5 Mind Blowing Use Cases of Deep Research

00:22:13
https://www.youtube.com/watch?v=ZzO2GMc-G3M

摘要

TLDRThe video explores the concept of 'Deep Research' in AI, comparing various models like ChatGPT, Gemini, and Perplexity. The speaker shares insights from extensive testing, revealing the strengths and weaknesses of each model. He introduces Jenspark, a lesser-known tool he believes is superior for deep research. Practical use cases for businesses are discussed, emphasizing the value of deep research for entrepreneurs and freelancers. A detailed prompting framework is provided to enhance research outcomes, making it easier to obtain in-depth and actionable insights.

心得

  • 🔍 Deep Research acts as a personal research assistant.
  • ⚡ Perplexity is the fastest for quick insights.
  • 📊 Gemini excels in methodical research.
  • 📈 ChatGPT offers depth but at a higher cost.
  • 💡 Jenspark is recommended as the best tool for deep research.
  • 📝 A structured prompting framework enhances research outcomes.
  • ⏳ Deep Research can save significant time in business research.
  • 📉 Users should verify the reliability of AI-generated information.
  • 🤝 Deep Research is valuable for freelancers and business owners.
  • 📚 Unconventional applications include client research and market analysis.

时间轴

  • 00:00:00 - 00:05:00

    The video discusses the limitations of current AI research tools like ChatGPT and Gemini, which often provide shallow or outdated answers. The speaker introduces a new deep research model that combines various AI models for better results. After spending over 50 hours testing these tools, the speaker will share insights on their effectiveness and provide a prompting formula for better outcomes.

  • 00:05:00 - 00:10:00

    The speaker explains the concept of deep research as a personal research assistant that actively searches the web and synthesizes information. They compare the deep research capabilities of different platforms: Perplexity is fast but lacks depth; Gemini is methodical and reliable but slower; ChatGPT offers comprehensive research but is the most expensive. The speaker emphasizes the importance of fact-checking regardless of the platform used.

  • 00:10:00 - 00:15:00

    A side-by-side comparison of the same prompt across different deep research models reveals significant differences in output quality and depth. The speaker highlights the strengths of each platform, noting that while Perplexity provides quick answers, Gemini offers more comprehensive reports, and ChatGPT excels in detailed analysis. The speaker also introduces a fourth model, Jenspark, which combines the strengths of various AI models for optimal results.

  • 00:15:00 - 00:22:13

    The speaker outlines five unconventional use cases for deep research in business, including researching new tools, client research for proposals, target audience analysis, regulatory compliance, and market opportunity analysis. They emphasize the time-saving benefits of deep research and provide a prompting framework to enhance the effectiveness of research queries, ultimately helping businesses make informed decisions.

显示更多

思维导图

视频问答

  • What is Deep Research in AI?

    Deep Research is a feature in AI models that allows for more thorough and detailed research, acting like a personal research assistant.

  • Which AI models offer Deep Research capabilities?

    Models like ChatGPT, Gemini, and Perplexity offer Deep Research capabilities.

  • What is the best tool for Deep Research according to the speaker?

    The speaker recommends Jenspark as the best tool for Deep Research.

  • How can Deep Research benefit businesses?

    Deep Research can save time and provide in-depth insights for decision-making, client proposals, and market analysis.

  • What is the cost of using ChatGPT for Deep Research?

    ChatGPT's Pro Plan for Deep Research costs $200 per month.

  • What is the prompting framework for effective Deep Research?

    The framework includes defining the scope, specifying output format, directing reasoning through conflicting information, and explaining how the information will be used.

  • What are some unconventional applications of Deep Research?

    Unconventional applications include client research, target audience analysis, regulatory compliance, and market opportunity analysis.

  • How long does Deep Research typically take?

    The time varies by model; for example, Perplexity can complete research in under 3 minutes, while ChatGPT may take 5 to 30 minutes.

  • What should users be cautious about when using Deep Research?

    Users should verify the reliability of sources and be aware that AI-generated information may not always be accurate.

  • Who is Deep Research primarily for?

    It is primarily for professionals in knowledge-intensive fields, business owners, and freelancers needing quick, reliable research.

