NEW Google 2.5 Deep Research Agents Update (FREE!) 🤯

Gemini and Google have just released a new AI agent for Deep Research. You can see the announcement just came out literally a few hours ago today. They've essentially said that now Deep Research is powered by Gemini 2. 5 Pro, the most intelligent AI model. Now, I'm going to

show you two different ways to build deep research agents. One of them for free and one of them. So, let's start with this method, which is using 2. 5 Pro Experimental. You can actually go directly onto Gemini as you can see right here. This is with the advanced version. So this is a paid model. We'll come on to the free model in a second. And then we can switch between all these different methods right here. Gemini is actually super powerful for just coding in general. But if you want to use deep research then you would select this from the dropdown and then you can say okay research for example I don't know the latest ID with all men code plus how it's different versus other IDs. It's going to be powered now by Gemini 2. 5 Pro which means that it's just got a better API. It's a more intelligent AI model and it's better at finding and synthesizing information providing more useful and insightful reports and analytical reasoning. Now, you can also do this on your Android device as well. So, you don't need to use this directly inside desktop. For me personally, I prefer to use desktop just cuz it's easier to get all the information right here. And then you can see now this is going off and doing its magic. And what you can actually do here is you can edit the plan before you post it. So you can edit what it's going to research and create before you actually post it if you want to directly. And if you just want it to go off and do its magic, you can say, "Okay, good to go. " Once it's done that, this usually takes I would say 5 to 10 minutes to generate from when I've tested this in the past. With the new upgrade, I'm sure it's going to be pretty much the same timings, maybe even slightly faster. The other thing that I would say here is if you're using something like chat GBT. I don't use chat GPT for deep research modes anymore. So if we go on to chat GPT, we can select deep research, but it just takes so long and it generates so many words that it almost defeats the object unless I need to go like crazy in depth like I'm launching a new product or I need to analyze a market or analyze my competitors etc. Right? And then you can see the sources that it's found right here. So here's an example and it's got all the sources. So actually finds a lot of different resources when it's researching the websites. You can actually see the thinking or you can see the sites that are actually browsed. And basically this AI agent is just going off searching the web doing the research for us and then it will generate the report once that's done. Actually insane how many websites it's researching. So it's researching 74 different websites to go and get the answers that we need which is pretty awesome to be fair. Now when we have a look at this as well, you can see this actually trending on Twitter right now. We can actually see the comparative between Gemini and OpenAI. So obviously Gemini is Google's AI agent and OpenAI's chat GBTS. And in terms of deep research, it's actually outperforming on many benchmarks here. So you can see for example for instruction following, comprehensiveness, completeness, and writing quality, deep research with Gemini, which is in blue, is outperforming on most of the benchmarks. This is absolutely crushing it right now. So, let's have a look what we've got back here. It's still doing the research. You can see the progress bar on the right hand side over here. And if

you want to do this for free, build your own AI research agents for free, there's a couple of different ways you can do it, right? So, one of the easiest ways is you can actually go onto something like shift and you can build your own AI agents with deep research. All right. So, let me just delete this previous pipeline. And we're going to just make sure that we've deleted the other chat bots that we've deployed in the past. And you can build one free chatbot or AI agent with vector shift using Gemini 2. 5 as well. All right. So, there's a bunch of different templates like you can see right here. Or we can just click on create pipeline. And inside here, we would just insert an input. All right. So, this is basically going to be the chat input. And then from here, we're going to select an LLM. You can choose between all these different LLMs, but if you want to use Google Gemini 2. 5 Pro for free as the LLM, you can insert that right here. And then from the dropown, we're just going to select 2. 5 Pro Experimental. Now, what you can also do

if you want to bypass the limits is you can go to aistic studio. google. com google. com and using AI Studio, we can start grabbing an API key, which again is completely free, create an API key here, and you'll get better limits if you plug that directly into vector shift. All right, so you can do it that way, but we're just going to stick with the defaults for now just to make it easier. And then in terms of the prompts, what we're actually going to do is we're going to type in the squiggly brackets like you can see up here. That is done with shift and then the square bracket on your keyboard if you want to select that. And then just make sure you select the input t text as the prompt. Right? So that's going to be the prompt that gives you the outputs. We're going to unselect use personal API key. So deselect that and just go with the default. And then from here we can also link this to web searches and that sort of thing, right? So let me show you an example of how to do that. So if you go to data loaders over here, we'll plug in the web search and then you can choose whether you want to use for example like Google search, xai or you. com. So we can click for example Google search right here and then you can just type in your query right. So what you can do is actually link this to your input. So we can take the response from that or we can just grab the keyword from the inputs over here. All right. So if we just select the input and the text now that's linked to our AI agent like you can see. All right. And then the output fields. So the URLs and the snippets we can actually plug directly into this. And then you just want to test this stuff as well when you're using it. But we'll come on to that in a second. All right. So we're going to delete that for the prompt. We can say okay make a you are a deep research assistant create a research report based on the Google search results from and then we'll just tag in the Google search over here. So you got the URLs and snippets. So we'll tag both create a beautifully formatted research report. based on the Google search results from this output. All right. So, we've got the outputs over here. And then for the actual responses, you can choose. So, you can have this just go straight to a Google doc if you want. Again, this is using the Gemini 2. 5 Pro experimental research. And the other thing here as well is like with the prompt you could also for example you could say to it all right with this prompt you can ask chat cheap or these other AI agents to create a prompt for deep research as well. So if you want to refine the prompt and just make something more detailed when it comes to deep research you can do it right here. But you can see how we're taking the input then we're running that through Google search and then we're going to use that to inform our AI agent on exactly how to perform. So, we're going to say, okay, create a deep research report based on and we'll take the Google search URLs and responses plus the input instructions from and then we'll take the input instructions right here. Okay. Now, if that doesn't work, there's a chance that it doesn't work. So, we can always take the output and test it in a second. We'll see how that goes. And then on the general section, you can choose how you want to output this, whether you want to save it, where you want to share it, etc. But just to keep

