Transcription
Gemini is the latest AI model developed by Google, a competitor to Chad GPT. And what set it apart is not only the fact that it is one of the most secure options on the market, being maintained by one of the industry giants and having a very straightforward HIPPA compliance feature, but also the fact it has a few very unique features up its sleeve that you probably haven't heard about, such as deep research, creating custom gems, or AI experts for your common workflows and live mode.
In this video, I'm going to go over how you can sign up for Gemini, how you can use it. I'll showcase all the main features, both for basic and advanced workflows. And so, by the end of this video, I would hope that you have a full understanding of what Gemini offers and how you can use it in your practice today. Let's get going.
Hey there, my name is Martin and I'm chief technology officer at Swans, where my team and I have helped hundreds of law firms implement AI and automation inside of their practices. As we go through this video, I'll also make sure to cover some of the doubts that you may have about using AI altogether inside of your practice, such as data security and chances of hallucinations, i.e. models making stuff up as you talk with them. So let's now get started and see how you can sign up for Gemini.
Now Gemini is included as part of Google Workspace. Meaning that if you're already using Gmail, Google Drive for your work and you have a Google Workspace account, Gemini is already actually included as part of your subscription. Meaning you don't need any extra cost. If, on the other hand, you're using Outlook, the Microsoft suite, in that case, you will need to sign up for Google Workspace, but the prices start at just $14 per month per user, which is also extremely competitive to some other providers such as Jupto cloud that only offer service. Obviously, with this, you also get all the other functionalities that Google Workspace provides.
So to sign up, if you haven't done so already, just navigate to workspace.google.com, click here, start trial. So you'll get 14 days to try out Google Workspace and all of its AI features, and after that, be billed depending on the number of users that you have. And as I mentioned at the beginning of the video, Gemini for Google Workspace is one of the most secure options on the market. Not only do you get all the security features that Google Workspace provides, such as two-factor authentication, IP restrictions, and any other extra extra security policy that you may need, you may want to define for a firm, but by using Gemini through Google Workspace, you're also making sure that your data remains your data. Uh, Google ensures that the data that you provide to Gemini will not be used to train its models.
If you then need additional security, you need to meet additional security standards such as HEIPA, then actually Google Workspace also has a very easy feature to enable that. I know that a lot of firms, for for example, in personal injury, refrain from using cloud or chpt. I mean, cloud doesn't have a really he compliant version unless you're an enterprise, and ch as well can really do it on a on a on a case by case basis. With Google Workspace, it's extremely easy to sign a BA with Google to make sure all your sensitive data is protected. Just navigate to the Google admin settings, account, then account settings, and if you scroll down in the list, you'll see security and privacy, additional terms, terms, and in here, you'll be able to review and accept the standard BA agreement, uh, for HIPPA compliance if you need that.
Once you have fully, uh, signed up, here's how the Gemini screen will look like. You can access it by going to gemini.google.com or navigating through your, um, app selector inside of the Google Workspace. The the interface is quite minimalistic. Here at the top on the left, you are able to select which model you would want to use, uh, for Gemini. So there are multiple different kinds of modes or models that you can use as you work through the program. Each one of them has its specific features, right? So, for example, two, uh, all flash is great for every detailed task. It is the most, it is the fastest model, right? So if you just need a quick response, this is really the best model to be using. 2.5 flash experimental uses some more advanced reasoning features, right? So you can throw some more complex tasks, tasks at it. Uh, 2.5 Pro is the the version you can use for most complex tasks, right? So it will really take longer, really to think about the questions that you are asking and come up with with a solution, right? So the execution of the query will be somewhat longer, but then you'll also get more in-depth and and thoughtful, let's say, responses. And finally, we have deep research with 2.5 Pro. Deep research is an extra feature which takes even more time, uh, which ensures that the model really scouts the internet for whatever question you're asking it about. We'll see exactly how each one of these works as we go through this video to make sure you're comfortable with using all of them.
