WEBVTT
00:00:03.725 --> 00:00:04.426
Welcome.
00:00:04.426 --> 00:00:18.532
This is the Public Relations Review Podcast, a program to discuss the many facets of public relations with seasoned professionals, educators, authors and others.
00:00:18.532 --> 00:00:22.167
Now here is your host, peter Woolfolk.
00:00:24.722 --> 00:00:30.545
Welcome to the Public Relations Review Podcast and to our listeners all across America and around the world.
00:00:30.545 --> 00:00:40.648
Apple has ranked this podcast among the top 1% of podcasts worldwide, so thank you to all of our guests and listeners for being the basis for this ranking.
00:00:40.648 --> 00:00:47.246
Now, if you enjoy the podcast, we certainly would like to have a review from you, and we'd like to ask you to share this with your friends.
00:00:47.246 --> 00:00:53.046
Now question is Google still the leading search platform for public relations practitioners?
00:00:53.667 --> 00:00:55.673
Can anyone overtake Google.
00:00:55.673 --> 00:01:07.302
Well, the new platform on the block is called LLMs Large Language Models and my guest today will expand on LLMs and why they are more impactful.
00:01:07.302 --> 00:01:11.277
So joining me today from Atlanta, Georgia, is Frank Strong.
00:01:11.277 --> 00:01:21.484
Frank is the founder and president of the sword and the script, a veteran-owned business focused on public relations, content marketing and social marketing for the b2b marketplace.
00:01:21.484 --> 00:01:34.846
He has implemented programs and led teams for public relations, content marketing and social media on the in-house side of the table with corporations ranging from startups to the mid-market to global organizations.
00:01:34.846 --> 00:01:49.650
He has also endured the rigors of billable hours, having grown up on the agency side, with gigs at PR firms small, large, including the top global firm, hill Knowlton, now HK Strategies.
00:01:49.650 --> 00:01:52.588
So, Frank, welcome again to the podcast.
00:01:52.588 --> 00:01:59.329
Now give us an overview of LLMs and the major impact it is having on public relations and searching.
00:02:00.331 --> 00:02:01.834
Hi Peter, thanks for having me back.
00:02:01.834 --> 00:02:02.715
It's good to see you again.
00:02:02.715 --> 00:02:06.504
Yeah, llm stands for Large Language Model.
00:02:06.504 --> 00:02:10.774
It is the model that drives kind of generative AI interface.
00:02:10.774 --> 00:02:21.329
So if you go to a platform like ChatGPT, perplexity, google's Gemini Claude, any of those platforms, they're all being driven by an LLM.
00:02:21.329 --> 00:02:30.317
And, without getting into all the ins and outs, an easy way to think about an LLM is it's kind of like a very sophisticated word suggest.
00:02:30.317 --> 00:02:38.945
So if you go to a traditional Google search engine and you start typing in a search term, it's trying to guess what you're going to ask.
00:02:38.945 --> 00:02:40.673
You know you say red cars.
00:02:40.673 --> 00:02:43.804
You could say red cars in Kansas, red cars in Brazil.
00:02:43.804 --> 00:02:45.049
You know fast red cars.
00:02:45.049 --> 00:02:49.431
So it's going to give you all of these suggestions and so LLMs kind of work the same way.
00:02:51.240 --> 00:02:58.274
Well, now that we've heard that sort of basic explanation of it, what is it actually doing?
00:02:58.274 --> 00:03:00.610
I mean, as people are using it, what are they getting?
00:03:00.610 --> 00:03:04.966
What are they seeing that is causing this rush over to LLMs?
00:03:05.848 --> 00:03:06.980
Yeah, I think it would be good.
00:03:06.980 --> 00:03:18.564
I'll just take a quick step back and explain kind of the old newspaper model and how that was displaced by Google and how that's being displaced by LLMs today, and the reason for that is, you know, it's conceivable.
00:03:18.564 --> 00:03:28.233
There are people today that didn't grow up seeing stacks of newspapers by a vendor and so a traditional newspaper you wanted to be what was called above the fold.
00:03:28.233 --> 00:03:33.211
If you hold a newspaper straight up and you fold it in half, they would stack the newspapers together.