things really simple for the sake of a YouTube video, we're just going to have the output over here. For the output, we'll say, okay, output the information from the Google AI agent. So the response here, all right, and you see how this has taken the input plus the Google search, plugged it into the deep research agent, and then it's going to output it over here, right? So we're going to hit deploy changes. All right, we've deployed that now. And then what we can do is just go over to run chatbot and you can just test this out. Okay, so we could say, for example, do a deep research report on Gemini 2. 5 and we'll wait for that to come back to us. As you can see, that's now processing the request. Probably the thing that's going to take the longest is the actual Google search over here. In the meantime, whilst we're waiting for that to load, let's come back to Gemini and we'll see where we're up to with the deep research report. So, you can see this is still going off and analyzing the SERs. But it's basically researched all the websites, defined what augment code is, come up with the common features, then it's figured out the next steps, and it's researched a bunch more websites down here. After that is figured out the AI feature insights, traditional ID capabilities and you can see just each step it's breaking it down and researching a crazy amount of websites. That's way more than I've seen for example sites like chat GPT research agents do but you can see the outputs right here. So it's reaching all it's researching all the sites coming up the details and that is only halfway through. So it's probably going to take I would say maybe this is going to be a lot slower than Gemini 2. 0. Looks like it's probably going to take about I would say 20 to 30 minutes to generate that on the deep research side. Seems a lot longer than usual. All right. And you can see it's researched 143 websites. Now, if you got a human to do that, imagine how many dozens of hours this would take to do, right? For a human to figure out all the sites to research, then go through each one individually, make sure that the human doesn't get tired or stops working on the weekends and that sort of thing. Like the efficiency with this, the fact that you can just go off and do all this stuff is just so much faster than any human. I honestly would never hire a human to do research anymore because you can just get something like Gemini Advance to do this magic for you. Now, if you want something quicker and cheaper, this is completely free like I was saying. And you can see here that it's done the deep research report. So, it's come up with the executive summary, the introduction, the scope, etc. And we're good to go on that along with all the references. Now, it doesn't reference everything, but it does go off and do its magic. So, if we, for example, check out this site, it's researched directly from the web, found this, and then used all these sources. Again, this is not as comprehensive as Gemini Advance, but Gemini Advance is going to take a lot longer, plus you have to pay for it. Whereas, if you want a free chatbot that you can deploy and use as a research agent, then you can just use this method right here. And you can see how simple and easy it is to do. You just plug in the input, do the Google search, program the agent to use its instructions to go up and take the research from Google and then output it onto a chatbot like you can see. All right, you can even do voice searches here which is pretty insane. And also you got like your standard inputs outputs over here. All right, pretty cool. And then if you actually want to export this as a chatbot and deploy it, you can click on export chatbot, type in the chatbot name. So for example, deep research agent with 2. 5 create chatbot. You can change all the branding and stuff right here. So we could go into the header name, type in deep research agent. You can add your own image here, add some messages, etc. And then once you're ready, hit deploy changes, export. We have the chatbot right here if you want to test it out. And you've got the deep research agent ready to deploy over here. and it can do Google searches and connect to the web and that sort of thing. Pretty cool. And that was incredibly easy to do. You can also embed this to your website. So, we can take the embedible code like you can see and send that over anywhere else. And also, you can connect this to Slack, which is pretty crazy. So, you can actually deploy this on Slack, have a Slack channel where this actually does the research for you and you just connect and deploy to Slack like you can see. So you can have a D research agent completely embedded for free inside Slack and it didn't cost you anything. All right. Whereas for example, Gemini Advance is awesome, but it's going to take a lot longer, especially with the 2. 5 Pro. I also think because this has just been recently deployed, you know, literally this update, it just came out like a few hours ago. I think it's going to take a lot longer than usual. usually within the first 24 hours or 48 hours a new update comes out with AI. It's typically a lot slower than you expect. All right, so thanks so much for

watching. If you want to get access to more workflows like this for Gemini, you can actually get that inside the AI profit boardroom and we have 31 different lessons on Gemini and Google Gemini, how to use it, autonomous agents, building anything with Gemini 2. 5 Pro, how to train new employees with an AI voice agent, free deep research agents, etc. If you want to learn all this stuff directly yourself with tons of SPs, then you can get that inside the AI profit boardroom. And this also comes with a community focus on making more money and saving time with AI. Additionally, you get weekly Q& A calls. There's actually one today. So, if you sign up today, you can join the Q& A call later today. And additionally, inside the community, if you ever get stuck, you can ask questions. And you can see here, for example, everyone gets their questions answered directly here. So, for example, like you see, everyone's commenting and helping each other out inside the community directly. All right? And then we're constantly updating this. So, for example, the new update that came out today, I'll add inside the SAP section and we'll add a new section for Gemini research agents plus an SAP on exactly how to use it. All right. Additionally, if you want to get a free one-to-one SEO strategy session that shows you how to get more leads, traffic, and sales with SEO, feel free to book that in. We'll show you how we take websites from zero to 145,000 business month and generate hundreds of thousands of dollars in sales on autopilot on this free link building acceleration session. You're going to get a free custom tailored game plan to generate more leads, sales, and profits from your website directly. All right, and this is based on how we've helped hundreds if not thousands of clients at this point. You can see all the testimonials right here. A lot of happy clients right there. And if you want to learn and grow your website with more free traffic, more sales, more customers, then feel free to book in an SEO strategy session. Appreciate you watching. Thank you very much. Bye-bye.