Latest / Inside Strategic Relations / Going beyond artificial intelligence

Transcript

Going beyond artificial intelligence · Inside Strategic Relations

• 0:01
  Speaker 1: Let's talk about using artificial intelligence to help organize information
• 0:05
  Speaker 1: rather than creating new ideas and concepts, and why that's
• 0:08
  Speaker 1: important if you're a subject matter expert, an entrepreneur, or
• 0:12
  Speaker 1: an executive. I'm just hit from inside strategic relations, where
• 0:16
  Speaker 1: we help you think critically and gain clarity about decision
• 0:21
  Speaker 1: making because you are using tools in an appropriate way.
• 0:26
  Speaker 1: So right up front, I will tell you that I
• 0:29
  Speaker 1: do use artificial intelligence, but I also use machine learning,
• 0:34
  Speaker 1: natural language processing and other tool text analysis tools to
• 0:39
  Speaker 1: look over my body of work that I've created for
• 0:41
  Speaker 1: the last thirty thirty five years and helps in using
• 0:45
  Speaker 1: it to help organize ideas and concepts.
• 0:48
  Speaker 2: However, I just a hit do not.
• 0:52
  Speaker 1: Use AI to create the actual content itself. And the
• 0:56
  Speaker 1: reason I don't do that is because, and you may
• 0:59
  Speaker 1: have I've seen these articles as well, is that forty
• 1:02
  Speaker 1: two percent of AI responses are blatantly incorrect. AI will
• 1:09
  Speaker 1: make up ideas, it will make up bullet points, it'll
• 1:13
  Speaker 1: come up with stats that don't exist, it'll have citations
• 1:16
  Speaker 1: that are incorrect, it'll hallucinate, it'll misunderstand the prompt and
• 1:21
  Speaker 1: provide you something that with cognitive So if you write
• 1:25
  Speaker 1: the prompt and it gives you a response that sounds appropriate.
• 1:29
  Speaker 2: There's a cognitive bias.
• 1:31
  Speaker 1: In your mind that this is the answer because AI
• 1:34
  Speaker 1: gave it to you.
• 1:35
  Speaker 2: And so all those.
• 1:36
  Speaker 1: Things together give you a forty two percent error rate,
• 1:41
  Speaker 1: in some cases a fifty percent error rate. You might
• 1:43
  Speaker 1: as well flip a coin to see whether or not
• 1:44
  Speaker 1: the output is relevant. And so for using it beyond
• 1:48
  Speaker 1: organizing factual information, then we put ourselves at risk of,
• 1:55
  Speaker 1: you know, looking stupid. And so if you make a
• 1:57
  Speaker 1: claim that X, Y and Z is a thing, and
• 2:01
  Speaker 1: you've talked to the machine about it, and you've talked
• 2:04
  Speaker 1: to the committees about it, and you've done very surface
• 2:08
  Speaker 1: level research, you may find yourself in a bad situation.
• 2:12
  Speaker 1: And in fact, there could be legal liability, reputational risk,
• 2:16
  Speaker 1: or other risks associated with that. And so, how do
• 2:19
  Speaker 1: we use artificial intelligence in a way that's going to
• 2:23
  Speaker 1: improve our ability to discern from information better decisions? And
• 2:30
  Speaker 1: I'm going to share with you how I'm doing it
• 2:32
  Speaker 1: with some of the modern tools that are available, But
• 2:34
  Speaker 1: I want to also give you some strategies so that
• 2:36
  Speaker 1: as tools evolve, because there could be a time where
• 2:39
  Speaker 1: AI is one hundred percent correct in the response. But
• 2:42
  Speaker 1: because you're prompt wasn't clear that response is not appropriate
• 2:47
  Speaker 1: to the action and outcome that you need. So we
• 2:49
  Speaker 1: start with outcome management. What is the outcome that we
• 2:53
  Speaker 1: want to create? So I have in front of me
• 2:56
  Speaker 1: all the material So I bought the rights to some
• 2:58
  Speaker 1: materials by a copyright named Patrick Quinn, and I've got
• 3:02
  Speaker 1: all of his materials.
• 3:03
  Speaker 2: I've got his unpublished books. I've got some.
• 3:07
  Speaker 1: Transcripts and everything I could find about Patrick Quinn in
• 3:12
  Speaker 1: Google Notebook. So Google Notebook LM is an AI powered
• 3:18
  Speaker 1: tool designed to help you organize, summarize, and analyze documents.
• 3:22
  Speaker 1: And so first off, we got to make sure we
• 3:24
  Speaker 1: got the appropriate tool. I don't want to have to
• 3:27
  Speaker 1: train an AI model on Patrick Quinn's materials and maintain
• 3:30
  Speaker 1: a separate AI model that's just this one author's materials
• 3:35
  Speaker 1: in order to create derivative works. So again we want
• 3:38
  Speaker 1: to choose the appropriate environment. Then we also want to
• 3:42
  Speaker 1: structure that data set. Now, Notebook LM lets us put
• 3:46
  Speaker 1: in all of our PDF documents and videos and past
• 3:51
  Speaker 1: newsletter articles into a library of sources. Now it doesn't
• 3:55
  Speaker 1: do a very good job letting me clean up these sources,
• 3:58
  Speaker 1: but I can always download source, clean it up, and
• 4:01
  Speaker 1: upload it again as a new source and delete the
• 4:03
  Speaker 1: one that it was old. The reason I'm telling you
• 4:05
  Speaker 1: this is because we can very often organize and structure.