Apart from that, we also, if you go over here in the hamburger menu, we can start a new chat. Also, if you have any already active chats, you'll be able to switch easily in between them. And then we have these gems over here, which is kind of custom AI experts you can define. We'll also cover them in in just a moment. Down below, you can manage your gems. Uh, you can see the account activity. And then here you can turn, um, on or off the the dark theme.
Let's now focus on the main window. And how do we actually start chatting with Gemini? So there are a few ways that you can interact with Gemini. First of all, you can just type in your question in here and send it. Or additionally, you can also use your microphone, in which case you may need to give permissions to them to do that. In this case, it will listen to whatever you are saying, right? And translate it into text, uh, so that you can quickly then type up, right, your prompt, uh, using this method. And additionally, you'll be able to add files. This can be files from your computer. It can be images, could be audio files as well, PDFs, docu, uh, docx documents. And additionally, if you connected your Google Drive account and that, um, feature is enabled, uh, you'll also be able directly to add documents from your Google Drive without having to upload them from your computer. Lastly, you can enable again deep research in here and canvas. It's another great feature, uh, which we'll see in just a moment, uh, as well.
So, let's, uh, actually write our first prompt. We'll switch to, uh, 2.0 of flash, the quickest model, and we'll ask it, what are the most common US visa types? Let's see how it does. So I'm just going to press enter to send it in, and as you can see, we get a response quite quickly. What it's doing right now, it is using a combination of knowledge that it has, in addition, actually, it's looking up and grounding its search on, uh, Google, Google search results. So whenever you see, um, a symbol just like this next to the answer, it means that it has sourced that information from either a document that you have provided it or actually from Google search. And so if you look it up, actually, there are two sources in here, right? Um, one, both from government websites. We can easily, you know, click on them and and actually check those responses, and we can obviously, you know, see which parts of it, of the actual response are referenced in here. This actually solves a big problem, uh, that, um, AI models experience, which is hallucinations, right? Meaning that they make stuff up that doesn't really exist, right? Such, for example, you know, creating fake cases, we have all heard stories about that. So this helps you ensure, right, that whatever data that it is providing back to you is actually grounded in some truth, grounded in some internet search results, and obviously then you can verify, you know, especially for more sensitive stuff, you can verify exactly if the model has given you the right response or not.
If you scroll down, right, we have this obviously, um, all this breakdown, and as well, we have any sources and related content in here, um, as well, as well displayed with all the sources that used for this response. An additional feature that the flash models have is actually, if you click on this three dots over here, it's to double check the response, which will do an additional search through Google and check whatever information the model has has provided to you. And in just a few seconds, as there we go, it marks in green all the text that has been able to verify through those Google search results, right? In addition to where that was sourced from. And in case that some of this information, it wouldn't be able to find references for, in that case, it will highlight it in in in orange, right? Making sure that you can see which, which stuff it is, it is less, less sure about.
Now, there are a few more things that we can do over here. We can just provide feedback, essentially to Google, you know, whether it's a good or bad response. Don't, don't necessarily bother about that. We can redo, right? So the the model will retry answering the questions. We can share and export. And the great thing about Gemini is it's, it's very well integrated in all of the, um, Google services. Meaning that you can easily from here export, for example, this into a Google doc and continue editing it. There we go. A new document created and continue editing it in Google Docs directly with all the features that obviously it provides, right? Very easily. Uh, you can also draft in Gmail, and we'll create a new draft in your Gmail account, and you're able to then obviously send it as an email.
In addition to that, on the three dots, uh, we can, as mentioned before, double check the responses, we can copy the text, right, to paste somewhere else, or we can also listen, uh, to, to the response if we would like to do that. We can also then obviously, you know, ask any follow-up questions. So, for example, um, convert the results into a table format. We'll do now. We'll just take all of that and create a table for you, right? With all those options, and as well, we can then easily export it into now Google Sheets and continue editing, uh, this table in there.