查看更多视频摘要

即时访问由人工智能支持的免费 YouTube 视频摘要!
字幕
en
自动滚动:
  • 00:00:00
    you can already do research with AI you
  • 00:00:02
    can ask Chachi PT or Gemini a question
  • 00:00:05
    and they'll give you an answer right but
  • 00:00:07
    here's the problem that answer is often
  • 00:00:10
    shallow outdated maybe or just plain
  • 00:00:14
    wrong sometimes and lately as you very
  • 00:00:16
    likely know these AI companies have been
  • 00:00:18
    hyping a Mode called Deep research
  • 00:00:20
    perplexity Gemini chat gbt all have this
  • 00:00:24
    mode all aptly named the exact same
  • 00:00:26
    thing across these models and by the way
  • 00:00:29
    later in this video I'm going to share
  • 00:00:31
    with you a free deep research model that
  • 00:00:33
    I think is better than all three of
  • 00:00:36
    those well-known ones it mixes together
  • 00:00:39
    all the different AI models agentically
  • 00:00:42
    within the research that it's doing and
  • 00:00:44
    you've likely never heard of this tool
  • 00:00:46
    and I love it so is deep research is
  • 00:00:48
    another overhyped AI feature does it
  • 00:00:50
    actually work well I spent over 50 hours
  • 00:00:54
    trying out this feature across all the
  • 00:00:56
    models that I've just talked about and
  • 00:00:58
    I've seen what works and what I think is
  • 00:01:01
    a waste of time and the results well
  • 00:01:03
    they kind of surprise me and in this
  • 00:01:05
    video I'm going to give you a clear
  • 00:01:06
    breakdown of each platform's
  • 00:01:08
    capabilities without all the marketing
  • 00:01:10
    hype I'll show you real examples of how
  • 00:01:12
    I've used deep research to solve actual
  • 00:01:15
    business problems in my business and
  • 00:01:17
    you'll learn exactly when deep research
  • 00:01:19
    is worth the time and money and when it
  • 00:01:21
    isn't and by the way these use cases are
  • 00:01:23
    going to be different from what you may
  • 00:01:25
    have seen already and by the end you're
  • 00:01:26
    going to have my step-by-step prompting
  • 00:01:28
    formula that gets results with the best
  • 00:01:30
    deep research models in minutes not
  • 00:01:33
    hours by the way if you're new here my
  • 00:01:35
    name is Rick M and for the past 11 years
  • 00:01:38
    I've run a multi- figure online business
  • 00:01:40
    and today I run the AI Playbook
  • 00:01:43
    community over on school and I'll link
  • 00:01:45
    to it in the description below it's for
  • 00:01:47
    online business owners who're looking to
  • 00:01:48
    leverage AI in your business to
  • 00:01:50
    streamline and increase profit now
  • 00:01:52
    before we dive in let me show you
  • 00:01:53
    something super interesting here's a
  • 00:01:55
    standard question that I just put into
  • 00:01:57
    chbt 40 about analyzing crms for online
  • 00:02:01
    businesses and it gave me a really
  • 00:02:04
    straightforward answer as you can see
  • 00:02:06
    here it just simply lists them out with
  • 00:02:08
    a few bullet points for each one of
  • 00:02:10
    these and so it's given me a list of six
  • 00:02:13
    and then a few different recommendations
  • 00:02:15
    for me uh based on my needs now check
  • 00:02:19
    this out using deep research also in
  • 00:02:21
    chbt here's the exact same question
  • 00:02:24
    using my deep research prompting
  • 00:02:27
    strategy as you can see here the first
  • 00:02:28
    thing that it did was it it asked me
  • 00:02:30
    follow-up questions to clarify some
  • 00:02:34
    different things to make sure that it's
  • 00:02:35
    giving me the best response possible now
  • 00:02:38
    here is the executive summary
  • 00:02:41
    comparative analysis of CRM options with
  • 00:02:44
    a nice table as you can see here going
  • 00:02:47
    all the way through and breaking down
  • 00:02:50
    each of these tools really really
  • 00:02:53
    indepth and as you can see here the
  • 00:02:54
    difference is massive one gives you
  • 00:02:56
    pretty general advice that you could
  • 00:02:58
    find pretty much anywhere the other one
  • 00:03:00
    deep research delivers very specific
  • 00:03:03
    in-depth actionable intelligence with
  • 00:03:06
    current data and citations and that's
  • 00:03:08
    the power of proper deep research
  • 00:03:10
    prompting and I'm going to show you
  • 00:03:12
    exactly how to do that so the first
  • 00:03:13
    thing you got to understand is what
  • 00:03:15
    exactly is deep research well
  • 00:03:17
    essentially it's a personal research
  • 00:03:19
    assistant working for you whenever you
  • 00:03:22