• 4:09
  Speaker 2: Data in order to get AI to give us a
• 4:12
  Speaker 2: better result.
• 4:14
  Speaker 1: So, if I have five documents that have biographical information
• 4:19
  Speaker 1: about Patrick Quinn, and I come back and write a
• 4:22
  Speaker 1: summary of those materials, even if it's a bullet point list,
• 4:25
  Speaker 1: and I add the summary linking to the in depth
• 4:29
  Speaker 1: materials notebook, LM will do a better job of giving
• 4:32
  Speaker 1: me nuanced responses. And so I can say, tell me
• 4:36
  Speaker 1: about Patrick Quinn's life in such and such time, and
• 4:41
  Speaker 1: it will. If I have a list of clients during
• 4:43
  Speaker 1: a time period on a timeline, if I have articles
• 4:47
  Speaker 1: in a timeline, then it can come back with something
• 4:50
  Speaker 1: relevant to that time period because the prompt is asking
• 4:52
  Speaker 1: about a time period. Now, when I ask you to
• 4:55
  Speaker 1: do something, I'm not going to say write an article
• 4:57
  Speaker 1: about I'm going to more appropriately, I'm going to ask
• 5:01
  Speaker 1: a question, and then I'm going to say, give me
• 5:03
  Speaker 1: a list of tips, ideas, and concepts. Because I do
• 5:06
  Speaker 1: know I know the information to a certain degree, but
• 5:11
  Speaker 1: I do know if it gives me a list of
• 5:12
  Speaker 1: twenty two tips, that some of those are just gonna
• 5:15
  Speaker 1: be blatantly wrong. Again, it could be forty two percent,
• 5:18
  Speaker 1: it could be ten percent, but I am making the
• 5:21
  Speaker 1: decision or creating the discernment necessary to digest this information.
• 5:28
  Speaker 2: Now, your error rate increases.
• 5:30
  Speaker 1: As you use AI generated materials to feed your overall model,
• 5:37
  Speaker 1: because basically notebook LM is creating small models.
• 5:40
  Speaker 2: And so what you have to.
• 5:41
  Speaker 1: Do here is to make sure you're just like when
• 5:45
  Speaker 1: you're doing functional research, and I get paid to commissions
• 5:49
  Speaker 1: to write research, so I'm really firm on the research methods,
• 5:54
  Speaker 1: coming up with a thesis, and then going and doing
• 5:57
  Speaker 1: your primary research, and then testing your thesis, and then
• 6:00
  Speaker 1: coming in and you know, looking at a counter thesis.
• 6:03
  Speaker 1: A lot of these elements help you make better decisions.
• 6:06
  Speaker 1: They take time. But if you go back to your
• 6:09
  Speaker 1: sources and refine your sources to be primary materials and
• 6:15
  Speaker 1: you don't feed it back other AI, not even in
• 6:18
  Speaker 1: the summaries, by the way, you have to go through
• 6:20
  Speaker 1: and read the materials and use your brain to rewrite
• 6:23
  Speaker 1: it in your own words in order to summarize those materials,
• 6:27
  Speaker 1: and then clearly give the document author attribution. So in
• 6:31
  Speaker 1: this case, Patrick quinn materials are clearly defined as Patrick
• 6:35
  Speaker 1: Quinn author, but any derivatives that I'm making are defined
• 6:39
  Speaker 1: as or any summaries or you know, mind maps or.
• 6:43
  Speaker 2: Outlines are clearly defined as just and hit.
• 6:46
  Speaker 1: I'm producing those things so that when I ask, you know,
• 6:52
  Speaker 1: provide a tip sheet of what Patrick Quinn recommends, updating
• 6:56
  Speaker 1: for modern systems and future proofing with proven advice. I
• 7:01
  Speaker 1: know right up front that I've said specifically, look at
• 7:04
  Speaker 1: Patrick Quinn's material there's kind of a test in there
• 7:07
  Speaker 1: to see if it's still relevant today. But that large
• 7:11
  Speaker 1: language model is going to test against average, and we
• 7:14
  Speaker 1: don't want average results.
• 7:16
  Speaker 2: So when I get a.
