رونوشت 🔗
WEBVTT
00:00:03.240 --> 00:00:13.358
Welcome to Business Mastery with Dawn Kennedy, your quick under forty five minute dose of expert insights and strategies to make a positive impact on your business and life.
00:00:13.358 --> 00:00:14.599
Let's get started.
00:00:14.599 --> 00:00:22.199
Hello and welcome to this episode of the Business Mastery Podcast.
00:00:22.199 --> 00:00:32.960
Today we're going to be talking about AI in the way that we should maybe be thinking about it as a tool and how we're using it to, quote, solve our problems.
00:00:32.960 --> 00:00:34.039
So I've got Leno here.
00:00:34.039 --> 00:00:38.399
We're going to talk about all the things, predictive modeling, what it means.
00:00:38.399 --> 00:00:43.478
And I'm really looking forward to this conversation and making sure that we give people the right way to use the tool.
00:00:43.478 --> 00:00:43.798
Right.
00:00:43.798 --> 00:00:45.640
Or a different perspective on the tool.
00:00:45.640 --> 00:00:47.600
So thank you so much for joining me today.
00:00:47.759 --> 00:00:48.000
Great.
00:00:48.000 --> 00:00:49.039
Thanks for having me.
00:00:49.000 --> 00:00:49.320
All right.
00:00:49.320 --> 00:00:51.560
Can you tell anyone who you are and who you serve?
00:00:51.798 --> 00:00:52.880
My name is Len Ward.
00:00:52.880 --> 00:00:56.158
I am the managing partner and head of AI of Conexus.
00:00:56.158 --> 00:01:04.000
Our focus is on B2B manufacturing, B2B service businesses, and we also work with quite a few law firms as well in the AI space.
00:01:04.000 --> 00:01:06.760
So we wear a full service digital marketing agency.
00:01:06.760 --> 00:01:12.760
We've now morphed over to more of an AI consultancy and AI systems installation company.
00:01:12.959 --> 00:01:13.718
Amazing.
00:01:13.718 --> 00:01:20.599
So you are the perfect person to talk to us about solving problems with AI.
00:01:20.599 --> 00:01:28.799
And when we were first talking, I have notes here about we're using it as search and retrieve, and that's probably not the best use of it.
00:01:28.799 --> 00:01:30.000
So can we just start there?
00:01:30.000 --> 00:01:36.599
What do you see AI being used for to solve your problems rather than just searching and retrieving?
00:01:36.840 --> 00:01:45.040
So I think maybe, maybe I'll re I'll rephrase my answer because I think things have changed a little bit since probably the first time we talked a little while ago.
00:01:45.040 --> 00:01:52.879
So I think the biggest issue we're facing is a lot of companies now don't understand how people are using AI, even though they're using it themselves.
00:01:52.879 --> 00:02:03.250
This is a classic example of when you're doing like digital marketing and you'll talk to a company and they're telling you about how their clients are searching and you're realizing, no, that's not the way there.
00:02:03.250 --> 00:02:04.688
That's not how they're finding your business.
00:02:04.688 --> 00:02:15.889
So what's happening is that most companies don't understand that people are using ChatGPT, Gemini Claude, and they're getting an answer, meaning they're kind of going in and they're solving the problem.
00:02:15.889 --> 00:02:18.610
So they're asking a question, they're working through it.
00:02:18.610 --> 00:02:24.968
They're either showing video of it, showing images to AI, and they're working all the way through from their problem, all the way down to a solution.
00:02:24.968 --> 00:02:30.810
Once you get to a solution at that point, then they may be trying to think about, okay, who am I going to retain or go look for?
00:02:30.810 --> 00:02:38.250
Once I get the solution to many businesses right now, I think are looking to find out how they can get at the top of that funnel.
00:02:38.250 --> 00:02:42.128
Like, okay, well, if they have a problem, how do I know they're going to come to me right away and we're going to solve it.
00:02:42.128 --> 00:02:51.610
We're morphing into a consumer that is going to walk in the front door and knows exactly what their problem is, and knows exactly what the solution is, and wants to know if you can handle that solution.
00:02:51.610 --> 00:02:55.128
There's no longer going to be like a sales cycle or a funnel.
00:02:55.128 --> 00:03:05.810
You're going to work them through that lived in the search and retrieve world to solve my problem world is you have a vastly more intelligent individual showing up at your law firm, at your business, whatever it may be.
00:03:05.810 --> 00:03:09.050
And they're going to be ready to kind of say, okay, this is what I need to do.
00:03:09.050 --> 00:03:12.530
So I don't know if that made any sense, but that's kind of where we are right now.
00:03:12.370 --> 00:03:18.288
Yeah, absolutely does, because I think it's important if we're going to have consumers coming in believing.
00:03:18.288 --> 00:03:25.810
And when I say believing, I mean they really do believe that's going to be the best solution because they've done all the research.
00:03:25.810 --> 00:03:33.449
I think it would be a mistake to brush off their AI research, or we would know better than the computer.
00:03:33.449 --> 00:03:36.729
I've seen some resistance, especially in law practice.
00:03:36.729 --> 00:03:39.330
I mean, let's be honest, and there's a lot of regulatory things there.
00:03:39.330 --> 00:03:44.169
And whether or not you feed all your stuff into chat, if it's discoverable or whatever.
00:03:44.169 --> 00:03:55.770
But the truth is, when we first meet with these people who are going to want their problems solved, we need to actually have a different perspective on.
00:03:55.770 --> 00:03:57.770
They looked, they went and looked.
00:03:57.770 --> 00:04:00.770
And this is what they've sort of drilled down to.
00:04:00.770 --> 00:04:03.729
And we need to be a little more receptive that they're not coming in as cold.
00:04:03.729 --> 00:04:11.050
Is that sort of accurate that we need to have some perspective that this research that
Load full transcript
they're doing is actually good for us.
00:04:11.090 --> 00:04:13.169
Yeah. No, no, it is.
00:04:13.169 --> 00:04:24.528
And because I think law firms who really have generating leads and so forth have lived in the search and retrieve world where literally the entire job of a marketing firm was just to get the lead in the front door.
00:04:24.528 --> 00:04:27.088
And then the lawyer, the intake team would qualify it.
00:04:27.088 --> 00:04:29.250
Or if it's a smaller firm, they will qualify it.
00:04:29.250 --> 00:04:35.088
And then the attorneys assume, okay, we're here to solve these people's problems and we're going to charge an hourly rate for that.
00:04:35.088 --> 00:04:36.490
And that model has been changed.
00:04:36.490 --> 00:04:36.810
Yes.
00:04:36.810 --> 00:04:42.850
You have to don't assume that more than ChatGPT or more than Gemini, because you don't anymore.
00:04:42.850 --> 00:04:47.930
I think as humans, we all have to have a come to Jesus meeting to realize these things are vastly smarter than us.
00:04:47.930 --> 00:04:48.810
They just are.
00:04:48.810 --> 00:04:51.970
And if you try to fight that, I think that's where your problem is going to be.
00:04:51.970 --> 00:05:05.259
What is going to help you as a law firm is yes, really being understanding of the individuals walking in the front door, understanding that they know the knowledge you have and that you're solely focused on their problem and how you're going to solve that going forward.
00:05:05.259 --> 00:05:07.379
It's going to destroy billable hours.
00:05:07.379 --> 00:05:10.019
Yes, it's probably going to reduce some headcount at law firms.
00:05:10.019 --> 00:05:21.338
I know this is all the stuff we don't want to talk about, but it's such a seismic shift that I'm I am seeing quite a few law firms that are still stuck in a search and retrieve mentality, and I don't think they understand how fast this thing is moving.
00:05:21.338 --> 00:05:24.338
And considering they probably spend more money on marketing than anybody.
00:05:24.338 --> 00:05:24.579
Yeah.
00:05:24.579 --> 00:05:29.139
Right now, stop the bleeding and really understand how intelligent this individual is walking through the front door.
00:05:29.139 --> 00:05:32.379
And if their intelligence is escaping them, they're going to pull their phone out.
00:05:32.379 --> 00:05:36.699
They're going to open up the app and they're just going to have the app listen to you, ChatGPT.
00:05:36.699 --> 00:05:39.540
And now you're literally put right on the spot as the law.
00:05:39.540 --> 00:05:40.778
Now lawyers may say, put that down.
00:05:40.778 --> 00:05:42.300
I don't want that in there, but.
00:05:42.540 --> 00:05:43.860
We don't want it discoverable, right?
00:05:43.860 --> 00:05:47.180
Like our conversations privileged and we don't want we don't want your.
00:05:47.180 --> 00:05:47.420
Yeah.
00:05:47.420 --> 00:05:49.220
And that's kind of an important point for lawyers.
00:05:49.220 --> 00:05:59.860
But there are other industries as well, even when in some of the products, businesses and things that people are coming in with ideas of what they want based on the searches that they're doing.