    want it unlike you know your standard
  • 00:03:25
    going into any of these AI models
  • 00:03:27
    limited to their training data deep
  • 00:03:30
    research is going to actively search the
  • 00:03:32
    web synthesize the information and
  • 00:03:34
    generate detailed reports for you kind
  • 00:03:37
    of like I just showed you and it's
  • 00:03:38
    different from standard web search
  • 00:03:40
    because it's going to mimic a especially
  • 00:03:42
    in chat gbt human research workflows
  • 00:03:46
    through a three-step process like number
  • 00:03:48
    one collecting sources from across the
  • 00:03:50
    web and any kind of documents that you
  • 00:03:52
    upload into if you're using deep
  • 00:03:54
    research within Chachi PT number two
  • 00:03:57
    reasoning through that information and
  • 00:03:59
    identifying patterns resolving
  • 00:04:01
    conflicting information and also drawing
  • 00:04:04
    conclusions for you and then number
  • 00:04:06
    three outputting detailed reports and
  • 00:04:07
    information from all that research and I
  • 00:04:10
    think it's really important super
  • 00:04:12
    important to understand how each
  • 00:04:13
    platform approaches this in terms of
  • 00:04:15
    deep research so perplexity uh Gemini
  • 00:04:18
    which was the first one and then Chachi
  • 00:04:21
    BT so let's talk about perplexity deep
  • 00:04:22
    research first it's going to be the
  • 00:04:24
    fastest one it's going to complete
  • 00:04:25
    research in under like 3 minutes which
  • 00:04:27
    is way faster than any of the other
  • 00:04:29
    models it's going to search the web it's
  • 00:04:32
    going to extract key facts for you and
  • 00:04:34
    then it's going to create clear reports
  • 00:04:36
    for you with all the citations now in
  • 00:04:39
    terms of strengths it gives you really
  • 00:04:41
    quick answers it does a pretty wide
  • 00:04:44
    range of sources although I kind of go
  • 00:04:47
    back and forth on that and it does
  • 00:04:49
    number the sources for easy verification
  • 00:04:51
    of the information that it's giving you
  • 00:04:54
    and the other thing if you're having a
  • 00:04:55
    really long conversation in perplexity
  • 00:04:57
    deep research it can have sort of poor
  • 00:05:00
    contextual remembering the context out
  • 00:05:03
    of the entire conversation if it's a
  • 00:05:05
    longer conversation and perplexities
  • 00:05:07
    deep research is by far my least
  • 00:05:09
    favorite of these deep research modes
  • 00:05:11
    that we've got available to us and the
  • 00:05:13
    next one Gemini deep research which was
  • 00:05:15
    the first Model to come out with the
  • 00:05:16
    Deep research mode it takes a very
  • 00:05:18
    methodical approach with the research
  • 00:05:20
    planning it will take your question it
  • 00:05:23
    will develop a plan for you which you
  • 00:05:25
    can edit accordingly if you want the
  • 00:05:27
    process is going to be longer than
  • 00:05:28
    perplexity it's going to typically take
  • 00:05:29
    like 5 to 10 minutes it excels at
  • 00:05:31
    understanding complex topics it's going
  • 00:05:33
    to use establish more reputable sources
  • 00:05:36
    uh it's going to provide trustworthy
  • 00:05:37
    information as always we need to be
  • 00:05:40
    making sure that we're double-checking
  • 00:05:41
    our facts and stuff like that but it may
  • 00:05:43
    lack diverse perspectives when you're
  • 00:05:46
    doing a deep research within Gemini and
  • 00:05:48
    it integrates really well with Google
  • 00:05:49
    drive because you can export the
  • 00:05:52
    research report right into a Google doc
  • 00:05:55
    if you want to then there's Chachi PT's
  • 00:05:56
    deep research obviously this is going to
  • 00:05:58
    give you the most comprehensive research
  • 00:06:01
    capabilities in my opinion it's going to
  • 00:06:03
    have the slowest processing time
  • 00:06:05
    anywhere between say 5 and 30 minutes
  • 00:06:07
    the ones that I've run and through all
  • 00:06:09
    these tests that I've been doing
  • 00:06:11
    generally in that 15 to 25 30 minute
  • 00:06:13
    range it is capable of analyzing files
  • 00:06:15
    and images it excels at Deep complex
  • 00:06:18
    topics it can produce extremely detailed
  • 00:06:21
    reports for you it will struggle
  • 00:06:23
    sometimes and this is obviously a big
  • 00:06:25
    thing when it comes to research and that
  • 00:06:26
    is struggling with Source reliability
  • 00:06:28
    but again as always regardless of what
  • 00:06:30