• 7:17
  Speaker 1: List back, I'm looking at what is appropriate to my audience,
• 7:21
  Speaker 1: what is appropriate to my audience within the context of
• 7:25
  Speaker 1: what I know about Patrick Quinn's materials. Now, if there
• 7:29
  Speaker 1: are any failures in there, it will be my understanding
• 7:33
  Speaker 1: of the materials that causes the failure. And ultimately it's
• 7:37
  Speaker 1: acceptable failure because I can do something about it. If
• 7:40
  Speaker 1: I take the AI stuff as gospel, because it's you know,
• 7:43
  Speaker 1: I've feted all this guy's materials and it said this
• 7:46
  Speaker 1: material contains such as now I gotta still fact check.
• 7:50
  Speaker 1: Because we want to bridge materials from the nineteen seventies
• 7:54
  Speaker 1: and nineteen eighties to modern application, I have to have
• 7:58
  Speaker 1: the subject matter expertise to do that. And so a
• 8:01
  Speaker 1: lot of times what happens, and I've talked about this
• 8:03
  Speaker 1: when we're searching the internet, is that if you have
• 8:06
  Speaker 1: subject matter expertise, you're going to work more efficiently. You're
• 8:10
  Speaker 1: going to work more effectively creating materials than AI would,
• 8:15
  Speaker 1: or you would doing general search on the internet, because
• 8:19
  Speaker 1: you've got a primary source. You've got your experience, you've
• 8:23
  Speaker 1: got your understanding, you've got case studies that you produced.
• 8:26
  Speaker 1: So your time and effort should be in creating those
• 8:28
  Speaker 1: primary source materials, which are case studies, definition of terms,
• 8:35
  Speaker 1: the documentation of processes and procedures, and then from there
• 8:38
  Speaker 1: having a staff member create derivative work. I used to
• 8:41
  Speaker 1: hire a lot of researchers and folks that understand research
• 8:44
  Speaker 1: methods to create that draft. And again you might be saying, well,
• 8:49
  Speaker 1: why not just have AI create the draft? Well, no,
• 8:52
  Speaker 1: I want writers. I want authors. I want human creators
• 8:56
  Speaker 1: because I have a human audience.
• 9:00
  Speaker 2: And if that seems.
• 9:01
  Speaker 1: Like a lot of effort or anything, then I need
• 9:03
  Speaker 1: to do that. So I can ask AI to give
• 9:05
  Speaker 1: me one hundred and one headlines based on these headline formulas,
• 9:10
  Speaker 1: and then I can pick and choose out of there.
• 9:12
  Speaker 1: And then for everything I pick and choose, I need
• 9:14
  Speaker 1: to go through and have a very clear reason why
• 9:19
  Speaker 1: I chose that material based on the outcome that I'm
• 9:22
  Speaker 1: looking for. Now, he might be saying AI promised to
• 9:25
  Speaker 1: make things easier. Well, so did the Internet, so did notebooks,
• 9:31
  Speaker 1: so did computers. Nothing makes it easier when the results matter.
• 9:36
  Speaker 1: And so when we're going through this process, and particularly
• 9:39
  Speaker 1: on notebook LM, we're curating the sources, we're thinking about
• 9:45
  Speaker 1: the prompts. We want the prompt to something that I
• 9:49
  Speaker 1: would have given to a human, I would have given
• 9:51
  Speaker 1: to a researcher, and then from there I have to
• 9:54
  Speaker 1: make decisions.
• 9:55
  Speaker 2: About whether the materials are useful or not.
• 9:57
  Speaker 1: Now, in one hundred percent of what I'm writing, this
• 10:00
  Speaker 1: is only like five to ten percent of what I'm doing,
• 10:02
  Speaker 1: because ultimately I need to take my experiences, my client's experiences,
• 10:07
  Speaker 1: lessons learned, primary research, and make it applicable to the
• 10:12
  Speaker 1: specific problems that my client's face. So again we're outcome oriented.
• 10:17
  Speaker 1: We've defined the problems up front that we're trying.
• 10:19
  Speaker 2: To solve for.
• 10:20
  Speaker 1: We're not looking for AI to give us new ideas
• 10:23
  Speaker 1: and concepts because it can't.
• 10:26
  Speaker 2: AI itself is pulling from a general model of knowledge.
• 10:32
  Speaker 1: It may have scanned hundreds of thousands of primary research documents,
• 10:37
  Speaker 1: but it hasn't evaluated every document for legitimacy.
• 10:42
  Speaker 2: And we do know that right now.
• 10:43
  Speaker 1: A lot of even you go to a Google Scholar
• 10:46
  Speaker 1: for example, a lot of these platforms have research papers
• 10:50
  Speaker 1: that are redacted all the time because they were not
• 10:53
  Speaker 1: well structured or organized research. But also AI is training
• 10:57
  Speaker 1: on historical materials as well as present materials, so there'll
• 11:00
  Speaker 1: be a conflict with material that may have been redacted after.
• 11:04
  Speaker 2: The model was produced.
• 11:06
  Speaker 1: So how else might I use AI? Well, I actually
• 11:10
  Speaker 1: don't primary use AI. I actually use some machine learning.