Let's now shift towards a more complex task of document analysis. So what I have done is I clicked on the menu, clicked new chat. So provides us a blank slate to be working with. Make sure that as you're working with the model, um, and asking different questions, you create new charts, which will mean that it won't reference any say responses or questions, um, any conversations that you had, uh, previously, right? In in the, uh, regarding other questions. So, um, I have here an acquisition agreement, um, in between two companies. It's 132 pages. The great thing about Google Gemini, as compared to some other models on the market, is also the fact it has quite a large context window, meaning it can take a lot of information in based on the files and text that you provide to it. And, um, as, as, as of May of 2025, that context window is actually 1 million tokens, which roughly translates to about 1,200 pages of of text in PDF format. So quite a lot of data we can actually input into the system.
And so what are we going to do is first of all, we're going to change our model to 2.5 Pro experimental, uh, which will, um, show us, which will essentially extend the thinking abilities of the model and and provide better results for our use case, which is now more complex. And I'm going to click the plus button here, select files, and select our acquisition agreement. You can also upload multiple files if that's required. And so let's, uh, for example, ask, highlight any key red flags/pitfalls in the agreement for the buyer. And so it will take somewhat longer to respond compared to the flash model. Um, again, the reason being that it is doing some more thinking behind the scenes. You can actually see and expand the exact steps, kind of like that, it is analyzing in here. And let's wait for its response. Depending on the amount of information that you throw into it, it can be as quick as a few seconds to potentially a minute, a minute or two.
So there we go. We have our response, thinking quite, quite thorough. And as previously with the Google search, right? Obviously, right now we don't have Google search. We're just working with the file. We still have these symbols over here, which means that that part of the text has been sourced or grounded in the information that we found in the document, right? So, for example, in the inter company agreements, if we click learn more over here, it will show us the pages where it has referenced that material, right? And even show a small excerpt about, for example, collected with the intercomp equity agreements, right? Um, we can see exactly, right, where that information is sourced. We can also quite easily then scroll down, source and related content. We can obviously open up the doc in here, or we can also open up the doc that we actually uploaded directly in the in the software and, you know, go to page 26 and if needed, we can check, right, that the intercompany agreements in here have been defined and what, um, other details exactly details about it to verify what the model has told us. So that's how you can do document analysis with with Gemini.
Um, let's now go into a more creative task that will leverage this, uh, canvas feature. So I will, I'll start a new chat. And let's imagine that I run a PI practice. I want, um, Gemini now to help me draft, uh, an article for for LinkedIn. I'll change the model now to 2.5 flash, and we could also perfectly fine use 2.0 or flash. And let's, um, say something like, draft a LinkedIn post for a lawyer in Arizona that wants to talk about how liens are handled in API to case. Target audience, a prospective clients. That's definitely not a great prompt, as well, but for the purpose of this discussion, that will work. And I will enable this kind of feature, and you'll see in just a second how, how that looks like. So let's, let's send it. Response should be quite quick. And there we go. We have the draft of this new post over here. As you can see, it opened up this new window. We can close it and reopen it just like this. Created in the canvas, and essentially canvas is kind of like almost like a Google doc, right? So I can easily make, as well, any changes that I want to the document in here. I can change the formatting, right? Um, I can add, you know, bulleted list, numbered list, etc. Um, and then easily copy that and paste it into another document or another software. Or I can also actually export to Docs, as you've seen previously, right? Google Docs and continue the editing in there.