    platform that you're using make sure
  • 00:06:32
    that you're factchecking with whatever
  • 00:06:33
    results that you get now this is also
  • 00:06:35
    the most expensive yes you can access
  • 00:06:37
    deep research with Chachi BT it's 200
  • 00:06:40
    bucks a month for the Pro Plan if you're
  • 00:06:42
    going to be using this a lot now you do
  • 00:06:43
    get access to deep research on the plus
  • 00:06:47
    plan which is the 20 bucks a month but
  • 00:06:49
    you are very limited to the number of
  • 00:06:51
    researches that you can do during the
  • 00:06:52
    month now let me show you a side-by-side
  • 00:06:54
    comparison of the exact same prompt
  • 00:06:57
    across all of these deep research models
  • 00:06:59
    and also I will show you the fourth
  • 00:07:02
    model that I think is the best one in my
  • 00:07:05
    opinion and it's also free to use okay
  • 00:07:08
    so here we are in chbt this is the
  • 00:07:10
    prompt that I will take you through in a
  • 00:07:12
    little bit this is the one that I've
  • 00:07:13
    come up with that I think works the best
  • 00:07:15
    for these deep research modes and I'm
  • 00:07:19
    asking about again as I mentioned
  • 00:07:21
    earlier CRM opportunities for online
  • 00:07:25
    businesses and as you can see here it
  • 00:07:27
    did it in 6 minutes and 20 three sources
  • 00:07:30
    which I don't think is very much yes it
  • 00:07:33
    does include the citations within the
  • 00:07:36
    content but it does give me a nice
  • 00:07:38
    executive summary uh a nice table here
  • 00:07:41
    of each of these tools and then it goes
  • 00:07:43
    through each of these tools and gives
  • 00:07:44
    summaries for each one which I think is
  • 00:07:46
    is really helpful it is a lot of content
  • 00:07:49
    here and then going down to the very
  • 00:07:52
    bottom I also ask it for scalability
  • 00:07:55
    options so it gives me a cost breakdown
  • 00:07:58
    which I really like
  • 00:08:00
    and I did ask for that though in my uh
  • 00:08:02
    in my prompt and at the bottom here it
  • 00:08:04
    gives me the executive insights and
  • 00:08:06
    Trends and CRM Solutions and then with a
  • 00:08:08
    conclusion at the bottom now here's
  • 00:08:10
    perplexity output for the exact same
  • 00:08:13
    prompt that I put in as you can see here
  • 00:08:15
    it did search 43 sources which is way
  • 00:08:18
    more than what chat gpt's version did
  • 00:08:21
    this is the entire report right here and
  • 00:08:26
    it is very brief it does give me I like
  • 00:08:29
    the table a whole lot better than what I
  • 00:08:32
    saw in chbt but as you can see here it
  • 00:08:34
    is super short it's nowhere near in
  • 00:08:37
    depth as these other deep research tools
  • 00:08:39
    and just like in the other models you
  • 00:08:40
    can ask followup questions of the
  • 00:08:42
    research and just like in perplexity
  • 00:08:44
    normally it gives you suggested
  • 00:08:45
    questions that you can follow up with
  • 00:08:47
    that is pretty cool and now we're
  • 00:08:48
    looking at Gemini's deep research on the
  • 00:08:50
    Le hand side here this is the actual
  • 00:08:52
    report that it created it searched
  • 00:08:57
    414 sites as you can see see here down
  • 00:09:00
    here now one thing I've noticed with
  • 00:09:01
    Gemini's deep research is the number of
  • 00:09:03
    websites that it is using as sources in
  • 00:09:05
    its researches has gotten way way bigger
  • 00:09:09
    so here's the report that I did come up
  • 00:09:10
    with here and also you can export it to
  • 00:09:13
    docs as I just mentioned here and so
  • 00:09:15
    this is the report over in Google Docs I
  • 00:09:17
    just exported over there it's 14 pages
  • 00:09:19
    long and as you can see here it's pretty
  • 00:09:21
    comprehensive and I like how it it just
  • 00:09:24
    broke it up really nicely and then
  • 00:09:26
    finally this is Jens spark. a this is my
  • 00:09:29
    favorite deep research platform to be
  • 00:09:31
    using I just like the output and some of
  • 00:09:34
    the options that it gives you within the
  • 00:09:37
    Deep research output so like for example
  • 00:09:39
    last week I got this really cool mind
  • 00:09:41
    map as part of the output something that
  • 00:09:44
    I've not seen ever in any of the other
  • 00:09:47
    deep research models so as you can see
  • 00:09:49
    here it went through 213 sources you can
  • 00:09:52
    ask questions to it right on the
  • 00:09:53
    follow-up and like the other models you
  • 00:09:55
    can ask questions about the research and