00:03:33.211 --> 00:03:36.288
It would look like a bale of hay and that's how they shipped them around.
00:03:36.288 --> 00:03:45.700
That's how it was displayed on the newsstands, and so the above the fold front page of a newspaper was the most visible place that you could possibly be For years and years.
00:03:45.700 --> 00:03:49.368
That's how people got most of their information until Google came along.
00:03:49.931 --> 00:03:51.112
Then it got a whole lot easier.
00:03:51.112 --> 00:04:06.836
Instead of sorting through a newspaper when you're looking for information or newspapers it could be many or going to the library and sorting through microfilm of old newspapers, you could just go to Google, ask a question and it would give a list of sources, right, and you could choose which source you want to look at.
00:04:06.836 --> 00:04:09.042
You could review which ones answered your question best.
00:04:09.042 --> 00:04:15.241
Those queries that answered your question best tended to rank higher over time in the search engine.
00:04:15.241 --> 00:04:25.333
That's all being changed by llms and the reason for that is llms are able to present information as just an answer.
00:04:25.333 --> 00:04:30.574
It it's almost like you're talking to a person and they're typing up a response to your question.
00:04:30.634 --> 00:04:33.182
You ask it a question and it gives you like a primer.
00:04:33.182 --> 00:04:36.870
That comes with some benefits and drawbacks too.
00:04:36.870 --> 00:04:38.802
It should be noted, like when you think.
00:04:38.802 --> 00:04:45.509
For example, when you went and did a traditional Google search, it gave you however many results it could be 20, 30, 100.
00:04:45.509 --> 00:04:49.870
And you got to choose what information you thought best answered your question.
00:04:49.870 --> 00:04:55.512
Llms present it as a briefing and it presents it from a position of authority.
00:04:55.531 --> 00:05:08.367
Like here is the answer to your question and that can have some drawbacks, mainly because A it's providing an answer and that may or may not be the correct one or the one that works best for you.
00:05:08.367 --> 00:05:14.305
And two, we know for a fact that generative AI programs do hallucinate.
00:05:14.305 --> 00:05:18.786
That means they just simply make up the answer and they're doing that because they're trying to.
00:05:18.786 --> 00:05:22.701
It's basically like an average of words, so you have to be careful what you believe.
00:05:23.663 --> 00:05:24.766
You know I have.
00:05:24.766 --> 00:05:38.668
Once I read your article, I looked into some of the other LLMs and I jumped onto a perplexity and I asked a very basically a question such as you know, how many listeners does the Public Relations Review podcast have?
00:05:38.668 --> 00:05:49.853
And while it was very, very quick with that coming back, I was shocked at how much information came back, and it came back in a format as though we were having had had a conversation about this.
00:05:49.853 --> 00:05:52.725
So I was, you know, pleased about it.
00:05:52.725 --> 00:05:55.773
It did not give me a specific number, as you said before it.
00:05:55.773 --> 00:06:00.261
Let you know, we went here and here's what we're here, here's what we got there and these other sort of places.
00:06:00.261 --> 00:06:10.411
So I was not only was I pleased, I was was just surprised at the format that it came back in, and I think this might be helpful to a lot of people who might consider using these.
00:06:11.160 --> 00:06:12.023
Yeah, I completely agree.
00:06:12.023 --> 00:06:19.973
It's incredibly amazing and I think your example brings up a good question, because you've had this podcast for a long time.
00:06:20.639 --> 00:06:29.595
So it's been indexed and these large language models have been able to be trained on all kinds of content and obviously, including your podcast.
00:06:29.595 --> 00:06:38.733
That's how it's able to return the answer, and so the question that it conjures is if LLMs are going to be the new front page, and there's some evidence suggesting that this is happening.
00:06:38.733 --> 00:06:43.512
Google's search traffic has dropped for the first time since 2015.
00:06:43.512 --> 00:06:56.911
And it was a small amount, but there is data that suggests referral traffic from LLMs Perplexity, chatgpt, claude Gemini, google Gemini that's grown, so we know that this is in transition.
00:06:56.911 --> 00:07:02.547
So the question that PR professionals should be asking themselves is how do we get into an LLM?