The great thing is it's not only an editor, but actually you can work as you do more prompting, right? As you speak with the model, it will actually be able to make changes, um, to, to this. So let's see how that looks like, right? Let's say, for example, we want to add more hashtags at the end, right? And so what's going to do, instead of writing up a completely new response, it will just modify this kind of thing here. And as you can see, it has added more hashtags in here. In addition to that, we can actually modify specific sections of the document, right? So, for example, right, I have selected this first paragraph, and what we can do is we can change its length. We can make it shorter or longer, right? Let's say very long. What it will going to do, it will going to extend now this this paragraph. And just like that, right, it has written up more text, and we can see all the new text that it has added in blue. In addition to that, what we can do is highlight the part of the text and change its tone. We can make it more casual or we can make it more formal, in a very similar manner to what I just explained with with the length. And we can also add some custom instructions, right? So, for example, without proper handling, these liens can significantly reduce the amount of money you actually receive from your settlement. Let's now suggest edits. And what it's going to do is it's going to actually come up with a few suggestions of how we can improve or potentially change the sentence, right? So in this case, for example, um, could be made more impactful for prospective clients worried about their recovery. Suggestion is significantly reduce to drastically cut the amount you actually take home from your settlement, right? So if we want to implement, as well, any of those changes, we can also just come here, select, and we're going to ask Gemini, make it more dramatic and include some more detail about about why the payouts to the client will be less. And so it introduces that text into the window. And just like that, it's going to make those necessary changes there. There you go. Without aggressive proper handling, these liens dodge suddenly can dramatically erode it, right? Definitely made it more dramatic and included some some more details. So, just like this, you can work with this canvas. Very useful feature, very cool for things such as drafting or creative content.
And another cool way that you can use Gemini is actually very similar to as well, for example, Chad GPT. You can generate images with it. We actually going to see how you can also generate videos at the later part of the video. But we can just open up a new chat and say, generate an image of an attorney walking into the office, right? Just like that. It's going to start generating the image. Takes a few seconds. And there we go. We have our our image. We can also refine it, right? So we can actually work with it continuously to match our style. So, for example, um, let's say like, replace those glass walls in the background with brick walls. Let's see how, how it handles that. And there we go. We have brick walls. The setting has has somewhat changed, right? So it's not, it's not necessarily perfect, like it's not exactly the same pose, right? But you can see how you can, uh, start working with this tool. Obviously, you know, great as well for any kind of, uh, content creation. Uh, you could even just use this in combination with your current method. So, for example, you know, generate a few assets, let's say with Gemini, and then place them in Canva or another, you know, editing software that you're using and combine them together, um, at, you know, the face of your actual attorney and whatnot, and use this tool more, let's say, collaboratively rather than just trying to generate the image, uh, one image from the get-go.
And now, let's actually talk about deep research. That's this other mode that we haven't yet explored in here. Um, what this is going to do is it's going to take a query and it's going to search a lot of different sources online, uh, for information, um, about, about this query, and then present its analysis, right? Using the more advanced, uh, complex reasoning model 2.5 Pro as kind of like a research report, which will be quite thorough, quite in-depth, right? That you can then go off from. So let's reuse actually one of the questions we had before about US visas and use now deep research to see how, what the output there would be. And so let's make sure that, you know, the deep research is enabled, and we're going to say, provide the full range of US visa options. And what's going to do, first, it's going to try to understand the question, and based on that, it's going to create kind of a research brief, right? For you of like, what exactly it's going to look for, right? And so in here, for example, we can see, research websites, access official US government websites, identify the list of primary categories of US visas, detail the various types of immigration visas, blah, blah, blah, right? Analyze the result and create a report. We can edit the plan if, if we don't like it for some reason, or we can click start research. And so what's going to do is then it's going to start with the research. This part can take a few minutes, can actually for more complex queries, can actually take up to 15, 20, 25 minutes, right? With very in-depth results. So let's wait for it to do its magic, and I'm going to show you the final result.