00:05:59.860 --> 00:06:07.860
So as a business, how are we setting ourselves up to receive the fact that they are not walking in with a blank piece of paper?
00:06:07.860 --> 00:06:08.338
Right.
00:06:08.338 --> 00:06:17.459
Especially, and I see this across maybe some industries like event planning and those kinds of things where people are going to be coming in and they're going to know pretty much what they want or what they think they want.
00:06:17.459 --> 00:06:34.420
So again, to your point, the sales cycle probably is going to change, but how do we optimize on our side to be able to predict or know what we think more people will be coming from, besides the things like industry trends, right?
00:06:34.420 --> 00:06:35.060
Everything.
00:06:35.060 --> 00:06:43.620
If it's one person or two people, it may be anecdotal that there's a new service or a new product or something that we should be looking into.
00:06:43.620 --> 00:06:56.819
But when do we know that this is the recommendation that's being made at a bigger scale, and that we should be paying attention to Chad's recommendations or the AI's recommendations as we're making our strategic moves.
00:06:57.540 --> 00:07:00.860
So I'll answer that by going to the prequel.
00:07:00.860 --> 00:07:19.338
If this was like a movie type thing, and I think to get to where you want to be, you companies have got to change their entire way they conduct business to solving somebody's problem, to understanding that everything we do is we are now going to solve my problem world, and I'll step back and describe it, search and retrieve.
00:07:19.338 --> 00:07:28.699
For the last twenty five years, people would go to Google, they would go to yahoo, believe it or not, back then, Netscape and all that, and they would search for solutions to their problem.
00:07:28.699 --> 00:07:38.699
And we were trained that they would read blogs, they would watch videos, they would maybe do a webinar, and it would really help educate them as to what their options are.
00:07:38.699 --> 00:07:50.740
They would land on your website and we would all say, okay, and they'd gather all the information, they'd sit around a table, maybe print some stuff out, or just look at it on their phone and say, okay, this is where I think maybe my direction is going to be.
00:07:50.740 --> 00:07:52.500
And every business was built like that.
00:07:52.500 --> 00:07:57.579
So it was always you'd call in and the sales team would pull up their CRM and give them a little bit of information.
00:07:57.579 --> 00:08:00.300
Customer service would say, okay, let's, let's walk through this.
00:08:00.300 --> 00:08:09.060
So and basically the entire job of companies was to literally just provide a bunch of information and say, okay, well, we'll help you work through it,
but here's all your information.
00:08:09.060 --> 00:08:14.899
Now solve my problem is individuals are just literally coming in and they know exactly what they're looking for.
00:08:14.899 --> 00:08:18.459
They're not looking at twenty five blogs or looking at twenty five videos.
00:08:18.459 --> 00:08:19.579
They're asking a question.
00:08:19.579 --> 00:08:21.778
ChatGPT is giving them an answer right away.
00:08:21.778 --> 00:08:24.459
So that fundamentally changes everything you have to do.
00:08:24.459 --> 00:08:26.379
It's going to change your website content.
00:08:26.379 --> 00:08:28.540
It's going to change the way your sales process is.
00:08:28.540 --> 00:08:29.860
It's going to change everything.
00:08:29.860 --> 00:08:33.860
So how do you optimize for it for something like that to be ready?
00:08:33.860 --> 00:08:39.220
I think it literally starts with really changing your business model from we no longer search and retrieve.
00:08:39.220 --> 00:08:40.940
We're now solved my problem.
00:08:40.940 --> 00:08:49.750
So your quarterly meetings, your annual meetings, or even your daily meetings should be along the lines of what are the typical bottlenecks we have to solving our clients problems?
00:08:49.750 --> 00:08:52.750
What is stopping this law firm from closing more cases?
00:08:52.750 --> 00:08:54.470
Are we not providing enough information?
00:08:54.470 --> 00:09:01.870
If we're a B2B manufacturer, what are the bottlenecks that engineers or architects are running into and we're not solving their problems.
00:09:01.870 --> 00:09:11.908
What typically where a business process from beginning to end for to grab a lead all the way to a to the conclusion where the sale that may take however many days, weeks or months.
00:09:11.908 --> 00:09:16.590
Now your consumer is going to expect that to happen within minutes and it's there.
00:09:16.590 --> 00:09:20.710
So there's a seismic shift that businesses have to make and a broader conversation.
00:09:20.710 --> 00:09:27.710
I won't get too far off topic where everybody's using AI on their speeding the process up, but all you're doing is speeding up the search and retrieve world.
00:09:27.710 --> 00:09:30.028
You're not doing anything to implement solve my problem.
00:09:30.028 --> 00:09:31.308
So I rambled a lot.
00:09:31.308 --> 00:09:33.028
Hopefully some of it made sense.
00:09:33.070 --> 00:09:34.308
Yeah, it does make sense.
00:09:34.308 --> 00:09:41.428
So where the model started before we can just use law firms if you want, because I think it's a pretty, pretty straightforward thing, right?
00:09:41.428 --> 00:09:46.750
We get to a decision point, take the case, don't take the case and then that isn't right.
00:09:46.750 --> 00:10:03.590
So there's there's less of the intake step is what you're saying that the time that you would have to educate people on their case, on their options, on their different strategic opportunities that may be available to them.
00:10:03.590 --> 00:10:08.629
So how do we change that upfront so that we also meet that consumer expectation?
00:10:08.629 --> 00:10:15.269
Because I would say that there are several industries that might be like, yeah, but they're coming in saying this is what they want.
00:10:15.269 --> 00:10:21.308
And then when I'm looking at these things, I'm still feeling very unsure, just giving them what they say they want.
00:10:21.308 --> 00:10:28.028
Because you're saying if we start to try to sell again or do consultations, or is there a way that we have to do that?
00:10:28.028 --> 00:10:31.389
Imagine you want to build an addition to a building or something like that.
00:10:31.389 --> 00:10:49.548
If they're coming in with ninety percent of sort of what they think they want, how do we restructure our business to meet their expectations while also making sure that we are doing what we need to do as his businesses to protect our reputations, protect our margins, but then also to protect these highly regulated industries.
00:10:50.870 --> 00:10:54.710
So I think that's like a multi-pronged question and I'll try to answer that.
00:10:54.548 --> 00:10:55.590
Sorry about that.
00:10:55.590 --> 00:10:56.428
No that's fine.
00:10:56.428 --> 00:10:58.668
Yeah that's fine.
00:10:58.668 --> 00:11:05.710
So I think what's going to happen is and I'll kind of bake all this together in one, one answer and see if this kind of makes sense.
00:11:05.710 --> 00:11:23.830
But we're not far off from the days of agent to agent marketing agent to agent bookings, which means if I have a legal case, I'm thinking about divorce, or I'm thinking about a lawsuit because I was hit by a UPS driver, or I'm thinking about a worker's compensation claim, I'm just going to kind of we'll do the regular attorneys and not get into the business type attorneys.
00:11:23.830 --> 00:11:35.190
But if I'm thinking about something along those lines, it's going to be very soon where the app on my phone, the wearable device, which I think is coming, or a foldable device, there's actually talk that the phone may go away and that they're trying to.
00:11:35.190 --> 00:11:40.028
I know it sounds crazy, but they're talking about various different apparatuses that we would have.
00:11:40.028 --> 00:11:41.918
That agent is going to know so much about us.
00:11:41.918 --> 00:11:53.278
It's going to know every single thing that we know, what we like, where we like to drive, what we want to do, the people we kind of like to interact with, the brands we like to look at online, that it's actually going to go out and market to another law firm's agent.
00:11:53.278 --> 00:12:00.320
And that agent within their website or within their repository is going to be literally doing a to a mark, a AI to AI marketing.
00:12:00.320 --> 00:12:01.558
None of us know how that looks.
00:12:01.558 --> 00:12:02.759
I know some people are trying.
00:12:02.798 --> 00:12:11.360
I just yeah, I just got the heebie jeebies a little bit that
they're, they're going to intercept human marketing using agents or we're going to use agents as well.
00:12:11.360 --> 00:12:12.879
Is that what you're saying on both sides?
00:12:13.080 --> 00:12:15.200
There'll be two lanes of marketing in my opinion.
00:12:15.200 --> 00:12:16.399
There'll be human to human marketing.
00:12:16.399 --> 00:12:18.240
And then there'll be AI to AI marketing.
00:12:18.240 --> 00:12:24.240
And I think we're going to get more and more reliant on a to a AI to AI marketing as the years and months go by.
00:12:24.240 --> 00:12:25.759
I honestly believe that.
00:12:25.759 --> 00:12:35.440
So what will happen is that they'll know AI is going to know about lawsuits, or you're thinking about doing something and it's going to present stuff to you, whether it's a text, whether it's a phone call.
00:12:35.440 --> 00:12:39.240
Because remember, a lot of AI's voice and it's going to say, these are the two law firms I like.