00:07:02.547 --> 00:07:06.423
You've gotten your podcast into an LLM and that is one approach, right.
00:07:06.423 --> 00:07:15.850
You have to have content, and I think a lot of the answers to the question of how do I pitch an LLM comes back to the fundamentals that PR people already know.
00:07:15.850 --> 00:07:23.949
It's having a strong point of view, it's being able to articulate a story so that it gets picked up, it gets a visibility and indexed.
00:07:23.949 --> 00:07:27.343
It's answering specific questions right.
00:07:27.964 --> 00:07:37.815
When people go to a traditional search engine and they type in a search query, that is, by definition, an expression of need I'm looking for information.
00:07:37.815 --> 00:07:47.855
So answering specific questions is a great way to get crawled and indexed by an LLM to provide that response great way to get crawled and indexed by an LLM to provide that response.
00:07:47.855 --> 00:07:53.822
And then, last but not least, the traditional aspects matter.
00:07:53.822 --> 00:07:57.875
So traditional media is going to be deemed highly credible by LLMs and so it's going to use that to train the data and return answers.
00:07:58.398 --> 00:08:00.064
All of the traditional things that PR people did.
00:08:00.064 --> 00:08:01.874
Press releases, for example.
00:08:01.874 --> 00:08:28.324
There seems to be some evidence suggesting that press releases are being used, and that makes a lot of sense to me, because when you think about press releases relative to the writing on the rest of the web and we can make fun of you know press releases are promotional and people do funky things with them, but the writing is probably a higher quality relative to the rest of the web, and so it's a really good data set on which to train these models, and I think that's why it's happening.
00:08:28.846 --> 00:08:33.605
If there's a caveat, it would be that at some point this happened with Google.
00:08:33.605 --> 00:08:35.640
Press releases were a good way to get into search engines.
00:08:35.640 --> 00:08:41.863
Google started deprecating the link value because they don't want you to be able to buy a press release to get into their search engine.
00:08:41.863 --> 00:08:43.465
That's kind of gets spammy.
00:08:43.465 --> 00:08:47.009
I suspect something like that will happen with LLMs at some point.
00:08:47.870 --> 00:08:50.375
Where else do you see LLMs fitting in?
00:08:50.375 --> 00:09:07.631
Because you know some of the things we've talked about here and I think I may have mentioned earlier that some organizations where consumers could call into the company and ask questions about the company or company products or those kinds of things Would this be an appropriate use for LLMs?
00:09:08.293 --> 00:09:10.506
Absolutely, and I think there's two ways to look at that.
00:09:10.506 --> 00:09:13.024
One would be to ask a question.
00:09:13.024 --> 00:09:21.972
Just like you asked the question of perplexity about your podcast, you can go to any public generative AI tool and ask a question about a company and it's going to return results.
00:09:21.972 --> 00:09:29.594
And that's the kind of media relations aspects that PR people need to be worried about is how are we being presented in these LLMs?
00:09:29.594 --> 00:09:30.582
That's one use case.
00:09:30.582 --> 00:09:41.571
The other use case along those lines is companies can purchase technologies, these generative AI tools, and train them on their internal documentation.
00:09:41.571 --> 00:09:54.028
If you have a product, for example, a technical manual, that can consume the whole technical manual, probably do it better than a human, and then it's trained on the answers that customers may have about your product.
00:09:54.028 --> 00:10:00.741
So come to a chat bot on your website and they ask a specific question and that AI has been trained on your product documentation.
00:10:00.741 --> 00:10:09.594
It can return a very good answer, which is a far it sounds like a far better approach than saying here's a bunch of help desk articles.
00:10:10.115 --> 00:10:18.591
Go ahead and search through these and see if you can find your answer and then, if you can't, submit a support ticket and we'll get back to you whenever we can and we'll schedule a call and go through your issue, right.
00:10:18.591 --> 00:10:29.404
So it's just improving the experience, I think, from a customer service perspective, and that's great use case, that's that's being entered now well you know, I certainly see that as a viable function.
00:10:29.445 --> 00:10:32.258
I'm sure there might be some companies using that already.
00:10:32.258 --> 00:10:34.630
As a matter of fact, I haven't see where they've got.
