Unlocking the Power of Gemini 2.5's Deep Research for Business Intelligence
Overview
In this video, the speaker reveals how anyone familiar with Gemini 2.5's deep research function can perform tasks typically handled by high-paid consultants. The speaker shares a step-by-step process for maximizing the tool's potential in business intelligence and marketing strategy, emphasizing the importance of well-defined research projects.
Key Points
- Understanding Deep Research: Gemini 2.5's deep research can rival the work of top-tier consultants, making it a valuable asset for businesses. For more insights on leveraging AI in business, check out Unlocking Business Growth: Mastering AI Strategies for 2025.
- Defining Research Projects: A well-defined research project includes a clear problem statement, audience, scope, output format, and constraints.
- Using Meta Prompts: The speaker introduces a 'meta prompt' to guide Gemini in creating effective research prompts.
- Generating Reports: After inputting the meta prompt, Gemini generates comprehensive reports that can include tables and market analysis. For a deeper understanding of AI's role in report generation, see Exploring GPT-4.5: A Comprehensive Review of Its Strengths and Weaknesses.
- Reviewing Reports: The review process is crucial, and tools like Notebook LM can help visualize and summarize findings.
- Finding Experts: The speaker discusses using Manis AI to identify approachable experts for feedback on research findings. To learn more about effective data-driven decision-making, refer to Developing HR Analytics Units: Essential Steps for Effective Data-Driven Decision Making.
- Creating Visuals: Tools like Napkin AI and V0ero.dev are recommended for creating engaging visuals and micro-sites from reports.
- Consulting Market Insights: The video concludes with insights into the high fees charged by consulting firms, suggesting that AI could play a significant role in future consulting practices. For a look at the future of consulting with AI, check out The Future of Business: Leveraging Autonomous AI Agents.
FAQs
-
What is Gemini 2.5's deep research function?
Gemini 2.5's deep research function is an AI tool that helps users conduct in-depth research and generate comprehensive reports, rivaling traditional consulting work. -
How can I define a research project effectively?
A well-defined research project should include a clear problem statement, audience, scope, output format, and any constraints. -
What is a meta prompt?
A meta prompt is a prompt designed to help create other prompts, guiding the AI to generate effective research queries. -
How can I visualize my research findings?
Tools like Notebook LM and Napkin AI can help create visual representations of research data, making it easier to understand and present. -
How do I find experts for feedback on my research?
Manis AI can assist in identifying approachable experts in specific fields who can provide valuable feedback on your research findings. -
What are some tools for creating engaging reports?
Napkin AI and V0ero.dev are recommended for creating visuals and converting reports into interactive micro-sites. -
What are the typical fees charged by consulting firms?
Consulting fees can range significantly, with top firms charging anywhere from $250,000 to over $10 million for projects.
Do you realize that anybody who knows how to use Gemini 2.5's deep research function can now do the work of some of
the highest paid consultants in the world? Today, I want to show you exactly how to use Gemini's deep research for
business intelligence, marketing strategy, and more. I've spent many hours banging my head against Gemini to
come up with a simple step-by-step process for getting the most out of this powerful tool. You see, while I was
running my marketing agency serving leading software companies, I interacted a ton with these top tier consultants.
I've been shocked at how Gemini's deep research can rival or often exceed some of the work I've seen from these folks.
Did you know that scientists and economists are now getting papers published in academic journals that have
been written primarily with AI tools like deep research? The trick is to make sure that your research project is
incredibly well-defined. This means clarifying the problem, defining the audience and scope, outlining the
desired output format, and identifying any constraints. In fact, a good deep research prompt itself can be over one
page long. I got to thinking, why not let AI do that heavy lift for us? And for that, I'm hopping into the cheat
sheet. I make a cheat sheet like this for every single video that I create. There are now over 125 of these
instantly available to anybody who supports me on Patreon. There's a link in the description, so check that out.
The first prompt I want to show you from this is what I call a meta prompt. Meta has nothing to do with the company Meta.
This is a prompt for creating a prompt. So, we're going to grab this. We're going to copy and paste this into
Gemini. This prompt basically just asks Gemini to create a prompt for us that we can use inside of Deep Research that
hits all those boxes that I just mentioned, including clear problem statement, scope, output, format, etc.
The key here is that we're really looking to steer the AI into creating a McKenzie style report. So, let's let
that run. Wow, look at this. I had my best results using 2.5 flash and it created this prompt that goes on and on
and on. I would say this is many pages but very clear on what we are looking to get out of our deep research project. So
I'm going to go ahead and copy and paste that. I'm going to start a new chat string and I'm going to go into this 2.5
Pro. Going to turn deep research on. I'm going to drop this right into the prompt there. going to clean it up a little
bit. I don't think you need the title. That could confuse things. And we're going to let Deep Research do its thing.
It's probably going to ask us some clarifying questions. And it's funny because generating the report is
actually the easiest part of the process. Once you have that meta prompt, uh you can really just let this thing
rip. You don't even have to pay the exorbitant $200 a month for the chat GPT deep research because Gemini these days
is performing as well or better than that in many ways. It's given us a plan here. What's really cool is you can edit
this plan if you feel like it is off base and I think it looks pretty good. So, we're just going to start the
research. This is really exciting because I think we're now entering a phase where these tools are getting a
little trickier to use. It's not just as simple as putting things into chat GPT anymore. And when I shut down my
marketing agency and decided to pivot my whole career around AI, it was this particular inflection point that I've