  • 00:09:57
    what I really like too is the T of
  • 00:09:59
    contents that it came up with now check
  • 00:10:02
    this out also at the bottom it gives you
  • 00:10:04
    a really nice summary also gives you a
  • 00:10:06
    really detailed summary gives you a
  • 00:10:07
    section called the takeaway and then it
  • 00:10:09
    also gives you supplementary video
  • 00:10:11
    resources yes perplexity will do that if
  • 00:10:14
    applicable I just like how this report
  • 00:10:16
    comes out and at the very bottom it
  • 00:10:18
    gives you this similar knowledge section
  • 00:10:19
    which gives you access to similar
  • 00:10:21
    researches so I really like Jens spark
  • 00:10:24
    another big thing why I love deep
  • 00:10:26
    research in Jens spark is it does this
  • 00:10:28
    thing called mixture of Agents where it
  • 00:10:30
    says it auto mixes the best AI models
  • 00:10:33
    for your specific task so as you can see
  • 00:10:35
    here it lists out all the the main
  • 00:10:37
    models and it will use the models that
  • 00:10:39
    it thinks is best for your specific
  • 00:10:42
    research question so as you can see here
  • 00:10:43
    each platform has its distinct strengths
  • 00:10:45
    and weaknesses when it comes to this
  • 00:10:47
    deep research mode depending on what it
  • 00:10:49
    is that you want to be researching now
  • 00:10:51
    let's talk about how businesses can
  • 00:10:52
    actually use deep research and no I'm
  • 00:10:54
    not going to give you the standard use
  • 00:10:56
    cases that you've heard a million times
  • 00:10:58
    with deep research here are five
  • 00:11:00
    unconventional applications that I found
  • 00:11:02
    incredibly valuable with deep research
  • 00:11:04
    and the first use case I've actually
  • 00:11:05
    already shown you that is researching a
  • 00:11:07
    new tool to use for your business
  • 00:11:09
    whether it's a CRM a social media tool
  • 00:11:11
    an AI tool maybe it's new equipment for
  • 00:11:13
    your business we're always looking for
  • 00:11:15
    new tools but rather than crawling
  • 00:11:17
    through websites trying to understand
  • 00:11:19
    and piece together different things
  • 00:11:21
    which one is best for you deep research
  • 00:11:23
    is going to do the heavy lifting for you
  • 00:11:25
    using deep research to do this kind of
  • 00:11:27
    research saved me hours in and it helped
  • 00:11:29
    me make a way more informed decision
  • 00:11:32
    based on my specific needs for the
  • 00:11:34
    second business use case for deep
  • 00:11:35
    research if you are doing any kind of
  • 00:11:37
    client research to get a proposal to try
  • 00:11:39
    to get new clients this is perfect for
  • 00:11:42
    you so Freelancers Consultants Etc deep
  • 00:11:44
    research can transform your client
  • 00:11:46
    preparation so instead of getting basic
  • 00:11:48
    company background information you can
  • 00:11:50
    get comprehensive insights on things
  • 00:11:52
    like Market positioning and competitive
  • 00:11:55
    landscape recent initiatives and their
  • 00:11:58
    outcomes potential pain points based on
  • 00:12:01
    industry Trends unique opportunities
  • 00:12:03
    like specific to their specific business
  • 00:12:06
    model so when you use deep research in
  • 00:12:07
    this way and you can get this depth of
  • 00:12:09
    information and level of expertise in
  • 00:12:12
    your proposal you're going to show a
  • 00:12:14
    whole other level of preparation and
  • 00:12:15
    depth of knowledge and research than
  • 00:12:18
    you're competitors so let's just say I'm
  • 00:12:19
    preparing a proposal for a sustainable
  • 00:12:22
    fashion brand looking to expand its
  • 00:12:23
    Market presence in North America the
  • 00:12:25
    purpose of this is to prepare that
  • 00:12:27
    proposal for marketing and branding
  • 00:12:29
    stres that are tailored to the specific
  • 00:12:31
    audience of this brand so this is part
  • 00:12:34
    of my prompt and as you can see here
  • 00:12:38
    it's gone through 32 sources which I
  • 00:12:40
    don't think is a ton but it's a very
  • 00:12:42
    in-depth research output here with a lot
  • 00:12:44
    of statistics and helpful information
  • 00:12:47
    that can go into the proposal now the
  • 00:12:50
    cool thing is if you're using AI to put
  • 00:12:52
    together your proposal you can use this
  • 00:12:53
    research or wherever you get the
  • 00:12:55
    research from whichever model and use it
  • 00:12:57
    in a knowledge Source or knowledge Bas
  • 00:12:59
    that informs your proposal which is