00:10:34.630 --> 00:10:47.403
Some companies have a on videos that can respond to consumer questions, and so I'm sure that I Other organizations are going to use this as a PR tool because, yes, we can answer your question right away.
00:10:47.403 --> 00:10:49.739
You don't have to wait for a real person to show up.
00:10:49.739 --> 00:11:04.124
You know, as soon as you call us, we can get back to you, or we can answer you right away even with someone talking to you, and in some cases, I wouldn't be surprised if they can take a real human face and animate it so that it looks like it's actually talking.
00:11:04.124 --> 00:11:06.639
In some cases, I think that is being done.
00:11:07.330 --> 00:11:08.475
Oh yeah, that's definitely happening.
00:11:08.475 --> 00:11:10.618
Some of the visuals are really compelling.
00:11:10.618 --> 00:11:15.376
I mean, they're able to make you know fully-fledged movies based on some generative AI prompts.
00:11:15.376 --> 00:11:16.320
It's actually crazy.
00:11:16.320 --> 00:11:27.852
One of the ways that you can notice, though, is that, for whatever reason, the same way that generative AI tends to hallucinate with text-based answers, it makes some wonky errors with visuals.
00:11:27.852 --> 00:11:33.576
Like, if you ask it to create a picture and you look at the hands it tends to have like seven fingers on it.
00:11:33.576 --> 00:11:35.043
It'll have like weird things like that.
00:11:35.043 --> 00:11:39.677
So these are like telltale signs and videos and images that you're able to spot some of this.
00:11:40.559 --> 00:11:47.503
Well, what are some of the things, some of the other things that you see in the future for public relations with these LLMs?
00:11:48.169 --> 00:12:01.100
Well, I think we're in the middle of a sea change right now, like now is the moment right If you were around when the Google search engine started becoming the dominant source, the first source of place that people went to for information.
00:12:01.100 --> 00:12:03.859
We're seeing that transformation happening now.
00:12:03.859 --> 00:12:07.421
Google, the traditional search engine, is very worried about this by the way.
00:12:07.830 --> 00:12:14.282
That's why they're investing in Google Gemini, and you may see what they call AI artificial intelligence overviews in search.
00:12:14.282 --> 00:12:17.416
If you go to Google right now and you type in a search, it's going to give you an AI overview.
00:12:17.416 --> 00:12:20.611
So they're working on this and they're playing catch-up chat.
00:12:20.611 --> 00:12:28.851
Gpt cause them by surprise, but I think this is an area that is emerging rapidly.
00:12:28.851 --> 00:12:36.421
We can see that it is taking over as a primary source because it provides a very neat, clean, simple answer when people ask questions, and so we got to be up on it.
00:12:36.421 --> 00:12:39.150
It means we're going to have to monitor our reputation there.
00:12:39.211 --> 00:12:43.472
How are these LLMs presenting the organizations that we represent, whether it's an employer or a client?
00:12:43.472 --> 00:12:46.922
How do we correct things if they go awry?
00:12:46.922 --> 00:12:49.798
Right, what happens when there's crisis communication?
00:12:49.798 --> 00:12:53.895
If it's getting something wrong and it sends off a tizzy, how are we going to go about fixing that?
00:12:53.895 --> 00:12:59.479
And then, generally, how do we get into the LLMs, like we've described already in the podcast?
00:12:59.479 --> 00:13:06.124
So this is emerging and it's going to unfold over the next 12 to 18 months.
00:13:06.124 --> 00:13:07.456
We're going to see an awful lot of change.
00:13:08.490 --> 00:13:15.724
And I also see it as a sales tool as well, because obviously you've got cars, tv sets or whatever else you might be selling.
00:13:15.724 --> 00:13:27.904
That probably is a lot easier for in some cases LLMs and whatever sort of visuals that they have with that can save a lot of time and answer a lot of questions quickly to consumers.
00:13:29.030 --> 00:13:30.056
Yeah, I completely agree.
00:13:30.056 --> 00:13:36.677
I think this is an extension of a trend that we've seen already and that is in the old days, you know.
00:13:36.677 --> 00:13:39.557
Let's say you had to buy a car In the old days, you didn't know much about cars.