• 11:15
  Speaker 1: So there's an addar index that can tell me about
• 11:18
  Speaker 1: documents that are linked to each other. I'm based on
• 11:20
  Speaker 1: those links, what other things should be probably linked to
• 11:23
  Speaker 1: those or what might be linked to it. Still have
• 11:25
  Speaker 1: to make a decision myself of what should be I
• 11:28
  Speaker 1: used your card, which is based on the context of
• 11:30
  Speaker 1: these materials, what other topics or documents in my note
• 11:34
  Speaker 1: set are related to that, And then of course I'll
• 11:37
  Speaker 1: do abstractions, I'll do outlines I.
• 11:39
  Speaker 2: Might do before I write a report.
• 11:41
  Speaker 1: I might look at the problems, the pains, the solutions,
• 11:44
  Speaker 1: the actions that I'm asking the audience to take, and
• 11:47
  Speaker 1: then I may build a book proposal if I'm writing
• 11:51
  Speaker 1: a book. Again, the AI is no different than hiring
• 11:56
  Speaker 1: some wet behind the year college student to go through
• 11:58
  Speaker 1: the library and dig through some materials. They have very
• 12:01
  Speaker 1: little discernment about what is appropriate and what is not
• 12:05
  Speaker 1: appropriate for the specific application of information. Now, if you're marketing,
• 12:10
  Speaker 1: for example, you can use AI to give you an
• 12:13
  Speaker 1: oppression of these materials. Hey, based on these materials, who
• 12:16
  Speaker 1: do you think within this audience is going to be
• 12:19
  Speaker 1: most motivated and why? And you can get their AI's opinion,
• 12:24
  Speaker 1: But ultimately you have to make a decision. Another factor
• 12:27
  Speaker 1: to understand is that there's certain structure to languish. Now
• 12:33
  Speaker 1: we talk more about this in AD briefings copywriting tips,
• 12:36
  Speaker 1: because there's certain structures that sell better than others, like
• 12:39
  Speaker 1: the AI DA formula, which is attention, interest, desire and
• 12:43
  Speaker 1: action is a structure which is used for advertising to
• 12:47
  Speaker 1: help take someone from general awareness of a topic into
• 12:52
  Speaker 1: a specific point of interest, and then from the specific
• 12:55
  Speaker 1: point of interest into.
• 12:56
  Speaker 2: Asking them to take action.
• 12:59
  Speaker 1: If we structure our information well or we ask ai
• 13:04
  Speaker 1: to extract into a structure that actually works pretty good.
• 13:08
  Speaker 1: And when I was in college, which I was in
• 13:11
  Speaker 1: college since I got out of high school until I
• 13:13
  Speaker 1: was forty forty something years old, I will take college
• 13:16
  Speaker 1: classes because MIT and at different universities have open classes.
• 13:20
  Speaker 1: I'll take a college class just because I'm interested in
• 13:22
  Speaker 1: a topic. But my biggest challenge has been either MLS
• 13:27
  Speaker 1: formatting or APA formatting. And so when I was at
• 13:31
  Speaker 1: Arizona State University, I ended up hiring teachers, assistants or
• 13:35
  Speaker 1: actual professors at other universities in order to review my
• 13:39
  Speaker 1: final document to make sure it was properly formatted, because
• 13:44
  Speaker 1: I've had fifty percent of the grade taken off because
• 13:48
  Speaker 1: I couldn't get the APA formatting right. I don't know
• 13:51
  Speaker 1: what the hell it is. I wasn't getting the APA
• 13:53
  Speaker 1: formatting right. But today, if I need an APA format
• 13:56
  Speaker 1: a document, I can go ahead and write my original materials.
• 13:59
  Speaker 1: I can do citations using zoto. I can have my
• 14:03
  Speaker 1: document written in PANDOC, I can process it through a bibliography.
• 14:07
  Speaker 1: And this may not make sense to you, but we
• 14:09
  Speaker 1: can help you with this if you're producing reports to
• 14:11
  Speaker 1: promote your consulting services, if you're writing technical documents and
• 14:15
  Speaker 1: you want to streamline that process.
• 14:16
  Speaker 2: But I could ask AI to give me feedback.
• 14:21
  Speaker 1: Now, I didn't ask it to transform the document to
• 14:24
  Speaker 1: make an APA format, but I could say, give me
• 14:27
  Speaker 1: feedback on where I am not aligned with APA version, whatever, whatever, whatever,
• 14:34
  Speaker 1: And I could get faster feedback than hiring a third
• 14:37
  Speaker 1: party to review my document. I could get some tutoring
• 14:42
  Speaker 1: associated on how to improve it and make it better
• 14:44
  Speaker 1: in the future, because I sincerely want to make sincerely
• 14:47
  Speaker 1: want to make it right the first time.
• 14:49
  Speaker 2: But it allows me.