  • 00:13:02
    going to make the putting together
  • 00:13:04
    proposal way easier and faster and again
  • 00:13:06
    here's that exact same question and
  • 00:13:08
    prompt into jpark getting the table of
  • 00:13:11
    contents and then scrolling down here
  • 00:13:13
    with all the citations by the way it did
  • 00:13:16
    141 sources in this result which again I
  • 00:13:20
    really like how the output goes here uh
  • 00:13:23
    gives you the takeaway and then again a
  • 00:13:26
    supplementary list of video resources
  • 00:13:28
    and by the way the way in Jens spark
  • 00:13:30
    before you get that nice report that I
  • 00:13:31
    just showed you this is what it looks
  • 00:13:32
    like on the output screen and this right
  • 00:13:35
    here is they call it the spark page now
  • 00:13:38
    I like the interesting facts at the
  • 00:13:40
    bottom here so check this out for the
  • 00:13:42
    141 sources examined in this by this
  • 00:13:45
    agent the total word count is estimated
  • 00:13:47
    so at an average reading speed of 200
  • 00:13:49
    250 words per minute that's nearly 9
  • 00:13:51
    hours of reading nearly a full day's
  • 00:13:53
    work that it would have taken you to do
  • 00:13:54
    the research that it did in a matter of
  • 00:13:58
    minutes so I think that's kind of funny
  • 00:14:00
    how they include this here use case
  • 00:14:01
    number three is all about target
  • 00:14:03
    audience deep Dives on deep research and
  • 00:14:06
    this is going to be different than what
  • 00:14:07
    you've likely seen so instead of surface
  • 00:14:09
    level demographics you can get a deep
  • 00:14:11
    report on your target audience including
  • 00:14:14
    things like uncovering behavioral
  • 00:14:16
    patterns of your target audience and
  • 00:14:18
    preference shifts these are things that
  • 00:14:20
    I would never find through a
  • 00:14:22
    conventional research so here's an
  • 00:14:24
    example of this kind of search and the
  • 00:14:27
    question I asked so basically you can
  • 00:14:28
    see here I'm researching in the
  • 00:14:29
    behaviors preferences and demographics
  • 00:14:31
    of my target audience to improve
  • 00:14:32
    marketing and product development
  • 00:14:34
    strategies the purpose is to inform
  • 00:14:36
    business Decisions by identifying key
  • 00:14:38
    customer segments pay points and the
  • 00:14:40
    best ways to engage with them and I
  • 00:14:43
    think perplexity deep research did
  • 00:14:45
    actually a pretty decent job on this
  • 00:14:47
    where it goes through the different age
  • 00:14:50
    and generational Trends psychographic
  • 00:14:52
    profiles key pain points etc etc and
  • 00:14:55
    then it also included strategies for in
  • 00:14:57
    effective audience engag engagement
  • 00:14:59
    based on the research that it's finding
  • 00:15:02
    which I really like so that's very
  • 00:15:04
    actionable for me giving me that
  • 00:15:05
    conclusion again all the different
  • 00:15:07
    resources that it searched I think it
  • 00:15:09
    was what 57 yeah 57 sources that it did
  • 00:15:12
    in this search eh it's okay in terms of
  • 00:15:15
    the number of sources but I did like how
  • 00:15:17
    it broke out this kind of research so
  • 00:15:19
    super helpful use case for deep research
  • 00:15:21
    for use case number four for your
  • 00:15:22
    business deep research is really good if
  • 00:15:25
    you need to do any kind of Regulatory
  • 00:15:27
    Compliance research so if you are in
  • 00:15:30
    regulated Industries and you're trying
  • 00:15:32
    to stay current on changing requirements
  • 00:15:34
    that's going to be crucial right so deep
  • 00:15:37
    research can monitor and synthesize that
  • 00:15:40
    regulatory changes across multiple
  • 00:15:43
    jurisdictions identifying really what
  • 00:15:46
    specifically applies to your business
  • 00:15:49
    now think about how long that kind of
  • 00:15:50
    research would take you that would take
  • 00:15:52
    you hours and hours or somebody on your
  • 00:15:54
    team hours and hours to do that or if
  • 00:15:57
    you're paying somebody else to that
  • 00:15:58
    research for you is going to be really
  • 00:15:59
    expensive because of the amount of time
  • 00:16:02
    that it's going to take to do that
  • 00:16:03
    research deep research is going to do
  • 00:16:05
    this in a matter of minutes and you've
  • 00:16:07
    seen the levels of research output that
  • 00:16:09
    these tools give you so we'll just move
  • 00:16:10
    right on to the fifth use case which is