00:12:39.240 --> 00:12:47.720
I'm going to do a little bit more research for you and I, because I think these are the law firms where I did some back channel and I really looked they're really good against UPS lawsuits like they have.
00:12:47.720 --> 00:12:50.278
They've done some stuff and they're kind of right up your alley.
00:12:50.278 --> 00:12:51.960
They're only about a mile from the house.
00:12:51.960 --> 00:12:53.399
And I think you would really like them.
00:12:53.399 --> 00:12:55.158
Like that's how it's going to work.
00:12:55.158 --> 00:12:56.798
I know that sounds crazy.
00:12:56.798 --> 00:13:00.038
So for law firms, like what, what do I do to get ready for that?
00:13:00.038 --> 00:13:02.759
Well, the first thing law firms have to do is you have to start building your brand out.
00:13:02.759 --> 00:13:09.519
And I don't mean ego branding, I don't mean putting billboards up or they don't need to be or flying banners or they don't need to be because it's an ego thing.
00:13:09.519 --> 00:13:14.000
Sorry, lawyers, I've been you've been my client for thirty years and yeah, that's how you operate.
00:13:14.000 --> 00:13:17.120
This is more you have to start doing like things we're doing right here.
00:13:17.120 --> 00:13:18.678
You have to start getting on podcast.
00:13:18.678 --> 00:13:20.278
You have to start clipping that up.
00:13:20.278 --> 00:13:22.399
You have to start doing, going on TikTok.
00:13:22.399 --> 00:13:26.918
I know you don't want to, but your demographic is racing towards TikTok, your age group.
00:13:26.918 --> 00:13:31.440
You have to start thinking about Instagram and maybe putting some video out there or a couple little clips and images.
00:13:31.440 --> 00:13:48.879
So the individual is going to from AI have all their answers, and then they're going to live where they all live, which is social media, and they're going to go into your social media, they're going to start looking at these types of things that is going to be way more important than you can ever think than just ranking number one for saying divorce lawyer near me.
00:13:48.879 --> 00:13:52.558
We'll get you the most out of your lawsuit or something along those lines.
00:13:52.558 --> 00:14:04.720
And that's while that was the old search retrieve world, the new solve my problem world is AI is going to arm the individual with every piece of information they want, and it's going and they're going to show up at your front door, and it's going to come back to the famous human element.
00:14:04.720 --> 00:14:06.879
Do I or do I not want to work with Don?
00:14:06.879 --> 00:14:07.720
Do I like Don?
00:14:07.720 --> 00:14:09.000
Do I not like Don?
00:14:09.000 --> 00:14:10.798
I mean, do I think she can help me out?
00:14:10.798 --> 00:14:23.639
Does she understand my situation with divorce or a lawsuit and house that they're even going to get to that front door is by looking at all your social understanding how you're a little micro influencer and saying, you know what, I really like what Don has to say.
00:14:23.639 --> 00:14:26.038
And I really hope she's like this in person as well.
00:14:26.038 --> 00:14:38.250
So it's a really weird, strange new world that we're heading into, especially with marketing, because I don't think there's anything more important in my opinion, than then marketing generating leads for a law firm.
00:14:38.250 --> 00:14:39.490
I don't think that's the lifeblood.
00:14:39.490 --> 00:14:41.090
If you don't have it, you're dead.
00:14:41.090 --> 00:14:43.450
And I think this is something you really have to pay attention to.
00:14:43.450 --> 00:14:48.370
So you, more than anybody, have to worry about optimizing and being prepared for AI answers.
00:14:48.370 --> 00:15:01.330
And that means you interacting and asking questions like pretending you're an actual person across the desk and on the right hand side, get yourself a microphone, get yourself a camera, get on podcast, start talking because people are going to expect it.
00:15:01.928 --> 00:15:02.928
Yeah, I love that.
00:15:02.928 --> 00:15:09.570
It's so interesting that the AI, the way I'm thinking about it, is like the humans on the top layer, you have to get the brand out there.
00:15:09.570 --> 00:15:16.370
But ultimately, at the end of the day, it may be an agent that's making that recommendation to your ideal client.
00:15:16.370 --> 00:15:26.649
And I think that, at least from my perspective, I'm sitting here like, wow, it feels a little weird that you're not going to have to do the top of funnel thing.
00:15:26.649 --> 00:15:41.490
But also, I think, and I've been very vocal about the idea of what is your conversion rate going to look like when you are receiving people who are not kicking tires, if they know you a little bit and they're going to probably convert, right.
00:15:41.490 --> 00:16:05.690
So even though we may not be comfortable not knowing what it's going to look like, I did see a video and I'm pretty sure quite a few people saw the video where I guess
an agent and an agent got on the phone to one was a personal assistant, one was a travel agency, and they got on the phone together and booked the guy's travel like he had an agent get up to book travel, which seems really advanced from where I'm sitting.
00:16:05.690 --> 00:16:22.690
I'm not a huge tech person, but it's an interesting thing to think about on the one side that the marketing may not necessarily be quote unquote easier or less expensive, but the conversion rate will probably be higher and faster.
00:16:22.769 --> 00:16:23.528
Oh yeah.
00:16:23.370 --> 00:16:24.850
The people who come through the door.
00:16:24.850 --> 00:16:29.808
So the good side of this, yes, you have to put yourself out there and get your brand out there.
00:16:29.808 --> 00:16:38.250
But what are your thoughts on how they're going to sift through and send you maybe people that like to your point, they're going to walk in the door and they're going to be ready to convert.
00:16:38.490 --> 00:16:43.009
Oh, I think the conversion rate is going to go up to ninety, ninety or ninety, ninety five percent.
00:16:43.009 --> 00:16:45.490
I don't think there's any question they're arriving because of that.
00:16:45.490 --> 00:16:46.769
And you made a good point.
00:16:46.769 --> 00:16:48.808
AI is going to wipe out the top of the funnel.
00:16:48.808 --> 00:16:53.330
AI is going to wipe out half of the middle of the funnel, the other half of the middle funnel that's going to happen.
00:16:53.330 --> 00:16:58.288
That's where social media picks up and where, and that's going to run all the way down to the bottom of the funnel.
00:16:58.288 --> 00:17:10.608
At the very last sliver, we know that actual conversion part of the funnel, the very small one percent of it, that's where them showing up at your front door, coming into your office or some having coffee and meeting them and so forth.
00:17:10.608 --> 00:17:12.890
That's where the conversion is really going to happen.
00:17:12.890 --> 00:17:18.808
I don't know if that answered your question, but that's like where I it's I just think we have to really start framing it that way.
00:17:18.808 --> 00:17:20.210
Where does advertising work in on this?
00:17:20.210 --> 00:17:22.808
Because somebody's going to say, well, leads are always going to be there.
00:17:22.808 --> 00:17:26.690
Yes and no. I do think lead generation is going by the wayside.
00:17:26.690 --> 00:17:35.980
So if you're a law firm for the last five years who have built your entire practice on really just having an amazing lead funnel, like you just knew how to generate leads.
00:17:35.980 --> 00:17:43.940
But if you were to go ask the average individual who is Mike Jones law firm, and they don't know who you are, you're the lawyer that's going to have a massive problem in about eighteen months.
00:17:43.940 --> 00:17:49.858
So you may think you're riding high right now, but that's not going to last because Google's even acknowledged pay per click.
00:17:49.858 --> 00:17:52.058
All this is all beginning to come down.
00:17:52.058 --> 00:17:54.618
And I know that there's going to be ads in ChatGPT and so forth.
00:17:54.618 --> 00:17:58.818
I get it, but I don't think they're going to be the ads that people think it's not going to be click here, click here.
00:17:58.818 --> 00:18:06.019
It's because again, that search, retrieve and ChatGPT, Claude and Gemini have all been crystal clear.
00:18:06.019 --> 00:18:08.578
They don't want you leaving their infrastructure.
00:18:08.578 --> 00:18:10.500
They want you staying within the infrastructure.
00:18:10.500 --> 00:18:12.500
So this isn't clicking out to a website.
00:18:12.500 --> 00:18:14.940
It's how do you provide the answers right in there.
00:18:14.940 --> 00:18:17.700
So that's kind of what we're looking at with that.
00:18:17.980 --> 00:18:30.180
Yeah, we did a few episodes at the beginning of this year on a e t and how you might have to start thinking about your website and how you can start to become recommended.
00:18:30.180 --> 00:18:35.980
You're saying within eighteen months, which is faster actually than we we thought of like six months ago.
00:18:35.980 --> 00:18:49.818
My question to you is, besides things like the website stuff and the a e t and changing your SEO to be found and making sure you have the FAQs, you're saying humans need to get out into the world, put your brand out there.
00:18:49.818 --> 00:18:56.098
And whether or not AI recommends you, would that still give you another avenue or another option?
00:18:56.098 --> 00:19:02.058
Will they recommend things like maybe a website or is it because it's predictive modeling?
00:19:02.058 --> 00:19:02.460
Right.