So after about 10 minutes, we got our research report, and as you can see, quite an in-depth, right, results. We have tables, we have all the main sources that used, right? A lot of detail. And in here, what we can do is we can export it to docs, we can copy it, we can actually generate an audio overview of it, and be able to actually listen to it, in addition to continuing to chat, right? So, for example, summarize, summarize the main, uh, options in a table. And obviously, what it's going to do now is instead of, you know, doing that from from the get-go, it will use the research report that already created and then, um, answer the question based, based on that. As with previously, we have seen with Google search, right? All of these things have the references, right? Where it collected the information from. And down below, we have the the list of all the sources that used. In addition, as well, kind of to all the sources and the process that it went through, right? So you can really see exactly what it is doing. Again, great for getting, let's say, lay of the lands or diving into more specific, um, specific, let's say, research question. Very cool feature. Anytime that I need to like research something in depth now, I first do this and then based on those results, um, already go and potentially do some more refinement of of of that and and and and dive in specifically deep into specific topics, um, to start off a research. This is, this is amazing.
Now, having looked at what deep research can do, let's now look at gems, or these custom AI experts. And so what are these? You have a few gems that are already predefined for some more generic tasks, such as brainstorming, career guidance, coding partner, outreach specialist, writing editor, etc. But you can also create your custom gems and your custom AI experts. How do we go about it? Let's click on here on job manager. And in here, we'll be able to create a new gem. I already have preset one, so let's just look at it. I'm going to click edit. And the setup is quite simple. The only things that you need to do is you're going to name it. You're going to provide some instructions to essentially provide some context about like what it needs to do, right? So in this case, for example, you're a professional M&A export. They will be helping the user owner of Dole Food Company. Analyze the proposed acquisition agreement and answer any questions about it using the copy in your knowledge. Be thorough and act professional. Always study your sources. Obviously, you can, this problem can be, as I mentioned, much more complex. In the knowledge, we are able to add files, right? So we can add up to 10 files in here of knowledge. It will be able to reference as we talk to it. And so previously, you know, we have done a similar task by creating a new chart, right? Explaining all the context and then uploading the file in there. With this, we'll be able to essentially reuse it, have multiple charts without having to kind of restate the context of the conversation and and reuploading those files. So that is very useful for things that you're doing on a repeated basis. You know, maybe here you can create a gem for creating your, let's say, LinkedIn content. Provides to it 10 examples of your previous posts, right? Explain what the model needs to do, and then based on that, right, we will be able to easily reference those past posts and provide outputs in a similar tone, right? In a similar language as you are using, without you having to constantly re-upload those files and provide those examples again and again in each one of the chats that you're going to do, uh, in here. We can play with it and kind of like do do a preview. Once it has been created, we can just click on it in here, under gems, and we'll open up a new chat. We'll be able to, as well, select whatever model that we want to be using with it, right? Um, so let's do it again. I also going to select the model and click into into the into into the gem, right? And so it's going to show any recent, uh, chats and, um, the name, so we, we know that you're actually using the gem. And let's, uh, let's do kind of a similar exercise we did before. List out all the red flags that you see, right? And again, in here, we don't no longer need to provide the file, right? We just send this query. It already has in it knowledge, the, uh, the file, right? So we'll be able to understand what we are talking about. And just like that, we have the response, similar to what we had before, and as well all the references in terms of the actual knowledge document that it used to come up with this answer. To hopefully see how this can be useful for again, tasks that you do on a more, let's say, mundane basis, repetitive workflows. The one limitation at this point in time that Gemini has is that you cannot really share these gems with your other Google Workspace members. So it will, they need to be defined on each user's account. Um, I, I do hope that Google will release that update and make it similar to what you're now going to do in Chad GPT and for example, cloud, where you can share those gems with with your, uh, with your workspace.