  • 00:16:12
    Market opportunity analysis so check
  • 00:16:15
    this out deep research is going to be
  • 00:16:17
    really good at identifying emerging
  • 00:16:20
    market trends for example before they
  • 00:16:22
    become mainstream so by analyzing all
  • 00:16:23
    the data sources that it does in its
  • 00:16:25
    deep research it's going to highlight
  • 00:16:27
    potential opportunities for you that
  • 00:16:28
    aren't yet saturated so for example if I
  • 00:16:31
    want to learn more about Ai and online
  • 00:16:32
    coaching emerging market trends
  • 00:16:34
    opportunities and challenges I found
  • 00:16:36
    Gemini to be the best output for this
  • 00:16:40
    specific search compared to any of the
  • 00:16:42
    other models that we've been talking
  • 00:16:43
    about here today it's really thorough
  • 00:16:45
    gives you a nice chart etc etc gives you
  • 00:16:48
    the opportunities I really like Gemini's
  • 00:16:50
    output for this now if you think about
  • 00:16:51
    it the ROI calculation for using deep
  • 00:16:54
    research is pretty straightforward right
  • 00:16:56
    so let's just say you value your time or
  • 00:16:59
    your team's time at a simple $50 an hour
  • 00:17:03
    deep research can save you like we saw
  • 00:17:05
    before 9 hours of research that's
  • 00:17:08
    $450 that it is saving you which is far
  • 00:17:11
    more than the subscription cost for
  • 00:17:13
    these platforms even Chachi BT Pro which
  • 00:17:17
    is 200 bucks so now let's talk for a
  • 00:17:18
    minute about who deep research is
  • 00:17:21
    actually for deep research is built
  • 00:17:23
    primarily for people who do intensive
  • 00:17:25
    knowledge work in areas like Finance
  • 00:17:28
    science policy engineering professionals
  • 00:17:31
    who need thorough precise and reliable
  • 00:17:34
    research but in reality like we're
  • 00:17:35
    talking about here today it's really
  • 00:17:37
    valuable for business owners and
  • 00:17:39
    entrepreneurs people who need to make
  • 00:17:40
    informed decisions really really quickly
  • 00:17:43
    who don't have access to a fullon
  • 00:17:45
    research team and people who need to
  • 00:17:47
    stay ahead of market trends or
  • 00:17:49
    consultants and Freelancers who need to
  • 00:17:51
    quickly understand client Industries
  • 00:17:53
    like we talked about here today people
  • 00:17:55
    who create datab backed proposals people
  • 00:17:57
    who need to differentiate themselves
  • 00:17:58
    with with deeper insights than what
  • 00:18:00
    maybe other companies who are trying to
  • 00:18:02
    get that same business are going to be
  • 00:18:04
    coming to the table with content
  • 00:18:06
    creators people who need well researched
  • 00:18:08
    factual content people who want to find
  • 00:18:10
    unique angles on topics the research is
  • 00:18:12
    great for this people who need to verify
  • 00:18:15
    information from multiple sources and
  • 00:18:17
    not just a few sources that you might
  • 00:18:19
    get within a normal you know Google
  • 00:18:21
    search deep research is really good for
  • 00:18:23
    Content creators now as I mentioned at
  • 00:18:24
    the very beginning of this video
  • 00:18:26
    standard prompting techniques often fail
  • 00:18:28
    with deep research because they don't
  • 00:18:30
    Leverage The Unique capabilities of
  • 00:18:33
    these Advanced modes in terms of the
  • 00:18:36
    deep research model so after doing all
  • 00:18:39
    these tests I put together a prompt
  • 00:18:41
    framework that I've been kind of
  • 00:18:42
    alluding to throughout the video here
  • 00:18:44
    that consistently delivers what I think
  • 00:18:46
    are superior results when using deep
  • 00:18:49
    research and here's that four-part deep
  • 00:18:51
    research prompting framework it begins
  • 00:18:53
    with defining the scope so you really
  • 00:18:55
    want to start by clearly defining the
  • 00:18:57
    entire scope boundaries of your research
  • 00:19:00
    question and you want to be super
  • 00:19:02
    specific about things like exactly what
  • 00:19:05
    you need to know the depth required in
  • 00:19:07
    the research that you want the model to
  • 00:19:09
    do any specific sources that you want
  • 00:19:11
    included or excluded be clear about that
  • 00:19:14
    in the first part of your query and also
  • 00:19:17
    if there's a time period relevance to
  • 00:19:19
    the research you want to be clear about
  • 00:19:21
    that in your prompt as well the second
  • 00:19:22
    part of the deep research prompting