00:13:39.557 --> 00:13:41.056
You went to the car dealership.
00:13:41.056 --> 00:13:43.291
The salesperson knew everything about the car.
00:13:43.291 --> 00:13:54.977
Then, when the internet got big, you could do all kinds of research and not from just official review companies, but you could see individual users would write reviews.
00:13:54.977 --> 00:14:02.682
You could get a whole lot more information and walk into a dealership prepared, knowing what you want, what the specs are, how it compares.
00:14:02.682 --> 00:14:10.028
I think we're going to see something like that become even more refined, more repolished and better answers from generative AI going forward.
00:14:10.028 --> 00:14:14.313
That's going to be a lot easier.
00:14:14.313 --> 00:14:18.229
Instead of spending hours and hours doing searches and trying to pull information together, it's basically presenting you with a dossier.
00:14:18.249 --> 00:14:21.417
That's an answer to whatever question that you ask, which is pretty incredible.
00:14:21.417 --> 00:14:24.712
Again the drawbacks are you've got to make sure it's accurate.
00:14:24.712 --> 00:14:28.852
So you mentioned perplexity In my observation.
00:14:28.852 --> 00:14:35.190
They were one of the first LLMs to provide source links, so it will tell you where it got the information.
00:14:35.190 --> 00:14:37.457
You can click the link and go check out the source and see.
00:14:37.498 --> 00:14:38.159
Is this credible?
00:14:38.159 --> 00:14:38.289
Do?
00:14:38.289 --> 00:14:51.998
I believe that, and all of the, as far as I can tell, all of the other LLMs have since followed suit, because you know, we know, working in public relations, having you know credible sources, citing your data, is important, and so we're starting to see the LLMs do that.
00:14:51.998 --> 00:15:02.450
There's also one other emerging property that LLMs are the companies that make them, are working on to combat the hallucination, right.
00:15:02.549 --> 00:15:17.836
So, for whatever reason they don't really understand why, but you can ask a generative AI tool a question and if it doesn't know the answers, once in a while this happens less more often than not it gets it right, but once in a while it doesn't know the answer and it just makes one up.
00:15:17.836 --> 00:15:23.436
It just gives you an answer that's completely fictional and it presents it with such confidence.
00:15:23.436 --> 00:15:26.881
You're like you know it'd be, like somebody like, well, here's the answer, I'm absolutely confident it is right.
00:15:26.881 --> 00:15:28.085
And you don't know.
00:15:28.085 --> 00:15:39.538
They don't really know why that's happening, but one of the things they're working on to develop is this new development called reasoning, and what reasoning is doing is it's showing you.
00:15:39.538 --> 00:15:47.960
It's showing you the steps of how the LLM got to the answer that it is, and I think all of these developments will make LLMs more credible, more conclusive.
00:15:47.960 --> 00:15:51.043
They're going to get better over time and deliver better answers.
00:15:51.083 --> 00:15:53.746
So it's something that the PR community really needs to pay attention to.
00:15:54.490 --> 00:15:59.571
You know, I would think so and I just keep thinking of some other areas that it can be used in.
00:15:59.571 --> 00:16:04.078
I mean public relations, sales, you know, just general information.
00:16:04.078 --> 00:16:28.774
You know, at a conference of some kind, a lot of different ways that or students can call up and or interact with it to get responses to particular questions, have to do classroom materials and so forth, and it comes through and I guess they could add visuals to this thing as we move along yeah, absolutely mean there's a couple of very definitive applications in public relations that are already happening.
00:16:29.216 --> 00:16:36.916
If you're using most of the big PR vendors, if you're using MuckRack, Cision just added it.
00:16:36.916 --> 00:16:39.522
Meltwater has got a huge AI team.
00:16:39.682 --> 00:16:45.591
I mean they've got scientists they've gone out and done acquisitions just to hire scientists with PhDs to help them work on this stuff.
00:16:45.591 --> 00:16:55.302
All of the big ones have that, so they can do things like if you're using media monitoring, it can better categorize, you know placements and put tags and categories of it.
00:16:55.302 --> 00:16:55.912
Is it a story?
00:16:55.912 --> 00:16:56.495
Is it a mention?