00:19:02.460 --> 00:19:08.098
It's the the models are trying to, I guess, predict how to solve the problem.
00:19:08.098 --> 00:19:10.019
What else do we need to be doing?
00:19:09.818 --> 00:19:11.058
Because that's.
00:19:11.058 --> 00:19:12.298
No, I know what you're saying.
00:19:12.298 --> 00:19:15.740
There's no this again, goes back to search and retrieve.
00:19:15.740 --> 00:19:15.940
Search.
00:19:15.940 --> 00:19:18.098
Retrieve is one hundred percent built on an algorithm.
00:19:18.098 --> 00:19:19.500
It was built on an algorithm of.
00:19:19.500 --> 00:19:21.700
The algorithm historically was always.
00:19:21.700 --> 00:19:23.778
This is an individual looking for something.
00:19:23.778 --> 00:19:29.940
And I'm going to put up a few different companies because I think this information is probably right for them.
00:19:29.940 --> 00:19:32.420
And we're going to all the people up looking for.
00:19:32.420 --> 00:19:34.220
I was an Uber car hit me.
00:19:34.220 --> 00:19:35.420
How do I file a lawsuit?
00:19:35.420 --> 00:19:37.858
Or can I sue something along those lines?
00:19:37.858 --> 00:19:41.140
And the algorithm was built for anyone looking for Uber cars.
00:19:41.140 --> 00:19:44.220
They didn't care what the person looked like or what they were
or what their interests were.
00:19:44.220 --> 00:19:48.858
It just was like you were an Uber, Uber car hit you or you were in an Uber car and it got hit.
00:19:48.858 --> 00:19:51.538
These are the law firms that provide the best information for you.
00:19:51.538 --> 00:19:52.700
That was the algorithm.
00:19:52.700 --> 00:19:59.019
The algorithm then morphed to where Google started identifying and even social media started identifying.
00:19:59.019 --> 00:20:04.220
You started looking at the other websites you're on, and then tried to start pairing you up with different types of information.
00:20:04.220 --> 00:20:09.538
But yet the information still coming back was information to read a blog and do certain things.
00:20:09.538 --> 00:20:11.940
AI is a very, very different process.
00:20:11.940 --> 00:20:13.538
It's I don't people call it an algorithm.
00:20:13.538 --> 00:20:14.460
I don't know if you want to call it.
00:20:14.460 --> 00:20:15.660
It's not an algorithm.
00:20:15.660 --> 00:20:21.900
It works on a tokenized process, which means if I tell you Mary had a little, you would think lamb.
00:20:21.900 --> 00:20:30.338
So what it's trying to do is it's trying to do predictive modeling in front of you where it thinks the right answer is you're not going to have an algorithm that's going to change what it thinks.
00:20:30.338 --> 00:20:30.700
The right.
00:20:30.700 --> 00:20:41.259
There's no outside influences that's going to change its answer to say, Mary had a little lamb, it's not going to be Mary had a little goat because I did enough backlinking here and I wrote enough blogs.
00:20:41.259 --> 00:20:44.660
And the algorithm now believes Mary definitely had a little goat.
00:20:44.660 --> 00:20:46.259
That's not how it works.
00:20:46.259 --> 00:20:55.858
So people have to like, get rid of the algorithm aspect of things, and they have to understand that it's predictively trying to answer questions that's specifically geared towards you.
00:20:55.858 --> 00:20:58.420
So you're not going to be able to optimize for these things.
00:20:58.420 --> 00:21:04.500
In my opinion, you're not going to be able to try to interject and say, hey, we're going to be able to answer their questions better than anybody else.
00:21:04.500 --> 00:21:06.058
That's not how it's going to work.
00:21:06.058 --> 00:21:16.858
I think you're going to be part of the gigantic ecosystem where your content and your and what you're talking about, more importantly on social than the written word anymore is going to be factored in.
00:21:16.858 --> 00:21:22.460
And in ChatGPT or whichever AI you're using is going to say, hey, Dawn's law firm.
00:21:22.460 --> 00:21:30.269
Really, this was they might kind of put some little things there to say, hey, this is a law firm that kind of helped us give us answer, but even that's going to go away.
00:21:30.269 --> 00:21:37.308
I think to the point where you're going to go in and chat GPT, if you're in there, you're going to pretty much know what the ninety nine percent accurate.
00:21:37.308 --> 00:21:40.588
What it's telling me is pretty accurate, and I don't need to know where it came from.
00:21:40.588 --> 00:21:43.588
I just want the answer because I'm not going to go read Dawn's blog.
00:21:43.588 --> 00:21:45.230
I'm not going to get more information.
00:21:45.230 --> 00:21:45.750
Makes sense.
00:21:45.750 --> 00:21:51.950
So I think, yeah, there's so many snake oil salesmen out there selling like, let's rank within ChatGPT and do all that.
00:21:51.950 --> 00:21:57.950
And it's like, when I hear that, I'm like, oh my God, you literally are literally missing what they're all telling us.
00:21:57.950 --> 00:22:02.150
And I, I get feel bad because law firms and everybody else are just paying money for this stuff.
00:22:02.150 --> 00:22:03.868
And it's like, you're being lied to.
00:22:04.028 --> 00:22:04.630
Right?
00:22:04.470 --> 00:22:04.868
Yeah.
00:22:04.670 --> 00:22:05.509
It's important.
00:22:05.509 --> 00:22:06.750
It's important information.
00:22:06.750 --> 00:22:08.269
This is evolving as well.
00:22:08.269 --> 00:22:12.190
Like you said, we don't know what this is maybe going to look like with a wearable later or whatever.
00:22:12.190 --> 00:22:22.588
But what has me thinking here is I have read some articles and I have looked at some things where AI always wants to solve your problem, and they always want to give you a positive sort of spin on it.
00:22:22.588 --> 00:22:26.308
Like it doesn't want to tell you necessarily that you're wrong.
00:22:26.308 --> 00:22:28.910
And it won't necessarily always push back.
00:22:28.910 --> 00:22:32.788
It'll say, we can see that point, but what about this over here?
00:22:32.788 --> 00:22:39.990
I was reading an article about how people like AI because it's in some ways a little affirming, right?
00:22:39.990 --> 00:23:09.308
So you put your your thing in there and I think it's this and then it comes back and it might affirm you a little bit rather than ask another stream of questions, or it'll give you an answer and then give you questions at the bottom, maybe to refine things a little bit about the quality of prompts and things that people are using, is there some education that we should be tuned into, or things where we can look at the quality of the prompts and how maybe we can be better prepared?
00:23:09.269 --> 00:23:09.950
Yeah.
00:23:10.308 --> 00:23:11.150
To serve.
00:23:11.150 --> 00:23:12.108
Yeah, I know what you're saying.
00:23:12.108 --> 00:23:17.750
It comes across almost like it's like a caring mother or father opposed to criticizing you where it needs to criticize you.
00:23:17.588 --> 00:23:19.788
Okay, so I'm not crazy reading these articles.
00:23:19.788 --> 00:23:20.788
That is true, right?
00:23:20.788 --> 00:23:23.190
That is, I heard something about some.
00:23:23.190 --> 00:23:25.150
I haven't dug into it, so I don't want to go too far.
00:23:25.150 --> 00:23:26.670
But it's
something about.
00:23:26.670 --> 00:23:28.588
I think they were calling it AI psychosis or something.
00:23:28.588 --> 00:23:29.509
One of those.
00:23:29.509 --> 00:23:30.068
Yeah.
00:23:30.068 --> 00:23:30.470
Okay.
00:23:30.470 --> 00:23:41.308
Can you elaborate then a little bit on this, and let's make sure that we put those guardrails up as well so that people don't, like you said, go all in, put their cash on fire without understanding exactly what's involved.
00:23:41.269 --> 00:23:52.308
So one of the best prompts you can do, you can throw in there, and I'll touch on prompts in a minute where I think that's headed, but is literally say, challenge me on everything I'm asking you criticize me where I need to be criticized.
00:23:52.308 --> 00:24:00.990
Ask me questions before you give an answer and say, I'm not looking for, are you looking for validation or say, I'm not looking for.
00:24:00.990 --> 00:24:01.828
Tell it what you want.
00:24:01.828 --> 00:24:02.950
I'm not looking for validation.
00:24:02.950 --> 00:24:04.150
I'm looking for clarity.
00:24:04.150 --> 00:24:15.788
So if you go in with those prompts and the power users per say, people that are like living, breathing, like building stuff on it, they know right away they go right in and say, treat me like dirt.
00:24:15.788 --> 00:24:19.230
Like just, I want to know, like, I want to know like where I'm at.
00:24:19.230 --> 00:24:22.920
But that's a good thing because it is challenging you now.
00:24:22.920 --> 00:24:29.078
I do believe as these things get to know you better, what I find amazing is Claude, which I've really gravitated towards.
00:24:29.078 --> 00:24:32.598
Claude just is really picking up on every single thing I'm doing.