  • 00:19:24
    framework is being very specific about
  • 00:19:26
    the output format that you are looking
  • 00:19:29
    for next in the prop you want to specify
  • 00:19:31
    exactly how you want the information
  • 00:19:33
    presented to you so tell it the
  • 00:19:34
    structure of the report that you want
  • 00:19:35
    tell it what sections that you want to
  • 00:19:37
    make sure are required maybe the
  • 00:19:39
    proposal that you're submitting to a
  • 00:19:41
    specific company for example needs to
  • 00:19:42
    include certain sections well you can
  • 00:19:44
    put that into your deep research prompt
  • 00:19:47
    if you have data visualization
  • 00:19:50
    preferences put it in there if you want
  • 00:19:53
    certain uh a certain citation style put
  • 00:19:55
    it into your prompt the next part of
  • 00:19:57
    your prompt and this is the critical
  • 00:19:59
    step that most people miss when it comes
  • 00:20:00
    to using deeper research is that you
  • 00:20:02
    want to direct how you want the AI to
  • 00:20:06
    reason through conflicting information
  • 00:20:08
    that it might be finding across all
  • 00:20:10
    these various data sources so give it
  • 00:20:12
    prioritization criteria for sources how
  • 00:20:15
    to handle that conflicting data what
  • 00:20:17
    constitutes sufficient evidence for the
  • 00:20:20
    conclusions that it's making and then
  • 00:20:21
    you want to be super clear about any
  • 00:20:23
    kind of areas that require deeper
  • 00:20:25
    exploration in your research and then
  • 00:20:27
    finally specify how you plan to use this
  • 00:20:29
    information which is going to help the
  • 00:20:30
    AI prioritize relevance in its research
  • 00:20:33
    output that it gives you so for example
  • 00:20:35
    your decision making context like why
  • 00:20:38
    are you having this research done in the
  • 00:20:39
    first place problems that you're trying
  • 00:20:41
    to solve what actions that you're going
  • 00:20:42
    to take based on the findings of the
  • 00:20:44
    deep research report uh your level of
  • 00:20:47
    expertise in the subject like what do
  • 00:20:49
    you already know so that's going to make
  • 00:20:51
    it easier for the AI to give you the
  • 00:20:53
    kind of information that's going to be
  • 00:20:54
    most helpful for you and this right here
  • 00:20:56
    is the prompt template that I put
  • 00:20:57
    together that contains the exact
  • 00:20:59
    framework that I just took you through
  • 00:21:02
    Now by structuring The Prompt this way
  • 00:21:04
    you're essentially programming a custom
  • 00:21:07
    research workflow tailored to your
  • 00:21:10
    specific needs and it's this level of
  • 00:21:13
    prompt with the framework I've just
  • 00:21:14
    taken you through that gets you the kind
  • 00:21:16
    of indepth output for your research that
  • 00:21:19
    we've been talking about here today and
  • 00:21:20
    by the way if you want this prompt
  • 00:21:22
    template I'll link to it in the
  • 00:21:23
    description below so which platform
  • 00:21:25
    should you use when it comes to deep
  • 00:21:26
    research well it's going to really
  • 00:21:27
    depend on what your specific needs are
  • 00:21:29
    as you saw here today perplexity is
  • 00:21:31
    great for Speed and quick insights which
  • 00:21:34
    kind of goes against what I think deep
  • 00:21:36
    research is really all about Gemini is
  • 00:21:38
    going to be great for methodical
  • 00:21:39
    research from trusted sources chbt for
  • 00:21:42
    depth and complexity if your budget
  • 00:21:43
    allows and then finally jpark AI which I
  • 00:21:46
    think is sort of the best of all worlds
  • 00:21:48
    among all these deep research AI models
  • 00:21:50
    but regardless of the platform proper
  • 00:21:53
    prompting like we've talked about here
  • 00:21:54
    today is going to be the key to
  • 00:21:55
    unlocking the Full Potential from your
  • 00:21:57
    deep research search all right thanks
  • 00:21:59
    for watching today's video if you are
  • 00:22:00
    interested in joining my school
  • 00:22:01
    Community the AI Playbook if you're an
  • 00:22:03
    online business looking to leverage AI
  • 00:22:05
    to streamline and increase your profit a
  • 00:22:07
    link to the community in the description
  • 00:22:09
    below thanks so much for watching and
  • 00:22:11
    I'll see you in the next video
标签
  • Deep Research
  • AI Models
  • ChatGPT
  • Gemini
  • Perplexity
  • Jenspark
  • Business Insights
  • Research Framework
  • Entrepreneurs
  • Freelancers