• 14:50
  Speaker 1: A little bit more flexibility to focus on the core materials. Now,
• 14:54
  Speaker 1: what I don't do is say, hey, rewrite this document
• 14:57
  Speaker 1: into APA format, because very often it'll rewrite the document
• 15:01
  Speaker 1: because your prompt is tokenized, and if it says rewrite
• 15:05
  Speaker 1: the document use APA format, it may change context and
• 15:10
  Speaker 1: character of the document, and then I'm going to end
• 15:12
  Speaker 1: up spending an extra time reading the document itself in
• 15:15
  Speaker 1: order to make sure it didn't change something that's important. Again,
• 15:20
  Speaker 1: AI makes up facts and figures. It may think it's
• 15:22
  Speaker 1: helping me, but it's not necessarily helping me. So again
• 15:26
  Speaker 1: I can write a document, I can get feedback on
• 15:29
  Speaker 1: that document. With AI, I can go through the sessions
• 15:33
  Speaker 1: and ask specific questions and finish the document off. Now,
• 15:38
  Speaker 1: when I was writing commissioned papers, so there was a
• 15:42
  Speaker 1: period of time where I was writing commissioned doctoral and
• 15:45
  Speaker 1: master's thesiss and in that case I would create the
• 15:49
  Speaker 1: materials appropriate within subject matter expertise of business development and
• 15:56
  Speaker 1: risk management, or areas I had intense interest, some agricultural,
• 16:02
  Speaker 1: some environmental. I did a big paper on art for
• 16:05
  Speaker 1: somebody who was getting a I guess they're getting a
• 16:07
  Speaker 1: master's in art. But long story short. Even then, I
• 16:13
  Speaker 1: would interview them to extract the details, create the primary
• 16:17
  Speaker 1: sources based on the assignment. The outcome is the assignment,
• 16:21
  Speaker 1: and then I would build out the document and I
• 16:24
  Speaker 1: would build out the drafts, and then I would have
• 16:27
  Speaker 1: them review the drafts. The primary reviewing the drafts, giving
• 16:30
  Speaker 1: me feedback, and then ultimately I would produce the document
• 16:35
  Speaker 1: for their approval.
• 16:37
  Speaker 2: Then I would get.
• 16:38
  Speaker 1: The formatting done for whatever the MLS or APA, mostly
• 16:42
  Speaker 1: through a contractor, and then I would give it back
• 16:45
  Speaker 1: to them with deliberate errors in the system. I would
• 16:49
  Speaker 1: ask them to finish the document off and then I
• 16:53
  Speaker 1: would review the document looking for the deliberate errors in
• 16:56
  Speaker 1: the system, and if they didn't fix the deliberate errors
• 17:00
  Speaker 1: in the system, because I told them there are things
• 17:02
  Speaker 1: in this document that I have made incorrect on purpose,
• 17:06
  Speaker 1: because again, if you're going for a doctorate or you're
• 17:08
  Speaker 1: going for a master's degree. A lot of these were
• 17:10
  Speaker 1: busy executives, entrepreneurs, people who really didn't have that much time.
• 17:14
  Speaker 2: But we could do maybe a three hour phone.
• 17:17
  Speaker 1: Interview from time to time, and then I would go
• 17:19
  Speaker 1: out and actually do primary research and build the citations
• 17:22
  Speaker 1: and stuff. Long story short, if they turned in that
• 17:26
  Speaker 1: paper and the professor actually read it, they're going to not.
• 17:29
  Speaker 2: Look very good.
• 17:30
  Speaker 1: And I would do that on purpose, because the integrity
• 17:34
  Speaker 1: of the document itself was their research, their thought, their mind.
• 17:40
  Speaker 1: The practical challenge was that they didn't have the time
• 17:43
  Speaker 1: or resources to go through and finish it. You know,
• 17:46
  Speaker 1: they could pay me five to ten thousand dollars to
• 17:48
  Speaker 1: do the work, of which I'd hire contractors and organize
• 17:51
  Speaker 1: everything for them, but ultimately it had to be their work.
• 17:55
  Speaker 1: If you're taking what AI gives you, there's a very
• 17:58
  Speaker 1: high probability of this. Some of these errors are built
• 18:01
  Speaker 1: in checksums in the system I personally, and part of
• 18:05
  Speaker 1: some of the services we provide our clients is we
• 18:08
  Speaker 1: will taint the LLM models, the large language models. Now,
• 18:12
  Speaker 1: what I mean by taming the model, it's basically curating
• 18:16
  Speaker 1: what the AI says about you as an individual consultant
• 18:19
  Speaker 1: or a marketer, or a copywriter or an author of
• 18:21
  Speaker 1: a book. Because AI is if you publish a book tomorrow,
• 18:24
  Speaker 1: AI isn't going to know about it.
• 18:26
  Speaker 2: The day after that.
• 18:27
  Speaker 1: So we do things to prompt AI to better understand
• 18:33
  Speaker 1: your position in context in the marketplace. So the book
• 18:36
  Speaker 1: I've put together about establishing instant credibility, while I don't
• 18:39
  Speaker 1: cover this topic, we are seeding the environment so that
• 18:43
  Speaker 1: the next time the language model refreshes, it knows about
• 18:47
  Speaker 1: you over anybody else. Now, if that's possible, then it's
• 18:52
  Speaker 1: going to attribute to a perpetual inaccuracy of information. If
• 18:57
  Speaker 1: you're going to make buying decisions, if you're going to
• 19:00
  Speaker 1: do primary research, if you're going to solve business problems.