00:24:32.598 --> 00:24:34.160
And I've had Claude.
00:24:34.160 --> 00:24:41.400
I use Claude all the time to build products, and a lot of times I get frustrated because the code I'm trying to bring over or what I'm trying to build is not working.
00:24:41.400 --> 00:24:46.960
Claude will push back on me because it's beginning to get know, get to know me where I get a little angry.
00:24:46.960 --> 00:24:48.880
And at Irish Fire hits pretty quick.
00:24:48.880 --> 00:24:51.000
But and it challenges me and I like that.
00:24:51.000 --> 00:24:57.240
So yeah, I do have a concern where ChatGPT in the past was and was telling you what you wanted to hear.
00:24:57.240 --> 00:25:02.240
I think sometimes and some of the answers about a year ago, you look at them like, that's not only not wrong.
00:25:02.240 --> 00:25:05.118
That's I think that's a little dangerous that that answer.
00:25:05.118 --> 00:25:12.118
I think we're beginning to move away from that as the models are beginning to, to tokenize processes getting better and better and better.
00:25:12.118 --> 00:25:13.519
Meaning it Mary had a little lamb.
00:25:13.519 --> 00:25:16.519
They're hitting on that way more often than not.
00:25:16.519 --> 00:25:20.920
Whether letting somebody know that
Mary had a little lamb is offensive or not.
00:25:20.920 --> 00:25:23.720
It's starting to get to that point of moving there.
00:25:23.720 --> 00:25:26.960
So yeah, what you're the point you're raising is definitely an issue.
00:25:26.960 --> 00:25:34.240
And my opinion, my advice to anybody is literally before you put a prompt in, if you're trying to work through a problem, say, I need you to challenge me on this all the way through.
00:25:34.240 --> 00:25:35.038
And it will.
00:25:35.598 --> 00:25:36.000
All right.
00:25:36.000 --> 00:25:37.278
So now let me ask you this.
00:25:37.278 --> 00:25:49.640
Do you think that there'll be more consumer education or more education to people at large so that when they come in and they're pretty sure they know what has to happen?
00:25:49.640 --> 00:25:53.960
And the reason I bring this up is there was recently a case and it's moving through there.
00:25:53.960 --> 00:25:56.680
There's something like three hundred thousand dollars in fees involved.
00:25:56.680 --> 00:25:59.038
Here's this is not a small thing.
00:25:59.038 --> 00:26:05.598
Somebody's basically fed a court document into ChatGPT and they said, oh no, no, no, you can get that set aside.
00:26:05.598 --> 00:26:06.278
You can go.
00:26:06.278 --> 00:26:08.118
You don't need to stay with that arbitration agreement.
00:26:08.118 --> 00:26:09.720
Here's how you're going to go do it.
00:26:09.720 --> 00:26:13.720
And this person was pro per and they filed and filed and filed.
00:26:13.720 --> 00:26:18.078
And the thing is, they were sure one hundred percent of the time the chat was correct.
00:26:18.078 --> 00:26:28.318
Now again, this is going back a little over a year or whatever, but we need to probably also educate, especially in the B2B space when they come up with the solution.
00:26:28.318 --> 00:26:31.200
Have they asked to be challenged?
00:26:31.200 --> 00:26:32.838
Have they done that?
00:26:32.838 --> 00:26:37.798
I'm going to call it challenged diligence, I guess, or push back diligence.
00:26:37.798 --> 00:26:47.920
Is there something that we could also be doing as a business, not just asking it to challenge us or to find gaps, or asking us to tear something apart again, from the business side.
00:26:47.920 --> 00:26:52.558
And you're saying we're going to have to expect the clients are coming in the door and they're really going to know what they want.
00:26:52.558 --> 00:27:04.160
Do we need to have some perspective or maybe some thought about whether or not this is chat giving them the answer they want versus having it pressure tested by the other side first?
00:27:04.400 --> 00:27:05.278
That's a good question.
00:27:05.278 --> 00:27:18.650
I think that's what the jobs of the attorneys are going to be, of what the sales team is going to be of the consultant or whoever it may be, even doctors to a point where you have to understand what somebody is handing over to you, the research they're handing over.
00:27:18.650 --> 00:27:26.328
Hopefully they're either saying, if you are going to use ChatGPT, in my opinion, if you're going to go heavily on that and you're going to sit down with an attorney, print out the answers, right?
00:27:26.328 --> 00:27:27.490
Don't just say what it is like.
00:27:27.490 --> 00:27:28.250
Attorneys are literal.
00:27:28.250 --> 00:27:28.890
They need to look at it.
00:27:28.890 --> 00:27:30.650
They're not going to go on hearsay type thing.
00:27:30.650 --> 00:27:34.730
So I think the job's going to be okay.
00:27:34.730 --> 00:27:38.130
I tell this all the time, your job is going to be to pilot these AI systems.
00:27:38.130 --> 00:27:40.769
And I don't know if the humans are ever not going to be the pilots.
00:27:40.769 --> 00:27:44.568
So I know I love how Microsoft branded the term copilot because I know exactly what they're doing.
00:27:44.568 --> 00:27:45.890
It's great what they did.
00:27:45.890 --> 00:27:53.690
So when an individual comes in is like, all right, my ChatGPT or my Claude or whatever, my Gemini kind of this is where it kind of took me to.
00:27:53.690 --> 00:27:56.650
You're going to have to be the educated pilot to say, that's right.
00:27:56.650 --> 00:28:00.848
However, if you keep going down this road, this is what's going to happen.
00:28:00.848 --> 00:28:05.848
Because the thing that I don't necessarily know these models are ever going to have is rationale.
00:28:05.848 --> 00:28:13.730
So your true leverage as a lawyer or great salesperson or consultant, whatever is going to be, you know, how to rationale something and you have experience.
00:28:13.730 --> 00:28:17.450
So yes, it may come back and say something, but it doesn't.
00:28:17.450 --> 00:28:24.170
So I think that's where the intelligence ends on the these models, these models ends when the human comes in with their rationale and comes in with their experience.
00:28:24.170 --> 00:28:28.048
I think because we all know everything is great until something completely changes.
00:28:28.048 --> 00:28:33.288
So just because ChatGPT is predicting this, life has a funny way to say, hey, this is going to happen now.
00:28:33.288 --> 00:28:36.528
So I think that's going to be the main job you have to do.
00:28:36.528 --> 00:28:40.690
So when an educated consumer comes in, respect the models, respect that.
00:28:40.690 --> 00:28:42.368
These things are really intelligent.
00:28:42.368 --> 00:28:52.608
But I also think consumers should be, to your point, educated to respect the individual humans you're going in to have a conversation with and to try to get these things to work together.
00:28:52.608 --> 00:28:58.288
Now that goes to the job thing that's going to eliminate this person that because you're going to go right to the attorney and that's what people are going to want.
00:28:58.288 --> 00:29:08.690
So attorneys are you're not going to be hiding behind walls too much anymore with a lot of people, like before the client gets there, you're going to
unfortunately, even the large law firms, when you have meetings, you're going to be even clients for the first time.
00:29:08.690 --> 00:29:10.490
And that's just kind of how it's going to have to be.
00:29:10.490 --> 00:29:15.650
But yeah, the thing we like to use a lot at Connexus is AI generated, human approved.
00:29:15.650 --> 00:29:20.730
We tell clients all the time, even if we generate something with AI, you have got to put your thumbprint on it.
00:29:20.730 --> 00:29:26.048
There's no way we're going to put something out in the wild or train a model unless a human is like, you know what?
00:29:26.048 --> 00:29:27.130
Okay, approved.
00:29:27.130 --> 00:29:35.930
And that's the type of thing I think we all have to start training ourselves on in a really, really quick manner, which comes back to that's another fundamental shift in business.
00:29:35.930 --> 00:29:38.490
My problem that your company needs to start doing now.
00:29:38.970 --> 00:29:41.568
Is that something interesting about training your AI?
00:29:41.568 --> 00:29:45.250
And it's funny, I started looking at mine as a pet almost, right?
00:29:45.250 --> 00:29:48.650
Like unconditionally loving pet in a way, always available.
00:29:48.650 --> 00:29:51.048
Two in the morning, you put a prompt in there.
00:29:50.890 --> 00:29:51.288
Take.
00:29:51.608 --> 00:29:52.568
Walks, right?
00:29:52.730 --> 00:29:53.528
It's amazing.
00:29:54.410 --> 00:29:57.048
Everybody has to train their own.
00:29:57.048 --> 00:30:09.450
So in this, let's talk about it from an operations and a business perspective, is there going to be one sort of agentic branded AI that understands your business?
00:30:09.450 --> 00:30:12.130
And then everybody kind of starts from that place.
00:30:12.130 --> 00:30:19.250
Or when we were talking before, you were saying that individuals may know how to use AI fairly well, but maybe the companies don't.
