AI tools

Beginners guide to use Google Gemini

Google has recently replaced Google Bard to Google Gemini, In this article we will deep dive into using Gemini for various use cases. Make sure you toggle on respond in real-time for quicker responses. Extensions are not available just yet for Google business accounts, so make sure you’re subscribed for future updates. Now that extensions are enabled.

You can use the @ shortcut to access a list of commands for Gmail, Google Docs, Google Drive, and YouTube.

Lets explore how you can use Gemini as your daily driver.

Access Gmail with Google Gemini

In Gmail, I like to ask Gemini to search for all emails related to one specific topic. For example, “@gmail give me all emails that mention Graphic Design,” and Gemini knows that I’m looking for emails that contain Design related stuff only. If I try to manually search for those same emails in Gmail, I would see a ton of unrelated results even if I use search operators like from and quotation marks. I would also need to right-click open in a new window to open multiple emails, whereas in Gemini, I can open these emails in new tabs directly and just start reading. Another practical example would be confirmation emails. “@gmail find all flight confirmation emails from this year,” and usually, these are annoying to find if you don’t use labels in Gmail, but Gemini finds them.

Access Google Drive with Gemini:

Moving over to Google Drive,We can add any search query. “At Google Drive, could you please find all documents related to Instagram Posts, a folder of posts where ive stored all my older Instagram post that ive mode for my clients in the past,” and boom, Bard understood the context and gives me all the Instagram posts I had saved ages ago. Whereas if I went to Google Drive and searched for Posts, tons of irrelevant results.

Use Google Docs with Gemini: Now on to Google Docs.

You can at Google Docs input the document name in quotation marks. For example, “EY year one self-assessment,” and ask Bard to only consider the content of this specific file in your answer. Your task is to take an action, in this case, analyze the file and share the top three takeaways. And Bard will pull information from that specific document and take that action. Pro tip: if you can’t remember the exact name of the file, you can start off with, “At Google Drive, find the self-assessment for my first year at EY. Ah, here it is. At Google Docs, please analyze and summarize the top three takeaways from this document.” You can take this a step further and include a time frame in your prompt as well. For example, “At Google Docs, what is the most recent position I held in my latest resume?” And Bard will consider the last updated date when deciding which file to pull information from.

Analyze Files and Data with Google Gemini:

Apart from the workspace commands, I also regularly use the image upload feature to distill important information. For example, if I were to present you with this chart, you might be overwhelmed by all the data points. Thanks to Bard, we can Command or Control + C to copy this image, Command or Control + V to paste it within Bard, and input, “Assume the role of a senior business analyst with over 20 years of experience. What are the top three key takeaways for this chart?” And we’re told that the pace of innovation in open-source LLMs is accelerating, the gap between open-source and private models is narrowing rapidly, and we’re on the cusp of a paradigm shift in AI. If we refer back to the original article, we find that Google Bard is correct. New research shows that open-source competitors are catching up to their private model counterparts.

Use case number two is something I do every day at work. I see a table with data in it that I want to analyze and play around with, but it’s in a non-editable format like a PDF. I take a screenshot, paste it into Gemini, and say, “Give me the data in table format. Be as accurate as possible.” And now with one click, I can export this entire table into a Google Spreadsheet to start my analysis. This literally saves me hours every week. Use case number three is for those of us who use text-to-image models. I find an AI-generated image I like, copy and paste it into Bard, and say, “Reverse engineer the text prompt for this image so that I can use that text prompt to get a similar image in DALL-E.” Bard outputs a simple prompt and a lot of additional details for us to consider. But even if I just take the simple original prompt and copy-paste it into ChatGPT, ChatGPT generates a pretty similar image using its integrational DALL-E.

