Google Cloud’s 10 New Bold AI Products: Gemini Robotics, AI Coach And Gemma 3

CRN breaks down 10 new AI products, features and tools Google Cloud launched in March that every partner should know about, including new Gemma 3 open models, an AI Coach and two Gemini Robotics models designed to control robots.

Google Cloud launched a slew of new AI agents, tools, large language models and other AI innovations at a blistering pace this month.

The $48 billion Mountain View, Calif.-based cloud company is seeking to become a market leader in the AI era by providing a full-stack AI portfolio of hardware, software, accelerators, LLMs, agentic AI technology, and end-user productivity and collaboration via Workspace Gemini.

This month, Google Cloud unveiled its new Gemma 3 open models, a unified console for building AI agents, two Gemini Robotics models designed to control robots, and a Gemini-powered tool designed to improve employee productivity called AI Coach, to name just a few.

Google Cloud generated $12 billion in revenue during fourth-quarter 2024, representing a 30 percent year-over-year increase. Google’s cloud business is the third largest in the world with 12 percent share of the market, trailing only Microsoft at 21 percent share and Amazon Web Services at 30 percent share.

CRN breaks down the 10 new Google AI products, features and innovation launched in March that every Google Cloud partner should know about.

New Gemma 3: Google’s ‘Most Advanced’ Models Yet

Gemma is Google’s family of open AI models launched in 2024.

The company launched its new Gemma 3 this month, which is a collection of lightweight, open models for customers and developers to build AI applications capable of running on anything. Google says Gemma 3 is the best in the world for running on a single GPU or AI accelerator.

“These are our most advanced, portable and responsibly developed open models yet,” said Clement Farabet, vice president of research for Google DeepMind, in a blog post. “They are designed to run fast, directly on devices—from phones and laptops to workstations—helping developers create AI applications wherever people need them.”

Compared with previous Gemma models, Gemma 3 has expanded from around 8,200 tokens to 128,000. Google also unveiled a new offering for image safety, dubbed ShieldGemma 2, which integrates with Gemma 3 to help block unwanted images.

Gemma 3 comes in a range of sizes from a text-only 1-billion-parameter model to a 27- billion-parameter version, enabling customers to choose the best model for their specific hardware and performance needs.

Click through to read the other nine new Google Cloud AI products recently launched.

New Gemini Robotics Model

Gemini Robotics brings Gemini 2.0 innovation into the physical world by allowing robots of any shape and size to perform a wide range of real-world tasks.

The first model is Gemini Robotics, an advanced vision-language-action model that was built on Gemini 2.0 with the addition of physical actions as a new output modality for the purpose of directly controlling robots.

Gemini Robotics can tackle complex, multistep tasks that require precise manipulation such as origami folding or packing a snack into a Ziploc bag.

It taps into Gemini’s advanced language understanding capabilities and can understand and respond to commands phrased in everyday, conversational language as well as in different languages.

“Gemini Robotics leverages Gemini’s world understanding to generalize to novel situations and solve a wide variety of tasks out of the box, including tasks it has never seen before in training,” said Carolina Parada, senior director and head of Robotics for Google DeepMind, in a blog post. “Gemini Robotics is also adept at dealing with new objects, diverse instructions and new environments.”

Google Cloud’s New AI Coach

Google’s new AI Coach is a Gemini-powered tool designed to improve employee productivity with real-time and contextual coaching and guidance.

“It provides contextual step-by-step agent coaching to guide customer service representatives through complex interactions, upsell opportunities, compliance steps, automated information retrieval and follow-up steps,” said Antony Passemard, director of product management, Applied AI, at Google Cloud, in a recent blog post.

AI Coach—which is part of Google Cloud’s Customer engagement Suite—also aims to speed up time to resolution for customers and improve satisfaction.

Gemini Robotics-ER

Alongside the Gemini Robotics new model, Google launched an advanced vision-language model called Gemini Robotics-ER [embodied reasoning].