• 19:05
  Speaker 1: And so again, the AI is an assistant. It can
• 19:09
  Speaker 1: consume massive volumes of information, but if you're not going
• 19:12
  Speaker 1: back into the primary sources to fully contextualize and understand it,
• 19:17
  Speaker 1: understand it, you will create materials that are blatantly wrong.
• 19:21
  Speaker 2: Now, again, I don't know if it's on purpose.
• 19:23
  Speaker 1: I don't know if it's just part of the fact
• 19:25
  Speaker 1: that it has such a broad scope of information, But
• 19:28
  Speaker 1: I do know that that process itself is going to
• 19:31
  Speaker 1: allow you to.
• 19:32
  Speaker 2: Speak and promptu or extraneously.
• 19:36
  Speaker 1: I don't know if that's the right word, but it's
• 19:38
  Speaker 1: going to allow you to speak on the topic with
• 19:40
  Speaker 1: intelligence because you're going to actually be training your model.
• 19:45
  Speaker 2: The actual intelligence is in your head. And so again, if.
• 19:49
  Speaker 1: I'm structuring primary materials in my note taking application, I'm
• 19:54
  Speaker 1: properly citing my sources. I'm going to primary sources in
• 19:58
  Speaker 1: different types of libra, or I'm building from case studies
• 20:02
  Speaker 1: I produce for myself. Then I'm asking AI to help
• 20:05
  Speaker 1: highlight key ideas and concepts that are relevant to a
• 20:08
  Speaker 1: specific audience. Because the original materials were written for you.
• 20:12
  Speaker 1: But now we've got to translate in order to write
• 20:15
  Speaker 1: for a specific audience, because the outcome we desire is
• 20:18
  Speaker 1: a senior level executive to hire our consulting firm or
• 20:20
  Speaker 1: our marketing agency over anything else that's available in the market.
• 20:25
  Speaker 1: We're not writing for general people typically, And if you're
• 20:28
  Speaker 1: using my establishing instant credibility, I talk about writing materials
• 20:31
  Speaker 1: for a specific dream one hundred audience and then putting
• 20:34
  Speaker 1: those materials in front of that audience.
• 20:37
  Speaker 2: That's marketing work.
• 20:39
  Speaker 1: And so we want to build quality materials because if
• 20:42
  Speaker 1: we hire a marketing agency to extract from those materials
• 20:45
  Speaker 1: to promote us, we want the original materials to be
• 20:48
  Speaker 1: clean enough that if they go off and they're sloppy
• 20:51
  Speaker 1: with the AI or they're sloppy with other things, that
• 20:54
  Speaker 1: there's still be contextual relevance to bridge from the problem
• 20:57
  Speaker 1: the prospect has to you as a possible solution, and
• 21:02
  Speaker 1: then their action specifically to contact you. So again this
• 21:07
  Speaker 1: is revealing behind the curtain how successful people create influence,
• 21:13
  Speaker 1: whether they're using the library, whether they're using search engines,
• 21:17
  Speaker 1: or whether they're using artificial intelligence, is because they're not
• 21:21
  Speaker 1: relying on the external tool. They're developing and packaging their
• 21:25
  Speaker 1: own abilities. Because think about this, you write a massive
• 21:29
  Speaker 1: paper on how to solve. So I wrote an anti
• 21:32
  Speaker 1: money laundering paper. It was strategies on how to identify
• 21:35
  Speaker 1: money laundering? What types of money laundering is happening in
• 21:40
  Speaker 1: the current time. So I could use AI to say,
• 21:42
  Speaker 1: give me a list of money laundering prosecutions from this
• 21:46
  Speaker 1: certain database, and it would give me a list of prosecutions,
• 21:49
  Speaker 1: and then I would build a story around those prosecutions
• 21:52
  Speaker 1: about maybe how did it happen, what are some of
• 21:54
  Speaker 1: the preventative measures, what are the.
• 21:55
  Speaker 2: Root causes to look for? And then at the end
• 21:59
  Speaker 2: it had kind of some guidance.