00:30:19.250 --> 00:30:30.170
So do you see that companies are going to be training it on the company at large, like an operation sort of bank of information that employees can use to start from?
00:30:30.170 --> 00:30:36.930
Or are you seeing that the individual employees will train their eyes in the way that they work, or both?
00:30:37.250 --> 00:30:38.650
That's a load of questions.
00:30:38.650 --> 00:30:40.288
I'm going to veer off to the right hand side.
00:30:40.288 --> 00:30:41.048
Okay.
00:30:41.130 --> 00:30:42.970
I hate to be so loaded this morning.
00:30:42.808 --> 00:30:43.650
No. It's fine.
00:30:43.568 --> 00:30:44.450
My second cut.
00:30:44.450 --> 00:30:46.410
So I'm surprised.
00:30:46.410 --> 00:30:47.288
So. No. So?
00:30:47.288 --> 00:30:58.088
So that what you just said is, I think the single biggest reason why we're going to have a problem in the unemployment world, and you're not going to be people are not going to be unemployed because they didn't learn AI.
00:30:58.088 --> 00:31:05.970
You're going to be unemployed because your company did not install the right foundational pieces with AI, and they are not going to be able to compete with the two or three person law firms.
00:31:05.970 --> 00:31:14.298
So the law, the monster law firms of five hundred to one thousand attorneys, you are going to have a major problem with the law firm of five attorneys, because they're going to leverage AI.
00:31:14.298 --> 00:31:15.940
They're going to know exactly how to do it.
00:31:15.940 --> 00:31:22.778
They're going to train all of their five team members to do it, and they're going to be able to whip you, and you're not going to be able to withstand that.
00:31:22.778 --> 00:31:24.460
And that's going to be the biggest problem.
00:31:24.460 --> 00:31:27.858
So the layoffs are not going to be we're laying you off because of AI.
00:31:27.858 --> 00:31:33.538
The layoffs are going to be our law firms closing down because we can't compete with these small little companies coming in.
00:31:33.538 --> 00:31:42.578
That's why I tell people, prepare yourself because the day of being your own entrepreneur, running your own business is upon you unlike anything we've ever seen in this country.
00:31:42.578 --> 00:31:48.778
You'll put twenty people in a room and ten are going to be running their own companies, five will be doing whatever, and five will be working for somebody else.
00:31:48.778 --> 00:31:53.098
So the large companies are going to start to shrink down a little bit.
00:31:53.098 --> 00:31:55.980
There's a big argument out there about what companies should do.
00:31:55.980 --> 00:31:57.740
Like, should you lock in on a model right now?
00:31:57.740 --> 00:31:58.980
Should you lock in on Claude?
00:31:58.980 --> 00:32:01.058
Should you lock in on copilot?
00:32:01.058 --> 00:32:06.338
It's really something copilot runs on open AI or should you lock in on open AI?
00:32:06.338 --> 00:32:07.818
I don't have an answer to that.
00:32:07.818 --> 00:32:21.618
I think if you have the fundamentals down, if meaning if you create your data rooms and make sure that all your information is in a data room and that you're constantly building out that data room, identifying individual right now, that would be, we'll call it your data room warehouse manager.
00:32:21.618 --> 00:32:34.220
That is going to be an actual job, which means their entire job is bringing data in, making sure it's organized to make sure they can plug any LL in Claude, open AI, even Grok or Gemini, and it can run that.
00:32:34.220 --> 00:32:37.980
That, to me is the foundational piece in AI that I think companies have to do right now.
00:32:37.980 --> 00:32:42.660
Because if you have all your data in one spot, you're going to be okay with AI.
00:32:42.660 --> 00:32:45.700
But yeah, there is a big question of what we should lock in on.
00:32:45.700 --> 00:32:51.700
I would recommend a company I wouldn't lock in on anything because if you were to tell me Claude is better than ChatGPT a year ago, I would have laughed.
00:32:51.700
--> 00:32:52.778
I'm like, that's crazy.
00:32:52.778 --> 00:32:56.618
Now I don't I have such a hard time using any model.
00:32:56.618 --> 00:32:58.220
But Claude, it's so strong.
00:32:58.220 --> 00:32:59.700
But that's not going to last.
00:32:59.700 --> 00:33:01.460
Gemini is going to pop up tomorrow.
00:33:01.460 --> 00:33:02.778
Open AI is going to pop up.
00:33:02.778 --> 00:33:07.778
So I think that question is going to come down to what search engine you like the most.
00:33:07.778 --> 00:33:09.858
Do I think one is going to emerge as like Google?
00:33:09.858 --> 00:33:12.180
Yeah, probably, but maybe not.
00:33:12.900 --> 00:33:19.900
Can you educate us a little bit because you said it doesn't matter, sort of which which model you plug into your data.
00:33:19.900 --> 00:33:24.740
Can you explain a little bit then about how LLMs do work differently?
00:33:24.740 --> 00:33:36.900
And even if they're coming from different developers and algorithms, and people are even starting to build out like their own agentic models for their businesses using LLMs.
00:33:36.900 --> 00:33:39.618
But like, take it down to the basic.
00:33:39.618 --> 00:33:46.259
Yeah, because you're saying that you don't know which LLM to use maybe on the data that belongs to the company.
00:33:46.259 --> 00:33:49.660
So yeah, how do they actually work?
00:33:49.940 --> 00:33:50.980
So it's not hard.
00:33:50.980 --> 00:33:58.180
They basically go in and they read through text, images, video, and they build out what's called a tokenized process.
00:33:58.180 --> 00:34:00.900
And each word is broken down into tokens.
00:34:00.900 --> 00:34:02.900
It doesn't look at things as necessarily as words.
00:34:02.900 --> 00:34:05.868
It's just tokens on how they basically are trying to.
00:34:05.868 --> 00:34:09.150
What token goes next to the next token goes to the next token.
00:34:09.150 --> 00:34:10.750
That is the level of my knowledge.
00:34:10.750 --> 00:34:15.190
If I were to have anything even beyond that, I would expect a lot more money to be getting paid.
00:34:15.190 --> 00:34:16.550
But I'm not so.
00:34:16.550 --> 00:34:29.429
But it's basically so they're able to take your content and based upon a query that's coming in, it can go through all of your content and formulate an answer by looking at all the words.
00:34:29.429 --> 00:34:32.550
It's not searching like for a specific answer.
00:34:32.550 --> 00:34:35.670
It's actually has ingested all of your information.
00:34:35.670 --> 00:34:39.909
And just imagine if the most the smartest lawyer ever to walk, that's what it thinks it is.
00:34:39.909 --> 00:34:44.110
And it's not because a lawyer is like trying to recite a case that they did like twenty years ago.
00:34:44.110 --> 00:34:51.389
That lawyer is going to give you information based upon fifteen thousand cases that they were in, and they're just going to come out with information saying, this is what you need to look at.
00:34:51.389 --> 00:34:53.070
That's how LLMs operate.
00:34:53.070 --> 00:34:56.389
They're not looking to search, retrieve and pull things back and say, hey, read this.
00:34:56.389 --> 00:34:58.869
They're saying, no, this is your answer.
00:34:58.869 --> 00:35:02.030
Because of all the information we've ingested from this company.
00:35:02.030 --> 00:35:07.750
This is what we believe their answer would be to you because of all the information, not a specific cases.
00:35:07.750 --> 00:35:10.190
Hey, we're going to cite this specifically for what it is.
00:35:10.190 --> 00:35:12.670
I hope that explains what that means.
00:35:12.670 --> 00:35:14.550
So I don't know if that does make sense.
00:35:14.550 --> 00:35:24.510
But so the goal is each LM for right now, the big ones you want to worry about are Claude OpenAI, which is ChatGPT and Gemini, which is Google.
00:35:24.510 --> 00:35:25.869
Claude is anthropic.
00:35:25.869 --> 00:35:27.670
So you could literally take any of those models.
00:35:27.670 --> 00:35:29.030
And if you had a data room.
00:35:29.030 --> 00:35:44.349
So the data room should be, I would say your F a, all your FAQs, all of your SOPs, uh, any type of marketing material, any type of collateral, any type of blueprints, any single possible thing that your company has created online or offline should be in this data room.
00:35:44.349 --> 00:35:45.510
It should be organized.
00:35:45.510 --> 00:35:50.550
You could drop any of those three models on top and start having a conversation with your data.
00:35:50.550 --> 00:35:56.590
It literally puts a voice on your data and will answer you with all the information it has.
00:35:56.590 --> 00:36:02.909
And if you want it to, you could let it tap into the wild, which means you could actually have it, use your data and combine it with other data.
00:36:02.909 --> 00:36:06.349
I don't know if I recommend that and then it would give you answers.
00:36:06.349 --> 00:36:11.989
So I think that's something that companies are not really leveraging.
00:36:11.989 --> 00:36:19.110
They're not law firms, especially like you should have every case you've ever been involved in, all in a secure data room.