00:16:56.495 --> 00:16:57.197
Is it a backlink?
00:16:57.197 --> 00:16:58.592
Pr?
00:16:58.592 --> 00:17:02.152
People are using it for brainstorming, right Like you just need to.
00:17:02.152 --> 00:17:08.324
You know, sometimes looking at a blank page is the biggest hurdle to getting started when you're doing a writing project.
00:17:08.470 --> 00:17:10.397
Well, get into a generative AI system.
00:17:10.397 --> 00:17:11.595
Ask it a bunch of questions.
00:17:11.650 --> 00:17:13.054
It's going to give you a whole bunch of ideas.
00:17:14.038 --> 00:17:18.155
Summarization Exceptional at summarization, like here's a big report.
00:17:18.155 --> 00:17:23.682
I don't have time to read this whole thing, but, you know, read it for me and give me the five key points I need to understand.
00:17:23.682 --> 00:17:27.694
Take me and give me the five key points I need to understand.
00:17:27.694 --> 00:17:33.053
Take my blog post or press release and give me 10 tweets that I can then go ahead and schedule, or Facebook posts or LinkedIn posts like a lot of use cases.
00:17:33.815 --> 00:17:41.101
There's a few vendors that are doing some interesting things where it has prompts that will help you write a pitch or help you write a press release.
00:17:41.101 --> 00:17:42.894
I'm not big into that.
00:17:42.894 --> 00:17:43.498
I really think.
00:17:43.498 --> 00:17:46.671
I think it's good for brainstorming, but I'd rather have a human write.
00:17:46.671 --> 00:17:49.861
I think humans write better than generative AI At least PR people do.
00:17:49.861 --> 00:18:08.670
But one of the interesting things that one of these tools does is it will look at the words that you are using as you're writing your pitch and then it will go out and review articles that are currently published by reporters and come back as you're writing your pitch and saying you should pitch these people because they've written these topics, and here's the link to that.
00:18:08.670 --> 00:18:11.219
That's a pretty cool application of generative AI.
00:18:12.011 --> 00:18:19.958
You know, one of the things that I have done with things such as chat, gpt, for instance, that I might lay out some things that I'd like for you to write for me on this.
00:18:19.958 --> 00:18:28.840
Then, once it comes out, then I will take it and make it my own, in other words, put in the words or other things, and that makes it as if it was coming from me.
00:18:28.840 --> 00:18:41.900
But it has saved me a lot of time and you know and I'm just tweaking it so that I'm comfortable with what it says rather than just ripping and running with it that could cause some problems down the road.
00:18:41.920 --> 00:18:42.801
Yeah, no, a hundred percent.
00:18:42.801 --> 00:18:52.477
I mean, I think what you said is really important, that if you're using these tools, you know, think about it as a first draft and then edit it and make it your own.
00:18:53.557 --> 00:18:57.359
The problem is is I mean, generative AI is basically a.
00:18:57.359 --> 00:19:05.707
These LLMs are giant models of words, and it understands how people tend to put those words in a certain order.
00:19:05.707 --> 00:19:16.971
Based on probability, it is probable that a sentence is going to read a certain way and have a certain structure, and so what that means is the answers that generative AI is giving you is, by definition, an average.
00:19:16.971 --> 00:19:23.836
Here is the average language that most people use, and we know in PR, we don't want to be average, we want to be different.
00:19:23.875 --> 00:19:27.095
We want to explain to people what's different about our products what's better right.
00:19:27.095 --> 00:19:31.071
So it's really important to make sure you have that human touch.
00:19:31.071 --> 00:19:35.040
But I think your application of the tool is a good one.
00:19:35.040 --> 00:19:39.759
I do that an awful lot where, if I don't like the way a sentence is written, I'll ask it.
00:19:39.759 --> 00:19:41.853
You know, rewrite this sentence, give me a better structure.
00:19:41.853 --> 00:19:43.155
Is there an easier way to write it?
00:19:43.155 --> 00:19:45.561
Or another application is?
00:19:45.561 --> 00:19:47.673
I don't really like this headline.
00:19:47.673 --> 00:19:48.935
I'm using the sub headline here.
00:19:48.935 --> 00:19:50.559
Can you give me five alternatives?