One last cool feature that I want to share, um, while we still the Gemini app, before going to the Google AI studio, which has some more advanced experimental features, is using, uh, apps, right? So these, uh, will potentially need to be enabled by, by your Google admin, uh, before you're able to use them. But what you're then going to be able to do is write an app symbol and then reference one of the other apps in the Google Workspace, such as Google Gander, Gmail, Google Drive, Workspace, right? And so let's now select Gmail. It will say, potentially, you know, you need additional permissions. And what we're then able to do is actually ask questions based on the contents, let's say, of your Gmail inbox, of your Google Drive, etc. And so let's see here, um, provide a summary of all the discussions I had with the founder of Marvy. Marber is another tool for M&A AI tool, and let's see how it works. So first, you need to connect to Google Workspace to turn on this app. So just click continue. You just need to connect the Google Workspace. Fantastic. And now it will source, it will look at your Gmail, and there we go. We have the summary. Conor uncle introduced Marvy and offered a demo. You requested recording of the demo. Uh, and then Connor sent an invitation for a meeting on April 29th. So, and, and you see in here as well, you can see any, let's say, sources where it, it, it took the information from, right? There we go. We got it from there. So, quite cool. Still now some feature that is still working on. But as you see, opens up, if you're using already using Gmail, you, Google, Google Drive, opens up quite a few possibilities of interacting with your data that you previously, uh, didn't, didn't have.
Let's now look at Google AI Studio, which is kind of a more advanced version of Gemini, which has some more experimental features, in addition to some more flexibility of how you actually interact with, uh, with Gemini. And so this can be accessed by going to a studio.google.com, again, with your Google workspace. Uh, this is covered and this is also covered under, under the heba. So let's get going. We have a few options in here. First of all, chat, which looks very similar to what we have seen in Gemini. But here now we have all of the settings on the right. And I won't go into too much detail into each one of those. But we have a bigger selection of the different families, let's say, of models that, uh, Google has released. We have the temperature, which will essentially influence how creative the model is. So the higher the temperature, the more likely it is to provide a a creative response versus the lower it is, the more, let's say, deterministic it's going to be and go for the safe option. In addition to that, we can do some more advanced stuff, providing structured output, execute code and and functions, and also grounding. We can enable or disable here grounding with Google search, right? So GX Gemini app by default does it grounds grounds all of its responses in Google search. In here, we can disable, enable this option, uh, in addition to some more advanced, uh, technical stuff. So it works exactly in the same way.
Apart from that, we have, uh, video gen, which is their new model V2 for video generation. So, for example, generate a video of an attorney walking into the office. There we go. It took about 30 seconds to generate this video. We can here on the right, we can also see how many results do we want to generate, select the model if there would be more models in the future, the aspect ratio, video duration, and then frame rate, resolution is is currently disabled, but then also negative prompt, if we don't see want to see a specific thing in the in the in the video, then we can also specify it in here. So, let's look at the video. And just like that, we have a video of an attorney walking into the office. And obviously, you can then refine it and and and work on it, right? You can, uh, refine your prompts and see what, what results you get. I won't go into the starter apps, as is just feature more so in development right now. But let's actually look at stream. And what stream allows you to do is kind of similar to, let's say, like the advanced voice mode of chat GPT, where you can talk in real time to the model. You can, as well, share your screen and and and show your webcam. And so again, someone more experimental at this point in time, but potentially very cool thing that you'll be able to use in the future. Uh, let's actually look at the use case. Is I'm going to, I'm going to start this conversation. I'm going to share my screen and then I'm going to ask it to for it to guide me in clear manage to solve a specific, uh, issue that I'm having. So in here on the right, we can see the output format, audio and text, or just text, which model we're going to be using this point in time that we just have Gemini flash, select the voice, the language, uh, media resolution that is uploading in there, and then also all those usual, let's say, tools and settings that we had previously in the chat window, uh, for this one, I'm going to enable grounding with Google search because most likely it's not going to have the right response for me based on just like general knowledge. It'll actually be need to look up the support documentation of Cleo Manage and guide me to the right solution. So I'm going to click share screen, entire screen, and it's now live. Fantastic.
So I'm an attorney working in GEO Manage is my practice management software. I'm a bit confused on how can I use a documents template feature. So if I go to the documents tab on a matter, click new, create document from template, I see that I can, you know, use a document template in here, generate a document, but I'm unsure where do I actually define and upload this document templates inside of the software. Can you please guide me in the software? Where can I do that? And also make sure to use, uh, Google search because probably you won't have this information in your knowledge.