• 22:01
  Speaker 1: All of that was built for and risk management organization
• 22:06
  Speaker 1: to get hired by banks and to get hired by
• 22:10
  Speaker 1: government agencies in order to identify money laundering. And so
• 22:15
  Speaker 1: it's not a massive reports. It's based on my understanding
• 22:19
  Speaker 1: of the audit process and my understanding of how money
• 22:23
  Speaker 1: I've done anti money laundering in financial services areas. It's
• 22:27
  Speaker 1: built on the clients insights, but ultimately as a persuasive
• 22:32
  Speaker 1: document to identify the big problem that money laundering exists
• 22:36
  Speaker 1: and it's likely going through your bank, and here's all
• 22:39
  Speaker 1: the horrible things that are going to happen if you
• 22:41
  Speaker 1: don't have good money laundering controls, and then it gives
• 22:44
  Speaker 1: practical examples, and then finally asks for a meeting. It says, hey,
• 22:49
  Speaker 1: if you want to talk specifically about your challenges, we
• 22:52
  Speaker 1: work under a non disclosure we are we're available for consultations,
• 22:56
  Speaker 1: and we can conduct audits. And by the way, here's
• 22:59
  Speaker 1: how you contact does. Again, there's a spectrum of subject
• 23:03
  Speaker 1: matter expertise, demonstrative ability, practical stories. So the story element
• 23:11
  Speaker 1: helps people understand complex information. And then marketing, which is
• 23:15
  Speaker 1: to call to action. AI is not going to consistently
• 23:19
  Speaker 1: do all of that for you.
• 23:21
  Speaker 2: Now, you can.
• 23:22
  Speaker 1: Break it down to a thousand little pieces and have
• 23:24
  Speaker 1: AI build these out. Hey, AI, give me a list
• 23:27
  Speaker 1: of problems this specific client might face if they're trying
• 23:31
  Speaker 1: to solve problems in money laundering or to stay compliant
• 23:34
  Speaker 1: with certain SEC filings or certain financial crime related filings.
• 23:41
  Speaker 2: Here are the definition of terms. AI might give you
• 23:45
  Speaker 2: the list of.
• 23:45
  Speaker 1: Problems, but you've got to pick which problems are relevant
• 23:48
  Speaker 1: to your audience, because some of those problems listed in
• 23:51
  Speaker 1: there are not going to be relevant to the decision maker. Again,
• 23:56
  Speaker 1: artificial intelligence gives us the ability to have that general
• 24:00
  Speaker 1: research done to search large volumes of data. But I
• 24:04
  Speaker 1: also can go back into my body of work and
• 24:06
  Speaker 1: use text analysis.
• 24:09
  Speaker 2: I could use ingrams, so I could look for.
• 24:14
  Speaker 1: Anti money laundering in ten or fifteen different variations. And
• 24:19
  Speaker 1: I can go through all my existing materials that I
• 24:22
  Speaker 1: wrote about money laundering, extract key ideas and concepts into
• 24:25
  Speaker 1: a map of content, and then from that just basically
• 24:29
  Speaker 1: just start typing and write out most of the report
• 24:32
  Speaker 1: without AI doing the work, and I'm going to have
• 24:35
  Speaker 1: a better understanding of the materials so that if the
• 24:37
  Speaker 1: client says, hey, this, so there's a money laundering thing
• 24:42
  Speaker 1: where people will use lottery pools in order to take
• 24:47
  Speaker 1: dirty money to buy lottery tickets, and then the lottery
• 24:50
  Speaker 1: winnings pay out the cash to replace the original money. See,
• 24:56
  Speaker 1: all they got to do is break even, but the
• 24:58
  Speaker 1: new money's clean. And so in some areas there are
• 25:02
  Speaker 1: certain people that are socially related to each other that
• 25:07
  Speaker 1: tend to win the lottery over others, and it's only
• 25:10
  Speaker 1: because they're using these schemes to wash dirty cash. It
• 25:14
  Speaker 1: is possible for the client to say that's not happening,
• 25:17
  Speaker 1: that's not possible. So again, I would use the story
• 25:21
  Speaker 1: of this actually was prosecuted, or this was actually suspected
• 25:25
  Speaker 1: by a FBI or an agency with authority. AI is
• 25:31
  Speaker 1: not going to build that out very carefully. It's going
• 25:33
  Speaker 1: to take you more work to get AI to give
• 25:34
  Speaker 1: you that. You know, if you put a bunch of
• 25:36
  Speaker 1: monkeys in a room on a typewriter, they're eventually going
• 25:38
  Speaker 1: to write some Shakespeare's sauna.
• 25:41
  Speaker 2: But saun it. But how many hundreds of years is
• 25:43
  Speaker 2: it going to take? And so again, if you.
• 25:47
  Speaker 1: Have subject matter expertise, a lot of the interviews that
• 25:50
  Speaker 1: I do are about gathering that primary source. Then you
• 25:54
  Speaker 1: can say, you know, the FBI has identified three cases
• 25:58
  Speaker 1: where this was being done, and here's the root cause,
• 26:01
  Speaker 1: here's the controls that you might use, here's some of
• 26:04
  Speaker 1: the issues you might be looking for. I can build
• 26:07
  Speaker 1: it faster with actual intelligence. And so again I do
• 26:13
  Speaker 1: some deep dives on how I use AI and writing content.
• 26:17
  Speaker 1: There are podcast videos to that I can try to
• 26:19
  Speaker 1: put a link in the description or in the comments.
• 26:22
  Speaker 1: I've also shown you some of the other tools I use,
• 26:24
  Speaker 1: because again I like to be tools agnostic. Notebook LM
• 26:30
  Speaker 1: might be great this month, but it may not be
• 26:31
  Speaker 1: good next month. Do I have the primary sources somewhere
• 26:35
  Speaker 1: to feed a different system?