been waiting for. I felt that work itself was going to change so drastically that I needed to be focused
on these tools as much as possible. And the whole goal of this channel is to help as many people as I can navigate
these changes. A quick side note, when it comes to Gemini's deep research, I found this tool incredibly helpful when
it comes to making any kind of buying decision. If you're looking to buy anything over a couple hundred dollars,
you might want to ask to do some deep research on that and then create an audio overview of its findings so you
can listen to that and decide, you know, what the best way to go is to spend your hard-earned cash. And here we have a
killer multi-page report that is pretty much exactly what I was looking for. And it goes on and on and on and includes
all these interesting tables with the total address of addressable market etc. You can export this to docs which is
probably a smart thing to do. You can create this audio uh overview which is pretty cool because these pieces of
information are so indepth that it's uh is kind of a challenge to review all of this. That's the step we're going to get
into next. But before we move on, I'm not sure how many of these top tier consultants you've worked with, but this
is the exact type of thing that they would turn in. And we're going to look at how much these folks charge for these
types of reports here in a second. But after working with a lot of these folks, I slowly began to despise them a little
bit because they would always create these reports and then hand them over to my agency team to try to make all of
their dreams come true, all these different strategies happen. So it felt like we were doing all the hard work,
the heavy lift uh and they were just sort of pontificating and creating these big uh strategic plans. So if generating
the report is the easiest part of this process, reviewing the report is probably the most difficult part of the
process. One really cool hack for this is to use Notebook LM. So I'm just going to copy and paste this into a fresh new
Notebook LM. You just go to notebooklm.google.com. I have many videos on this. So, we're going to paste
in this text here of this report. I got to give it a title because of my OCD. And here's where the new mind map
feature comes in really handy. Just clicking this little mind map with this one source uh selected. You can create a
mind map for whatever sources are selected. If you have multiple sources selected, it's going to create a mind
map that uh pulls in all those sources. Or you can go one by one and create mind maps just specifically around one source
at a time. You can also do a uh another type of audio overview here by doing this generate. I get the feeling this
one is going to be a little bit longer than the one that you would create directly in Gemini. Here's our mind map.
And I think this is a really cool way to get a feel for what's going on in the report. This could be the first step of
your review of this report because it visually highlights all of the major areas rather than having to scroll and
kind of dig through everything. This gives you insight into how those break down and you can keep clicking through
and when you get to the bottom of the uh mind map, this is pretty crazy. I haven't seen it go this many levels, but
you can click on any one of these and it will pull from that source all of the references related to whatever you
clicked on in the mind map. So, anytime you're processing a bunch of, you know, information and these days there's no
shortage of information to be processed, I like to pull it in and get sort of that visual overview of what's going on
in here. I did this really in a helpful way while I'm trying to learn some JavaScript programming. I used this
discover feature which is another great feature. Just asked it to pull in a bunch of different resources all related
to JavaScript which it did. And then I created a mind map around that that just gave me kind of the fundamentals of what
all of those resources um you know agreed upon. So can be a great way to you know get a handle on a big topic.
But again, if you're handing this over to a client, the review step is critical because as these AIs have gotten more
and more powerful, they've gotten better and better. It also makes it harder to spot when they're off. So, ideally, if
you're doing this on behalf of somebody, you have that expertise. You kind of know that industry that you're working
in uh to some degree. But if not, you can go out and find folks that may have that expertise. It's a lot easier to
hand an expert a report to review than ask them to go build that report for you. So, you can just shave down a lot
of their hours that way. And while we're on this topic, I've been pondering this Ray Dallio quote recently. He says, "One
of the most important decisions you can make is who you ask questions of. Make sure that they're fully informed and
believable. Find out who's responsible for whatever you're seeking to understand and then ask them. Listening
to uninformed people is worse than having no answers at all. So, in this world of AI, you know, getting other
people to vet what you're creating can be really a superpower. And the tool that I like the most for finding people
that can help me is this Manis AI. This thing is killer at list building, going out and finding contact info for people.
Here's another prompt that I want to share with you from the cheat sheet. It says, "I'm looking for experts in the
consumer healthc care space. I plan to launch a SAS app in this space. Can you help me find a list of experts,
including contact info who may be interested in helping me review my goto market strategy? Here's the key. It's
important that these folks are approachable. Very high-profile people are not applicable as I wouldn't be able
to reach them. So, I'm going to copy and paste this right into Manis here. Going to let this run. That has run and it has
pulled together all of these different files here. Let's open this one up. it's gone out and it has found a couple folks
that seem like they might be the type of folks that could help me. Um, in other instances, I've had it create entire
spreadsheets of people with their contact information, but I think it's more important to get uh the right
information rather than a lot of information. So, I really nudged it to help me find people that it felt like
were approachable. Okay, so we've got our report. we've vetted it with uh either our expert opinion or found other
experts out there. At this point, I went and asked the AI about, you know, what else do these uh consultants do? It's
really just creating these analysis and these reports, right? And it said something very funny. It says, "Yeah,
that's what they do, but they do it with flare, which I thought was really funny." So, let's see if we can add some
