Earlier in 2024, Google introduced Gemini, a multi-modal large language model (LMM) designed to integrate across Google Workspace applications. From Gmail to Slides to Sheets, Gemini offers AI capabilities through a chat window that appears with a simple click on its omnipresent icon.
The MG Labs team has been putting Gemini to the test across key Workspace tools—Docs, Sheets, Drive, Slides, Gmail, and the standalone Gemini web and mobile apps—to evaluate its real-world impact on marketing and PR workflows. In this blog, we explore its strengths, limitations, and overall utility.
Note that the summary findings that follow are for Gemini 1.5, which is currently available in Google Workspace as of this writing. Google has announced that Gemini 2.0 will be widely released sometime in early 2025.
For Gmail
The Gemini feature for Gmail is located in the righthand corner of the screen and can be activated with a simple click of a button, prompting users to make comments or ask questions. Unsurprisingly, its search capability is the shining star—far surpassing the results that Gmail’s standard search box renders.
For example, commanding Gemini to “find all emails about Mobile World Congress from December 1,” a search relevant to a major event we support each year for clients, will quickly show any emails that were sent or received on the aforementioned date that pertain to the conference. This is particularly helpful when searching for emails that have nondescript subject lines or numerous threads, some of which cover different topics.
In addition, Gemini for Gmail enables users to seamlessly outline the body of an email by clicking the “summarize this email” button, located below the subject line. This comes in handy for longer emails and e-newsletters as it highlights the key themes or messages within a few bullets. Gemini for Gmail encapsulates the idea of “working smarter.” It does all of the monotonous searching and summarizing for you, freeing up time to focus on more strategic work.
One area where Gemini for Gmail still struggles is email composition. Although the tool exists, it typically needs a substantial amount of guidance to formulate a well-thought-out email. In fact, it took Gemini about the same amount of time to formulate an articulate email response as it did for us to write a similar response manually. The purpose of Gemini is to maximize work efficiencies but candidly, the composition tool isn’t quite ready for regular use. We recommend limiting the use of Gemini for Gmail to its strong suits for now—email search and summaries—until its composition capabilities improve.
For Slides
While using Google Slides, Gemini offers three categories of use cases: create, summarize, and ask. Within create you can build both images and slides, but neither is quite ready for extensive use. Visuals tended to be generic and often irrelevant, significantly trailing other LLMs’ image-creation capabilities at this stage.
Slide creation was only useful when the tool was asked to provide a “wrap-up” or “key themes” slide. This gets to the primary takeaway of our evaluation: the value of Gemini for Slides at this stage is summarization. Even the “ask” bucket prompts the user to request various summaries of existing content rather than helping to build anything new.
The challenge is that good decks are inherently supposed to be high-level summaries of strategies, plans, and initiatives—particularly at our agency. Outside of long, content-heavy decks like analyst reports or company overviews that we receive from a third party and are asked to digest quickly, the value of Gemini for Slides is still relatively narrow.
For Docs
Gemini is conveniently embedded into Google Docs for easy access, but its performance presents strengths and limitations that affect overall effectiveness. Gemini has proven useful for brainstorming, summarizing, rephrasing, and making stylistic changes. Its ability to search Google Drive for reference documents also adds value.
However, its reliance on highly specific prompts for more complex tasks can hinder efficiency. In addition, slow response times and frequent output discrepancies often require multiple attempts to achieve the desired results.
Overall, Gemini for Docs shows potential but needs refinement to fully realize its capabilities. Its current limitations, especially in terms of efficiency and output quality, impact its usefulness for daily tasks.
For Sheets
Gemini is seamlessly integrated directly into Google Sheets, accessible via the Gemini button at the top right corner. At the time of our evaluation, its effectiveness was somewhat limited, especially with complex or poorly labeled spreadsheets (which can be common). It struggles with nested data sets, cannot read multiple tabs simultaneously, and occasionally produces inaccuracies or “hallucinations” in its outputs. Response times can also be slow, taking up to 20 seconds per query.
Gemini can help create tables from subsets of data, generate complex formulas, and draft initial email summaries based on spreadsheet content. However, given its current limitations, it’s advisable to use it with simple, well-structured spreadsheets and always double-check its outputs for accuracy.
As of today, it can be used as a supplementary tool rather than a primary resource in Sheets. As the tool evolves, we will continue to experiment with it to discover more efficient use cases that could enhance productivity in the future.
For Drive
Gemini for Drive is embedded within Google Drive, easily accessed via the Gemini button or by right-clicking on a file and selecting “Ask Gemini.” It’s good at summarizing and exploring documents such as PDFs, Google Docs, and Google Slides (although it’s currently limited to a handful of documents at a time).
Users can ask detailed questions about the content, with Gemini identifying commonalities across documents and providing summaries. However, its capabilities are limited when dealing with more than a handful of files, and its search functionality can be inconsistent. Gemini for Drive may return irrelevant results or struggle with specific search prompts. Translation features are also confined to small text snippets vs. full documents, and response times can be slow.
The tool’s value, however, lies in its ability to summarize lengthy documents, compare content across multiple files, and generate emails based on document content. It saves users time when reviewing and synthesizing information without manually sifting through each file. We recommend incorporating Gemini for Drive into regular workflows for document exploration and summarization while being mindful of its limitations.
Gemini Web and Mobile App
The Gemini web app is a chat interface that gives you direct access to Google AI. Users can take advantage of its features and use it to brainstorm ideas, develop plans, learn about new topics, and support writing. According to Google, the goal of this tool is to enhance your productivity, creativity, and learning. Users can access it on their desktop or mobile device by signing in with a Google Account.
The tool also has different “Gems,” or specialized chats, that are created for more in-depth and specific circumstances. These include use cases such as “career guide” and “coding partner,” In addition, it can generate images based on user requests. While the application is powerful in some areas, it is lacking in others—especially when compared to competitor AI chatbots.
Specific areas of improvement for the application include delayed response time, limited ability to answer complex questions, and limited image generation capabilities. It is not hard to stump the AI or give it a prompt that results in a very limited response. While we recommend experimenting with the application, users should also compare and contrast Gemini’s answers against other AI apps or traditional search engines to ensure Gemini’s outputs are accurate and reliable.
Summary and Overall Value
After our analysis and evaluation of Gemini 1.5 for Google Workspace and the standalone app and website, we came away with mixed results. The LLM works better with some applications—like Gmail and Google Drive—than others—like Google Sheets or Google Slides. It has clear limitations around image generation, understanding highly unstructured spreadsheets, or assisting on complex tasks with multiple steps or input vectors.
However, as with most LLMs in the market, Gemini continues to improve. We will revisit this evaluation with the general availability of Gemini 2.0 for Google Workspace and report on those improvements.
For more information on AI use cases for marketing and PR, visit the MG Labs page, where we analyze the latest AI tools, platforms, and features. And don’t forget to check out our Artificial Intelligence page to learn how we’re helping AI-focused companies bring their solutions to market.