The new Gemini Robotics-ER enables a robot to have an advanced understanding of the space around it, known as spatial understanding. It lets developers run their own programs using reasoning abilities offered by Gemini 2.0.

Gemini Robotics-ER improves Gemini 2.0’s existing abilities like pointing and 3-D detection by a large margin, Google said.

It can instantiate entirely new capabilities on the fly. For example, when shown a coffee mug, the model can intuit an appropriate two-finger grasp for picking it up by the handle and a safe trajectory for approaching it.

“As part of our efforts, we’re partnering with Apptronik to build the next generation of humanoid robots with Gemini 2.0. We’re also working with a selected number of trusted testers to guide the future of Gemini Robotics-ER,” said Google DeepMind’s Parada.

New AI Trainer

Google Cloud’s new AI Trainer tool provides AI-powered simulation, personalized training and coaching to address service challenges, information overload and difficulty handling feedback.

“This allows new hires and less experienced support staff to learn in a safe, controlled environment to avoid inadvertently delivering lower-quality responses and risking negative impact on an organization’s brand, while providing self-service tools for them to continually improve on quality,” said Google Cloud’s Passemard.

Agent Trainer aims to boost business performance and customer satisfaction by delivering ROI through lower training costs, tailored training recommendations and faster agent proficiency—ultimately reducing turnover, Google said.

New Agent Desktop Application

Targeting the Contact Center-as-a-Service (CCaaS) market, Google Cloud has created a new stand-alone agent desktop application.

The new stand-alone agent desktop client provides an out-of-the-box, omnichannel and multimodal interface that seamlessly integrates with Google’s Agent Assist capabilities.

“By seamlessly integrating with Agent Assist capabilities, like generative knowledge assist and post-interaction summarization, it helps improve agent efficiency and productivity,” said Matthew Clare, outbound product manager, customer experience, at Google Cloud, in a blog post.

“It provides a flexible, customizable layout, new UI modules that integrate with third-party data sources and easily integrate with a system of record like a customer relationship management system,” Clare added.

The agent desktop application is part of Google’s Customer Engagement Suite.

Two New Contact Center-as-a-Service Features

Google launched two new capabilities around its CCaaS offerings, including a new web and voice Co-browse feature that allows support representatives to see what customers are viewing.

“This feature helps customer service representatives see what customers see so they can provide more effective service, lower abandonment rates, and help improve both customer satisfaction and loyalty,” said Google Cloud’s Clare. “Co-browse delivers an improved digital experience with real-time, collaborative browsing that allows support representatives to deliver seamless digital customer experiences and improve first contact resolution and customer satisfaction.”

The other new CCaaS capability is Google’s new customizable dashboarding and reporting tools to improve access to information and help supervisors and managers better measure and manage their operations.

Enhanced reporting, powered by Looker that’s embedded in the product, provides new customizable dashboarding and reporting tools to improve access to information.

New Conversational Agents Console

Google Cloud has launched a new unified console for building AI agents that combines generative AI and rules-based controls to enable rapid building of AI agents with realistic, natural-sounding inflection and expressive conversations for self-service experiences.

The Conversational Agents console also provides new evaluation capabilities to benchmark agent performance and improve reliability and quality at scale.

“We also introduced observability, evaluation and test-case instrumentation into the Conversational Agents console,” said Google’s Passemard. “This enables customers and partners to validate agent quality at scale, with tools to consistently monitor and improve their self-service experiences.”

New Google Cloud Region In Sweden To Enable AI

Google Cloud is opening up its AI portfolio to a new customer base by launching a new cloud region in Sweden this month, marking Google’s 42nd cloud region globally and 13th in Europe.

Google is now offering organizations in Sweden a new way to build resiliency, reduce costs and accelerate sustainability impact through the smarter use of data and AI.

The new cloud region provides Swedish businesses and individuals with a new platform powered by Google Cloud’s AI, machine learning and data analytics technologies.

This region also directly addresses data residency requirements and digital sovereignty concerns, removing key barriers to cloud adoption for many Swedish organizations. Customers can harness the full potential of Google Cloud's services while maintaining control over their data's location.