Upload PDFs onto Google Gemini:

What about PDFs? Can I upload a PDF file onto Google Bard? The answer is technically no, but definitely yes. Meaning, although you can’t upload a PDF file, there are two workarounds. Method number one, if the PDF is already hosted online, like this BCG report, you can simply paste the link into Bard and say, “Analyze this report. Tell me who the intended audience is and give me the top three takeaways. Be as specific as possible in your response by using examples from the report.” And Bard will be able to take action directly. Method number two, upload the PDF onto your Google Drive and using the @Google Drive command, ask Bard to pull information and take action for that uploaded file.

Google Gemini’s YouTube Vision Feature:

Moving on to the YouTube Vision functionality, you might have seen other videos talk about how you can ask Bard for YouTube video recommendations, and in my opinion, that’s kind of silly because it’s much faster and more accurate to just search directly on YouTube. But where Bard shines is if you want to analyze a video. For example, “Why did this PopCorners advertisement perform so well with over 5 million views, 250,000 likes, and over 6,000 comments?” I can use the @YouTube command, paste the video URL, and input, “Analyze this advertisement and tell me who the target audience is, what the user insight is, and why it performs so well and why the engagement level is so high.” And Bard summarizes and answers my questions and expands on those insights. Another way I like to use the YouTube Bard integration is to take a longer video that I’m interested in but I don’t have time to watch. For example, this 2-hour podcast between Huberman and Goggins on how to build immense inner strength. I go to Bard at YouTube, paste in the YouTube URL, first, share the top three takeaways from this video, and second, what are the five actionable steps I can take to build inner strength, give me specific examples from the video. This allows me to quickly distill learnings from the video, and now I can just focus on a specific 10 to 15-minute segment to dive deeper into a particular concept.

Gemini’s User-Friendly Features:

That’s actually a very underrated tip. In addition to these practical applications, let’s go through some of the features that make Google Bard so user-friendly. For example, if you open up a new chat, you can and should click on these default options to see recommended prompt structures from Google. For example, this one: “Understand a complex topic.” I just need to replace the text in blue here. “Briefly summarize this concept.” Let’s choose large language models. “You’re an expert on this concept and able to explain it in a clear, concise manner.” Blah blah blah. And the biggest benefit to using these default prompts by Google is the more we read and use them, the better we get at writing and structuring our own prompts. Google is basically giving us training wheels for prompting. I can also choose to modify the response by clicking the icon here to make it shorter, longer, simpler, more casual, more professional. How do you professionally say, “I’m not accepting that meeting because it’s a waste of my time?” Declining this invite as I don’t feel my participation in this discussion is required. I can also click the Google icon here to fact-check the response by Google Bard. Green highlight means Google is fairly accurate, and if you see yellow, it means it’s a little bit iffy.

I actually click this drop-down menu here to view other drafts quite often because other drafts might give you a better format. For example, this one, draft three. “Here are three key things about LLMs,” versus the original draft just tells you what LLMs are able to do. So it really depends on your personal preference. You could just cycle through the different drafts if you prefer. To interact with Bard via voice, you can click on the microphone icon here to speak your prompt. “Give me the funniest and most hilarious dad joke you can think of.” And after the answer is generated, click on the speaker icon to listen to the output. Did you hear about the restaurant on the moon? Great food, but no atmosphere.

You know what makes this super realistic? Gemini actually goes ahead and explains the joke, how “atmosphere” has a dual meaning. That’s what dads do when no one laughs, right? Anyways, if you click on the share button down below, you can choose to either share your latest prompt and response or the entire chat publicly. And I personally like to generate a more informative headline before sharing the link with others. You can even find a repository of all your public links by clicking on the settings icon, your public links. Pro tip: I like to save prompts that have generated high-quality outputs by clicking the three dots here, clicking pin, giving it a new name, adding an emoji, and clicking pin so that I can easily reference this in the future

By Hrushikesh Paygavhan

UX designer at Customer Capital. With 7+ years of Design experience, I belive in passing the knowledge to everyone. I help UX designers upskill themselves, and stay updated in the design field.

Leave a Reply

Your email address will not be published. Required fields are marked *