuery to leverage the power of advanced language models such as Gemini to query databases, generating deeper and more valuable insights. As Kashyap points out, this opens the door to a multitude of use cases, including:

Creating enriched features for AI models

• Performing advanced customer segmentation
• Discovering valuable, previously inaccessible insights
• Through this method, users will be able to ask questions intuitively and get actionable answers. This shift to a more conversational and accessible approach marks a paradigm shift in the way we interact with data.

The AI ​​Query Engine’s Prospects in the Technology Landscape

At the same time, giants like Amazon Web Services, Microsoft, and Databricks are moving in the same direction, integrating similar features. The need to provide a seamless and intuitive user experience is becoming increasingly pressing, revealing a general trend in the industry toward integrating AI-assisted data analysis strategies.

Competition

While this article doesn’t present a definitive statement, it highlights the major challenges Google Cloud will need to address to realize this ambition of a “family” of AI agents. Harmonizing agents to break down silos, while optimizing data processing through technologies like AI Query Engine, positions Google as a key player in the digital transformation of businesses. In a constantly evolving field, the next steps will be crucial to ensure these innovations are not only relevant but also aligned with user expectations.