00:36:19.110 --> 00:36:21.869
And IT companies can help you with data rooms are not that hard.
00:36:21.869 --> 00:36:27.190
And you should be literally talking to that like all everything should be right there.
00:36:27.190 --> 00:36:32.429
Because a lot of times you're finding more companies and law firms, they just ask ChatGPT a question.
00:36:32.429 --> 00:36:33.349
They ask Claude a question.
00:36:33.349 --> 00:36:42.909
My first question is, why aren't you putting all your data first in there and asking that the question first, take your answer, then bring it over to ChatGPT, see what they say.
00:36:42.909 --> 00:36:43.989
Bring that back over.
00:36:43.989 --> 00:37:01.599
So long story short, it should start with all of your company's history, all the information you've, you know, built law firm, whatever all over
how many years and put a voice on that data first with an LLM before you go out and just start asking random questions, You want to learn where the answer may or may not be right for you.
00:37:01.800 --> 00:37:02.559
Right?
00:37:02.559 --> 00:37:12.039
So then, this is the sort of final question in this series is what about AI generated things informing AI?
00:37:12.039 --> 00:37:16.000
So we see that AI will generate questions.
00:37:16.000 --> 00:37:21.199
It'll generate images, it can generate solutions, it can generate all the things.
00:37:21.199 --> 00:37:26.599
But a lot of people are very quick to say, well, I know that's AI generated or I can see that's AI generated.
00:37:26.599 --> 00:37:29.719
So at what point is AI generated?
00:37:29.719 --> 00:37:36.000
Types of things going to inform future AI responses, if you will, inside your company.
00:37:36.000 --> 00:37:37.639
And is that a good thing?
00:37:39.320 --> 00:37:40.840
Trying to think what is your what is the question?
00:37:40.840 --> 00:37:43.320
You mean like when will we be able to.
00:37:44.400 --> 00:37:51.518
So for example, let's say you loaded all of your marketing campaigns for the last two years, right?
00:37:51.518 --> 00:37:55.960
And all your emails that you send, just the whole marketing campaign, whatever it looks like for your company.
00:37:55.960 --> 00:38:02.518
And we put it up in there and it says, oh, this content, let's say that you wanted to confirm your pillars or whatever.
00:38:02.518 --> 00:38:15.559
And I know top of funnel is going away, but and it comes back and it says, well, let me generate for you and it generates for you based on what you fed it originally, more AI content, lots and lots of content.
00:38:15.559 --> 00:38:19.960
Does that content that it generates go into the bank still.
00:38:19.960 --> 00:38:24.518
So it'll start to pull from itself for future questions and things.
00:38:24.518 --> 00:38:29.039
If do we just accept it and say, yes, this is great, and then it'll put itself in the library?
00:38:29.039 --> 00:38:33.360
Or does it start always from human generated content?
00:38:33.199 --> 00:38:37.320
No, that's a great question because that that's actually probably a great way to end it.
00:38:37.320 --> 00:38:38.599
That's where we're headed.
00:38:38.599 --> 00:38:39.840
Synthetic content.
00:38:39.840 --> 00:38:45.239
So AI is starting to create content for its own self, which to train itself.
00:38:45.239 --> 00:38:55.239
So we actually have built a couple products where we actually trained a couple models we like, we kind of did like a couple dummy small companies where they're not even real things.
00:38:55.239 --> 00:38:59.159
It's like we call it the lab, and these things don't even exist.
00:38:59.159 --> 00:39:06.920
But I created I went in and I created, I grabbed a bunch of documents, I put it in there and I was like, okay, I want you to learn about this business in this industry.
00:39:06.920 --> 00:39:08.960
Did I see s t a m s like everything?
00:39:08.960 --> 00:39:10.239
We throw it in there, it was all made up.
00:39:10.239 --> 00:39:11.039
It's fake.
00:39:11.039 --> 00:39:12.920
But what happened then?
00:39:12.920 --> 00:39:17.920
I said, okay, now I need you to start creating more information for me based upon what you know, to build out the data room.
00:39:17.920 --> 00:39:19.280
And it started doing it.
00:39:19.280 --> 00:39:24.000
It started reading all the information and it started creating, creating, creating so much to the point where I started.
00:39:24.000 --> 00:39:25.159
We call them golden reports.
00:39:25.159 --> 00:39:27.400
I started like twenty five Golden reports.
00:39:27.400 --> 00:39:30.960
There's thousands of pages in there right now, one hundred percent AI generated.
00:39:30.960 --> 00:39:32.000
We just click a button.
00:39:32.000 --> 00:39:33.760
It generates more content, click a button.
00:39:33.760 --> 00:39:42.039
So what's going to happen is these LMS, they're going to look at all your information and they're going to say, the number one thing we have to do is build up the data room.
00:39:42.039 --> 00:39:42.320
Okay.
00:39:42.320 --> 00:39:43.800
That's so that's how smart they're going to get.
00:39:43.800 --> 00:39:46.199
These things are pretty much almost here right now.
00:39:46.199 --> 00:39:49.119
They're gonna look at your data room, which may be sparse or maybe it's really vast.
00:39:49.119 --> 00:39:55.489
They're going to say they're going to build little LMS underneath it, and those little LMS are going to start creating content individually.
00:39:55.489 --> 00:39:57.050
This is everything we know about bankruptcy.
00:39:57.050 --> 00:39:59.730
This is everything we know about divorce and so on and so on.
00:39:59.730 --> 00:40:00.610
So you're like, okay, that's great.
00:40:00.610 --> 00:40:01.329
Creates all that content.
00:40:01.329 --> 00:40:02.409
We're gonna take it a step further.
00:40:02.409 --> 00:40:07.210
Then it's going to create the content on content that it's already created, and it's going to keep going down and down.
00:40:07.210 --> 00:40:16.610
So yeah, the amount of content and the amount of things it's going to create, that's where you start hitting superintelligence because now, because remember how it comes all the way back full circle.
00:40:16.610 --> 00:40:18.369
I said, well, you got to have rationale.
00:40:18.369 --> 00:40:18.969
You have to have this.
00:40:18.969 --> 00:40:20.090
You have to have that.
00:40:20.090 --> 00:40:28.409
Now, from a predictive standpoint, it is literally hitting almost every single plausible outcome possible, and it's going all the way down.
00:40:28.409 --> 00:40:30.690
And the question is, can it get to one hundred percent?
00:40:30.690 --> 00:40:34.090
Because if it does now humans are humans and life is life like we talked about.
00:40:34.090 --> 00:40:34.329
Yeah.
00:40:34.329 --> 00:40:36.809
So there's always a one percent, but that's where we're headed.
00:40:36.809 --> 00:40:44.010
So once you get the superintelligence, which is where they're all trying to get to, I'm a little concerned about that, to be honest with you.
00:40:44.010 --> 00:40:44.730
That's the one thing.
00:40:44.730 --> 00:40:52.570
If something were to give me pause at my fifty three years of age and maybe twenty more years of work or whatever, if I look in and say, what do you think's the thing that scares you the most.
00:40:52.570 --> 00:40:59.610
If it's super intelligence and then it goes beyond super intelligence, where does that leave the human?
00:40:59.610 --> 00:41:00.929
That's my concern.
00:41:01.170 --> 00:41:02.289
Yeah, I hear you.
00:41:02.289 --> 00:41:04.329
And I know that's in the future.
00:41:04.329 --> 00:41:18.130
And they're trying to solve it today, though as a business owner, how should we be treating, let's say, AI generated content off of other content that we've done is it's still our role to go in and edit, be the pilot.
00:41:18.130 --> 00:41:18.409
Yeah.
00:41:18.409 --> 00:41:30.768
Before we put anything out there and that responsibility, I think I want to talk a little bit maybe about that responsibility for the fact that people can tell things are AI generated now or generally.
00:41:30.768 --> 00:41:31.090
Right.
00:41:31.090 --> 00:41:37.690
I've seen so many things about Em dashes that it's and I love the em dash oh my gosh, I use the em dash like it's like it's a comma.
00:41:37.690 --> 00:41:41.369
Like I just, I absolutely love having that hanging thought and then filling it in.
00:41:41.369 --> 00:41:42.530
That's just my nerdiness.
00:41:42.530 --> 00:41:44.889
But what is our role?
00:41:44.889 --> 00:41:52.889
And I guess because you were talking already about layoffs and we've talked about scaling down, but being faster, not being able to optimize everything.
00:41:52.889 --> 00:42:03.690
But if we get this thing trained and it's making content on our content, where is our responsibility still to lie as the business owner, as we're embracing this?
00:42:03.690 --> 00:42:09.969
Again, wrapping it back up, number one, we have to have the perspective that these people are coming after doing their research.
00:42:09.969 --> 00:42:15.768
And that's an amazing thing because the education side sort of sort of narrows out a little bit.
00:42:15.768 --> 00:42:17.570
But what is our responsibility?