flare to this. So, back inside of Notebook LM, I've got this uh source selected. And now I'm going to grab this
prompt. Please recommend a few areas of this document that might lend themselves to visual representations. I'm copying
and pasting this right into Notebook LM. And it's highlighted a few areas that would be great for visuals. So the uh
market size, market trend analysis, competitive landscape. Let's click through to the citation and get the
exact data that it's talking about. So now we're going to go to a tool called Napkin AI. And if you've never used
this, it's really cool. Once we sign in, we're going to click new napkin and blank napkin. You don't want to use that
AI version cuz it's going to start to rehash whatever you put in here. We're just looking to copy and paste our
information in and not have it rewrite it. Seems like every tool now wants to rewrite whatever you put in there using
their AI. We don't want that. We want the blank napkin. And we're going to copy and paste in this information from
Notebook LM right here. Grabbing all that. copying and pasting that into napkin. These are two separate
visualizations. So, I'm going to just select this first one and click this little bolt tool here, this uh lightning
bolt. And it now gives us some different options for how we want to display this data, which can be really cool. This is
the one I think this makes uh the most sense. This type of a bar graph showing the total addressable market. Let's do
the same thing with the value. And here are market size estimates by region. I'm zoomed in a little close here, so it's
not fitting it all in, but this is a great way to create your graphs. There's a lot of different visualizations you
can do. This is a different example here, um, where I've gone through like a process flow,
uh, something like this, different funnels, obviously the bar graphs, etc. And you can get in here and edit these
all till your heart's delight. You can export them in ways that they you can make them uh editable in other pieces of
software. Really flexible. Great tool. Napkin AI. Not being paid to say that. I just love this tool. So from there, I
did a bunch of research into AI tools that might be able to help with the layout of the document. And
unfortunately, I was surprised to find that there really aren't too many uh AI tools out there that can help you, you
know, lay a document like this out with, you know, the different graphs and, you know, some nice visual treatments, you
know, cover page, that type of thing. But one thing that I did find that worked really well is this tool from
Versel. This is called v0ero.dev. And this tool is, I think, trying to be a little bit like a replet type of um AI
agent. So, I haven't used it for that. What it really excels at is making things look sweet. This prompt is going
to convert our report into a stunning micro site. Let's copy and paste this into this V0ero. And I'm going to put
the report in here in a second. But first, I like to separate that with these three little quotation marks so it
knows the prompt from the context. Here's our report. I'm grabbing that. copying pasting that whole thing in. And
the prompt just says, "Here's a report that I'm working on. Can you convert this into a highly engaging micro site?
It must include an interactive layout in different columns along with places where you can recommend pictures.
Important. Do not make any changes to the material and make sure that it adheres to some of the most modern and
impressive site designs you have access to. So, let's see what v0.dev comes up with here. I tried a
similar thing inside of Gemini and I found that this actually works much better. It's building out our micro
site. It's funny that it's easier to build a micro site now than lay out a document. I just love watching this
stuff code. It's so cool. Look at this. It's now building out multiple pages. That's something you can't do in Gemini,
but you can do in Firebase, which is uh if you wanted to stay entirely in the Google ecosystem, you might want to try
something like this in Firebase. Awesome. Look at how sweet this is. Let's open this in a new tab by clicking
this button. Now we can see this micro site. And as we've instructed it, it's given us some area here where we can add
the images. This isn't going to generate images. You can do that either in napkin or straight up in Gemini. Create some
interesting images. But this is just the layout and the creation of the micro site that you can then pass off to your
um client or your team or whoever it might be. But this has now given us a beautiful layout. I think this is
actually better than, you know, some sort of a PDF because people can click through. It has all these different
pages that show you um the different aspects of the market analysis. Really, really cool stuff here. So, while I was
making this video, I just had to dive in and figure out how much these people are actually charging, how much money is on
the table for projects like this. And this is pretty crazy. So when you look at McKenzie their projects are anywhere
from half a million to 5 million. Boston Consulting Group quarter of a million to 5 million. Bane and company f you know
half a million to 10 million. So pretty clear there is some money to be made here. You know topping out at that 10
million plus. I know that some people like to argue that hiring these high-end consultants is really just a way that
CEOs and other leaders can cover their ass when it comes to making these big decisions. This phenomenon is referred
to as decision laundering or credibility outsourcing. Some even call it CYA consulting where the CYA stands for
cover your ass. But I'd like to argue that AI is going to be playing that role in a bigger and bigger way as we move
forward. This cheat sheet is absolutely packed with a ton of good information, including everything we went through
here today with my process. But there's a separate section here related to starting your own consulting business.
This goes deep into finding your own niche and target client identification, figuring out your target client profile,
pricing strategies, how you might differentiate yourself, and a whole lot more. So, if this high-end consulting
business is something you think you might be interested in, definitely check out my Patreon, support this channel,
grab this cheat sheet and over 125 others that are instantly available. There's some coaching options in there
as well. There is a link in the description to all that. And if you haven't checked out my video about
becoming a oneperson startup, there's a link to that here. A lot of people have been loving that. I'll see you over
there. Make your dreams come true.
Heads up!
This summary and transcript were automatically generated using AI with the Free YouTube Transcript Summary Tool by LunaNotes.
Generate a summary for freeRelated Summaries