00:19:51.000 --> 00:19:51.301
And it'll.
00:19:51.570 --> 00:19:53.556
it'll belt them out and I can choose to use it.
00:19:53.556 --> 00:19:57.454
I can rearrange, I can mix and match, and that's helping me to make my content better.
00:19:58.136 --> 00:20:09.664
Well, you know, actually with this podcast now it, it records everything and then it cranks out not only the transcript for me, but it also suggests titles for each episode.
00:20:09.664 --> 00:20:10.934
Now I'll read those.
00:20:10.934 --> 00:20:15.868
Either I like them, or I don't like them, or I'll make alterations to them, but it gives me all of that.
00:20:16.170 --> 00:20:25.262
It gives me blurbs for Facebook and other social media posts and those kind of things.
00:20:25.262 --> 00:20:25.742
I read them.
00:20:25.742 --> 00:20:26.605
If I like them, fine.
00:20:26.605 --> 00:20:27.506
If not, I change them.
00:20:27.506 --> 00:20:32.301
But these things are a great time saver is what I'm really seeing here.
00:20:32.301 --> 00:20:50.857
There's a huge amount of time being saved and helping you get the job done faster and, in some cases, maybe even more accurately, because it may come up with some information that you had overlooked, because, as you said, these things reach out to a wide range of places to pick up the information and add it to what it's giving to you.
00:20:51.440 --> 00:20:52.082
Yeah, that's right.
00:20:52.082 --> 00:20:59.638
I mean, it's almost cliche to say this at some point, but the topic comes up over and over is people are like are we going to be replaced by generative AI?
00:20:59.638 --> 00:21:05.834
And I don't think we are going to be replaced as individuals.
00:21:05.834 --> 00:21:07.520
Generative ai is very good at specific tasks, but it can't do a job.
00:21:07.520 --> 00:21:12.719
You can't say go be my pr person for me right, I can't do that, but jet pr.
00:21:12.739 --> 00:21:23.564
People need to pay attention to this stuff because, while you won't probably won't be replaced by generative ai, you could be replaced by somebody that has learned how to use it effectively and been able to be more productive.
00:21:23.564 --> 00:21:24.875
That's a real risk.
00:21:25.710 --> 00:21:39.702
You know the other thing, when I look at a lot of times about public relations, public relations a lot of times is a problem-solving exercise, and if the platform doesn't know what the problems are, it can't offer you any solutions to them.
00:21:42.054 --> 00:21:45.417
So you know questions like how do we fix ABC and D?
00:21:45.417 --> 00:21:48.650
That's right, you can use it to brainstorm possible solutions.
00:21:48.650 --> 00:22:03.526
I completely identify with the problem solving because I'm obviously a consultant now, but I spent 10 years on the in-house side and anytime somebody got a question that they didn't know how to answer, they sent it to one of two places Corporate communications or the legal department.
00:22:03.526 --> 00:22:07.538
So you're constantly getting these crazy wild requests that you're like why am I getting?
00:22:07.558 --> 00:22:07.638
this.
00:22:07.638 --> 00:22:08.520
I don't know what to do.
00:22:09.109 --> 00:22:13.811
But if you have an LLM system that's trained on your internal documentation, it's a great place to start.
00:22:13.811 --> 00:22:17.195
Start brainstorming, I mean, you know, here's the problem, how do I solve this right?
00:22:17.195 --> 00:22:23.540
It's just a good way to show how you can be more efficient and effective and productive using these tools.
00:22:23.922 --> 00:22:25.363
Great Well, frank.
00:22:25.363 --> 00:22:28.625
You've provided us with an awful lot of information really on LLMs.
00:22:28.625 --> 00:22:39.051
Are there any closing?
00:22:39.071 --> 00:22:41.902
remarks you think that we need to cover so that our listeners can be completely not completely, but even more up to speed on these LLMs.
00:22:41.902 --> 00:22:43.407
You know, I think the key is to get out there and experiment with them.
00:22:43.407 --> 00:22:46.777
Make it a point to set some time aside and try to use some of these tools.
00:22:46.777 --> 00:22:48.461
It doesn't matter, really, which one you use.