00:42:17.570 --> 00:42:23.170
What should we be doing, especially now, let's say, over the next eighteen months to make sure that we're leveraging it?
00:42:23.170 --> 00:42:26.809
You're saying don't lock anything in, build that data room.
00:42:26.809 --> 00:42:43.809
What should we also be doing to make sure that as these things are evolving, we are not left hidden in the World Wide Web, and nobody can ever find us because we didn't put in the things or so much of our content that is now going out was generated first with AI, with slight tweaks.
00:42:43.809
--> 00:42:46.409
Did they treat that content differently?
00:42:47.250 --> 00:42:51.090
Well, I think before you put anything out in the wild, like whatever content.
00:42:51.090 --> 00:42:58.369
And I know there's a lot of AI videos right now, AI images, you want to go in and literally make sure that you approve it, approve and deny it.
00:42:58.369 --> 00:43:07.570
We are seeing companies that are just kind of throwing stuff out there, and that that could potentially open you up to a lawsuit if you're putting out information that's wrong.
00:43:07.570 --> 00:43:15.969
But I think you talk about the golden age attorneys with wrong information coming out on some of this stuff because of what humans created, just because AI generated it.
00:43:15.969 --> 00:43:20.409
If you don't approve it or you tweak it a little bit, you could open yourself up to legal action.
00:43:20.409 --> 00:43:23.809
If you are somebody who harms himself or they make a wrong decision.
00:43:23.809 --> 00:43:29.768
So I'd be interested to see where the laws go for this because we know they're coming and there should be there should be legal guardrails.
00:43:29.768 --> 00:43:31.489
Let's be crystal clear on that.
00:43:31.489 --> 00:43:37.849
But I think our biggest responsibility is before you put anything out in the wild, you literally need to make sure that you're putting your eyes on it.
00:43:37.849 --> 00:43:45.289
If you're not creating that content anymore and the dirty secret of nobody doesn't know the average business right now is not creating their own content anymore.
00:43:45.289 --> 00:43:47.380
They are having ChatGPT do it now.
00:43:47.380 --> 00:43:50.780
my recommendation is have a conversation with ChatGPT.
00:43:50.780 --> 00:43:55.820
Start typing stuff down, brainstorm it and say, hey, here's all my past blog posts.
00:43:55.820 --> 00:43:57.460
Here's all my past podcasts.
00:43:57.460 --> 00:43:59.018
I want you to really get my voice.
00:43:59.018 --> 00:44:02.380
And every week I'm going to come in and give you my thoughts on everything.
00:44:02.380 --> 00:44:05.980
Can you then structure that as a blog post, as an email campaign?
00:44:05.980 --> 00:44:06.539
That's perfect.
00:44:06.539 --> 00:44:07.500
I think that's perfectly fine.
00:44:07.500 --> 00:44:11.179
That's just not basically having an assistant taking all your thoughts and writing.
00:44:11.179 --> 00:44:11.500
Right?
00:44:11.500 --> 00:44:11.860
Same thing.
00:44:11.860 --> 00:44:12.340
Ghostwriter.
00:44:12.340 --> 00:44:13.219
So that's okay.
00:44:13.219 --> 00:44:14.980
But I think that's the number one thing you have to do.
00:44:14.980 --> 00:44:22.579
But businesses to me right now, if you're if the one question is, what can I do right now to make sure that they're not left behind and so forth?
00:44:22.579 --> 00:44:28.059
It comes back to the famous marketing thing, man, you really, really got to start embracing social media.
00:44:28.059 --> 00:44:30.659
You've really
got to start embracing putting yourself out there.
00:44:30.659 --> 00:44:32.380
I can't stand to hear myself talk.
00:44:32.380 --> 00:44:35.219
I can't stand to watch myself at all in any capacity.
00:44:35.219 --> 00:44:36.699
But we eat our own dog food here.
00:44:36.699 --> 00:44:40.340
Like if I can't tell my client to do it and yet I'm not doing it.
00:44:40.340 --> 00:44:45.579
So to to make sure that you never disappear in this new world that we're going in.
00:44:45.579 --> 00:44:55.739
I'm not saying you have to be constantly out there and like Mr. Beast and all that and kind of doing all these videos, but I think you have to make your voice and your brand's voice relevant on social media.
00:44:55.739 --> 00:44:58.900
And before you put anything out that's been AI generated, absolutely.
00:44:58.900 --> 00:45:00.420
Make sure that you approve it.
00:45:00.420 --> 00:45:03.260
Even like a little create a little company tag underneath.
00:45:03.260 --> 00:45:06.980
Like we always said, everybody can steal that AI generated human approved.
00:45:06.980 --> 00:45:09.699
And this was the human of thumbprint and put the lawyer's name on there.
00:45:09.699 --> 00:45:11.018
I think that's the way to do it.
00:45:11.099 --> 00:45:12.940
I love it, I love it.
00:45:12.940 --> 00:45:22.420
Where can everybody find you and learn more about your company and how you work with businesses and where to to learn, I guess, social media for you?
00:45:22.260 --> 00:45:22.980
Yeah, yeah.
00:45:22.980 --> 00:45:24.659
So you can just go to connexus dot com.
00:45:24.659 --> 00:45:27.420
Anybody that fills the form out, you come to me directly.
00:45:27.420 --> 00:45:30.579
We, we actually, it's don't mean to come across in a bad way.
00:45:30.579 --> 00:45:33.260
We're very selective on how we take just because our team is small.
00:45:33.260 --> 00:45:34.820
We like our bandwidth the way it is.
00:45:34.820 --> 00:45:39.900
So normally when one client graduates on what we're training them on AI, we build infrastructures and systems.
00:45:39.900 --> 00:45:41.980
Once they're in flight, we then bring in the next person.
00:45:41.980 --> 00:45:43.579
So think of it kind of like a ride.
00:45:43.579 --> 00:45:45.340
That's how we operate right now.
00:45:45.340 --> 00:45:46.659
If you want to just follow me.
00:45:46.659 --> 00:45:50.139
A lot of times I put up clips on what I'm talking about and I put some notes up as well.
00:45:50.139 --> 00:45:51.699
In AI, you can just find me on LinkedIn.
00:45:51.699 --> 00:45:52.739
It's probably the easiest way.
00:45:52.739 --> 00:46:00.619
And the last thing I say, if you have a general AI question, just something that's stumping you, or maybe you just have a question, you're not really looking to do business or something.
00:46:00.619 --> 00:46:02.380
I love getting inboxes on that stuff.
00:46:02.380 --> 00:46:07.780
I may not be able to solve your problem per se, but maybe I can point you in a direction and say, hey, why don't you go look at that?
00:46:07.980 --> 00:46:08.860
Amazing.
00:46:08.860 --> 00:46:11.059
So we're going to put all that down inside the show notes.
00:46:11.059 --> 00:46:14.739
So if you find this on the day that we launch, you can find Len.
00:46:14.739 --> 00:46:19.500
If you wait two years, I don't know how relevant this is going to be in two, two and a half years.
00:46:19.500 --> 00:46:20.460
We may be out.
00:46:20.460 --> 00:46:25.460
This may be one of those episodes that actually becomes less evergreen, which is interesting to me.
00:46:25.460 --> 00:46:28.139
I actually just did to see how this pans out.
00:46:28.139 --> 00:46:30.139
I have notes on that, but you'll be.
00:46:29.980 --> 00:46:32.619
Able to have a couple podcasts that I did like a year ago.
00:46:32.619 --> 00:46:33.699
I like want him to take them down.
00:46:33.699 --> 00:46:35.179
I'm like, my God, these things are irrelevant.
00:46:35.179 --> 00:46:37.579
I feel like it's a dinosaur running around like it's bad.
00:46:37.579 --> 00:46:38.699
So yes.
00:46:38.539 --> 00:46:49.110
It's interesting that some some evergreen things are not so evergreen, but I your perspective on how we should be seeing the clients coming in with their research, even if you're not a lawyer.
00:46:49.110 --> 00:46:56.869
I think that's an incredible perspective for us to start having that it's okay that everybody else is also educated in what we do before we have a conversation.
00:46:56.869 --> 00:46:59.230
I think that's actually a beautiful thing.
00:46:59.230 --> 00:46:59.510
Yeah.
00:46:59.510 --> 00:47:02.630
So thank you so much for being with me today.
00:47:02.630 --> 00:47:02.909
All right.
00:47:02.909 --> 00:47:06.869
We'll talk to you next time on the next episode of the Business Mastery Podcast.
00:47:06.869 --> 00:47:07.710
Take care.
00:47:07.710 --> 00:47:17.750
Thank you so much for listening to this episode of the Business Mastery Podcast.
00:47:17.750 --> 00:47:32.710
If you want to learn more about me, you can go to Don k Kennedy dot com and you can now check us out on YouTube as well as, of course, any of your favorite platforms that host podcasts.
00:47:32.710 --> 00:47:33.590
Take care.