• 26:37
  Speaker 2: And then even then, what are my outcomes? What are
• 26:40
  Speaker 2: my deliverables?
• 26:42
  Speaker 1: The project management side, AI is not going to manage
• 26:44
  Speaker 1: the project for you in a way that's appropriate for
• 26:47
  Speaker 1: the client's needs and desires. Yes, you could feed AI
• 26:51
  Speaker 1: a contract and you can say, tell me what are
• 26:54
  Speaker 1: the key points, what are the key deliverables, and what's
• 26:56
  Speaker 1: the time expectation. You can have it to an analysis
• 26:59
  Speaker 1: on a game chart. But you can also just use
• 27:02
  Speaker 1: some project management software and outline the project.
• 27:06
  Speaker 2: Again, folks, AI sounds great because.
• 27:08
  Speaker 1: They're marketing it just like websites and dot com was.
• 27:13
  Speaker 1: Eventually it'll drop back and some of the tools will disappear.
• 27:16
  Speaker 1: But ultimately you still need to deliver high value decisions
• 27:20
  Speaker 1: that are going to create key, profitable customers. You still
• 27:23
  Speaker 1: need to create deliverables that deliver on the promises you
• 27:26
  Speaker 1: made for clients. You still need to create persuasive materials
• 27:31
  Speaker 1: so that you can develop your business and gain a
• 27:34
  Speaker 1: portfolio of clients. You still need to have goals and
• 27:37
  Speaker 1: objectives in order to guide the work so that you're
• 27:42
  Speaker 1: using these tools in a way that produces meaningful results,
• 27:46
  Speaker 1: that's focused on specific outcomes, and that ultimately is done
• 27:50
  Speaker 1: in a cost effective manner, and we can help you
• 27:53
  Speaker 1: with this modeling. So in my business development consulting, we
• 27:57
  Speaker 1: do talk about for ten million dollar plus businesses, how
• 28:00
  Speaker 1: do we streamline things. A lot of companies are just
• 28:03
  Speaker 1: dumping employees because AI seems to give them relevant results.
• 28:07
  Speaker 1: And that's because the manager didn't know whether or not
• 28:09
  Speaker 1: the employees are given you irrelevant results or not, and
• 28:12
  Speaker 1: so they're dissatisfied with performance because they are I've talked
• 28:17
  Speaker 1: to some of these folks if they getting a bad
• 28:19
  Speaker 1: response from an employee or a response they don't understand
• 28:23
  Speaker 1: from an employee, getting a bad response or a response
• 28:25
  Speaker 1: they don't understand from AI is the same thing to them,
• 28:28
  Speaker 1: and so there's organizational and structural problems there. Maybe it's
• 28:31
  Speaker 1: the manager that ought to get laid off. We can
• 28:33
  Speaker 1: talk about that another webinar or another podcast, But my
• 28:37
  Speaker 1: point being that is we do have consultations where you
• 28:40
  Speaker 1: start talking about, Hey, what tools are you using, how are.
• 28:42
  Speaker 2: You using these tools?
• 28:43
  Speaker 1: Let's codify some of this, let's extract best practices, Let's
• 28:49
  Speaker 1: help you deliver on those outcomes. Let's let you deliver
• 28:53
  Speaker 1: on those promises you make to clients. Let's help package
• 28:57
  Speaker 1: materials so that you have the primary sources you can
• 28:59
  Speaker 1: use to create marketing campaigns and other content that drives
• 29:03
  Speaker 1: those leads, that drives those sales. If you're interested in that,
• 29:06
  Speaker 1: visit us at www dot inside strategic relations dot com.
• 29:11
  Speaker 2: Otherwise I'm just a hit. I'm a strategic consultant.
• 29:14
  Speaker 1: I hope I've given you some ideas and concepts on
• 29:17
  Speaker 1: how to better use these tools to create better results.
• 29:21
  Speaker 2: But don't ignore.
• 29:23
  Speaker 1: Text analysis, graph analysis, those other types of tools that.
• 29:27
  Speaker 2: Build from the relationship of content.
• 29:30
  Speaker 1: Because if you've got a body of work, we can
• 29:33
  Speaker 1: help you monetize it.
• 29:35
  Speaker 2: There are so many valuable tools today that amplify the
• 29:42
  Speaker 2: quality and value you can deliver to clients. And so
• 29:45
  Speaker 2: if you're a subject matter.
• 29:46
  Speaker 1: Expert, an entrepreneur, or an executive, and you're launching products
• 29:49
  Speaker 1: and you're positioning in a marketplace as an expert, you're
• 29:53
  Speaker 1: in the right place.
• 29:55
  Speaker 2: Thanks for listening. I'd love to hear your comments.
• 29:57
  Speaker 1: Contact me direct at www dot inside strategic relations dot com.
• 30:03
  Speaker 2: I'll see you in the next podcast.