Okay, I can guide you on how to define and upload document templates in Cleo Manage. Here's how you can add document templates. One, go to documents, then categories and templates, and then templates. Two, click new template. Three, choose the document file with the inserted merge fields. Four, name the template and optionally select how, how do, how can I know, uh, what merge fields can I use?
Okay, to know what merge fields you can use, follow these steps. One, go to settings, then documents, and then document template merge fields. Sorry, can you repeat? Can you repeat which, which, uh, which option should I click here on the screen?
To find the available merge fields, you will need to click on settings, which is on the left navigation bar. It looks like a gear icon. From there, you can navigate to documents and then document template merge fields, as I previously mentioned.
Okay. Documents. And let's see. Document merge fields. All right. And so, how does this work? Could you clarify what you would like guidance on?
How can I use this document merge field? It looks like you want to know how to use the document merge fields. First, you'll need to choose a matter from the select a matter drop-down menu. After you select a matter, the merge fields will display the information associated with that specific matter. You can then copy and paste these merge.
So, there we go. As you can see, quite, quite cool. Can be definitely somewhat clunky. Took me like two tries to to make this actually demo work. Definitely quite, quite exciting that potentially in six to 12 months, we'll be able to have just like this AI system that can look at our screen and and help us brainstorm and resolve any tech issues that we're having.
And let's now look as well at Notebook LM, which is a separate product from from the the two mentioned previously that is also leverage in Gemini. Important thing here, Notebook LM is not HIPA compliant. So don't use it for your PI or you know any other cases that you have that have that need. But what this allows you to do is to actually add many more sources of information in various formats. So that could actually include as well YouTube videos, that could include links, that could include obviously the files that you upload, and you can add up to 300 sources at this point in time if you are using Notebook Alam through through Google Workspace. And so you'll be able to upload all of your sources in here and then based on that, chat with those documents, also generate some, let's say, predefined, um, documents from it. It's just a briefing doc, FAQ, study guide, obviously now see as well, like all the sources it's actually using, right? And its responses, and even generate an audio overview of like almost like a podcast, uh, that explains that topic, right? So very nice for for doing some like more generic, more generic research. For specific legal research, I'm not sure that's still the best tool. Probably, you know, specialized software still is still the best, the best bet there, but definitely something quite cool. I'll just briefly show you like all the sources that you can add. Create a new notebook, and then as you can see, you can upload files, you can add Google Drive files, you can add websites and YouTube, and as well like post base text, up to 300 sources in here, and then after that, generate a deep dive conversation and generate any of these documents in here, and then obviously chat with it as well, uh, as you can Gemini.
And the great part about Gemini being integrated with Google Workspace, it actually also works outside of those apps that we previously mentioned, in your Gmail, in your Google Docs, in your Google Sheets. So, for example, in Gmail, you can, when you're composing a new email, there's this button over here, help me write, where you can actually explain what email you want written. It will take into context, if it's a thread of emails, any previous emails, and suggest a draft for you. Um, in addition to that, in Google Docs, as well, you can use a similar feature. In Google Sheets, you can actually use Gemini to help analyze data. So Google is really integrating into all of its different products more natively. So you don't even have to go to a separate app to execute some of those functions.
So I hope now that you have a good understanding of all the features that the suit of Gemini products provides, how you can actually start using it. And so I'm very excited to see what, what, how you're going to implement it in your practice, how you're going to use it. Uh, please let me know about your experience, um, either in the comments or actually join our free legal tech community, where we post videos like this. We have discussions with with in between each other about different tools that are coming into the market, about different use cases, uh, where you can also mingle with like-minded attorneys and law firm owners and managers. For that, check the link down below. I hope that this video was helpful for you. Thank you very much for watching and have a great day.