“Gemini for Google Workspace” is fueled by Gemini Google’s most capable set of large language models (LLMs) it has ever built, setting a lead for others to follow, according to new research from GlobalData.

“Gemini for Google Workspace represents a significant upgrade to the Workspace platform,” says Gregg Willsky, principal analyst at GlobalData. “Gemini is positioned to become Google’s flagship AI tool, powering many products and services within its portfolio.”

GlobalData analysis finds Gemini is a special LLM capability being built deep into the Workspace set of apps and bringing a new level of generative AI (GenAI) functionality. For example, creating custom sales pitches from transcripts of video calls; assembling tables in Sheets driven off raw data and content; and building a slide deck from text prompts.

The Gemini set of LLMs is natively multimodal unlike other LLM models and the new Ultra 1.0 version is the first solution to outperform human experts on MMLU (massive multitask language understanding) tests meaning it can use its “reasoning abilities” to “think more carefully” before it answers questions.

“A question that has plagued Google and its rivals is what makes their ‘brand’ of AI different?” says Willsky. “Now, Google is the first to have an answer – the Gemini set of LLMs.”

Productivity enhancements have been delivered through a proliferation of virtual assistants from every major competitor in the cloud-based collaboration space. These assistants represent the centerpiece of a “permeate the platform” approach to AI that has seen the technology applied across competitors’ full portfolios.

“Google’s rivals will need to evaluate the LLMs powering their respective collaboration platforms and consider upgrading them,” says Willsky. “Gemini’s ability to process a range of data types and the effectiveness at which it does so is well worth mimicking.