Unlocking Business Growth: Mastering AI Strategies for 2025
In this comprehensive video, discover how to effectively leverage AI, particularly ChatGPT, to enhance your business operations and marketing strategies. Learn about common pitfalls in AI content creation, the importance of authenticity, and actionable prompts to generate engaging content that resonates with your audience.

A Step-by-Step Roadmap to Mastering AI: From Beginner to Confident User
This video provides a comprehensive roadmap for anyone looking to start their AI journey, emphasizing the importance of understanding core concepts before diving into tools. It offers practical tips on building an AI learning system, developing critical thinking skills, and strategically selecting AI tools to enhance productivity.

Unlocking Creation: How to Make Videos with Nvidia AI for Beginners
Learn how to easily create captivating videos using Nvidia AI technology, from scripting to voiceovers, in just minutes.

Exploring GPT-4.5: A Comprehensive Review of Its Strengths and Weaknesses
In this video, we dive deep into the capabilities of GPT-4.5, highlighting its impressive creative writing skills while addressing its shortcomings in coding. We also discuss its pricing strategy and how it compares to other models like Claude and Gemini 2.0.

Unlocking the GPT Store: A Beginner's Guide to Creating AI Agents and Making Money
In this comprehensive guide, Liam Otley introduces the newly launched GPT Store, drawing parallels to the early days of the App Store. He shares essential skills and strategies for creating valuable GPTs, emphasizing the importance of unique offerings and effective marketing to stand out in a competitive landscape.
Most Viewed Summaries

Mastering Inpainting with Stable Diffusion: Fix Mistakes and Enhance Your Images
Learn to fix mistakes and enhance images with Stable Diffusion's inpainting features effectively.

A Comprehensive Guide to Using Stable Diffusion Forge UI
Explore the Stable Diffusion Forge UI, customizable settings, models, and more to enhance your image generation experience.

How to Use ChatGPT to Summarize YouTube Videos Efficiently
Learn how to summarize YouTube videos with ChatGPT in just a few simple steps.

Ultimate Guide to Installing Forge UI and Flowing with Flux Models
Learn how to install Forge UI and explore various Flux models efficiently in this detailed guide.

How to Install and Configure Forge: A New Stable Diffusion Web UI
Learn to install and configure the new Forge web UI for Stable Diffusion, with tips on models and settings.