On Tuesday, Google introduced a new agent AI feature aimed at data engineers and data scientists.
The Tech Giants have announced their agents at Google Cloud at the Tokyo 2025 conference.
The new products include: BigQuery data engineering agent. Data Science Agent for BigQuery Notebooks. Conversational Analytics Agent and Code Interpreter. Spanner migration agent. Google has also introduced the Conversation Analytics API and Gemini Cli Github Action. Everything is in the preview.
Building Agent AI
With a focus on agent AI, Google comes as excitement and what some call hype, as it has little proven autonomous AI agents.
AI experts predict that most significant companies will pursue agent AI initiatives in the coming years.
“The long-term journey a company is doing is rethinking it, essentially dismantling it, rebuilding it in a native context of AI and agents,” said Chirag Dekate, analyst at Gartner.
Part of achieving that long-term goal is finding ways to help data engineers and scientists use agent AI technology to increase productivity. Foundation model providers like Google, AWS, Microsoft, and OpenAI are among the vendors considering this issue.
Previously, Agent AI has focused on customer service and developer-centric applications and applications. This is primarily because data conversion workflows can be complicated, Dekate said.
Agent AI for Data Scientists and Engineers
“When you look at data engineering agents, data science agents, and so on, what's starting to appear now is to start attacking more complex problems that are common in corporate contexts,” he said. Such agents help data scientists and engineers build data pipelines and processes.
“I have to say as an analyst and as a practitioner, Godsend, to use generator AI as a means of augmenting these processes, and to automate as much as possible.”
He added that data scientists and engineers spent most of their time searching for data in their data repository instead of creating applications, workflows and data infrastructure.
So, he said that agent AI tools like Google will have a huge benefit to data scientists and engineers.
Google's new agent
One new tool is Google's Data Engineering Agent for BigQuery in preview. Agents allow data engineers to simplify and automate complex data pipelines. Agents allow users to use natural language prompts to streamline workflows, from ingesting data from sources such as Google Cloud Storage to converting data and maintaining data quality.
Another new feature is the Spanner Migration agent. This AI-equipped service allows data engineers to simplify migration from database systems such as MySQL.
Meanwhile, the Data Science Agent, powered by Gemini's leading language models, includes features such as exploratory data analysis, data cleaning, machine learning prediction, and capabilities. Agents can plan and run the code and provide reasons for the outcome, but data scientists can provide feedback and collaborate, Google said.
Code interpreter is for business users and analysts. Convert challenging natural language questions into Python code.
Google also launched the Gemini Data Agents API to coordinate various agents, allowing data scientists and engineers to connect agents to the system. The first API in this group is the Conversation Analytics API. Users can integrate natural language processing and code interpreter capabilities of Google's Analytics Platform Looker into their applications and products.
Data scientists and engineers can create custom agents using the Agent Development Kit and the Data Agent API.
Starting point
Companies looking at Google's Agent AI Toolkit will find that data science and engineering teams can become more agile and accurate for simple workflows. However, for complex workflows such as ecosystem integration, vendors are likely to address these complexities in the future, Dekate said.
“This is a good starting point for companies to think about ways to improve data engineering and data science team productivity,” Dekate said.
Shimmin said the trick with these Agent AI and Generate AI (Genai) models is that users have to trust what they're doing.
“These frontier-based models are better at using tools and are good at structured output learning,” he said.
He added that AI agents understand data better than before, but users need to remember that the company is running with semi-structured information that is either between the agent process or moves.
“It's all a massive language model that understands this kind of semi-structured content and data, and how to take things in turn, understand them for context, understand their meaning and take action,” he said. “It's amazing.”
In addition to agent tools for data scientists and engineers, Google has introduced Gemini Cli Github Action, a free AI coding agent.
Other developments
Google has also expanded its collaboration with Wells Fargo. Companies are partnering to help financial institutions use Google Agent Space to build AI agents in a variety of sectors, including businesses, investment banks and customer service.
Meanwhile, Google has recently launched several Gemini Advances.
On July 29th, Google revealed that its video generation model, VEO 3, is generally available on the Vertex AI Genai platform. VEO 3 Fast, a faster way to turn text into video, is now generally available with Vertex AI. Both VEO 3 and VEO 3 FAST will offer image-to-video functionality this month.
On August 1st, Google Deepmind revealed that Gemini 2.5 Deep Think has become more widely available to Google AI Ultra Subscribers.
Esther Shittu is an Informa TechTarget News Writer and Podcast host that covers AI software and systems.
