The Rise of Agentic AI [AI Today Podcast] AI transcript and summary - episode of podcast AI Today Podcast: Artificial Intelligence Insights, Experts, and Opinion

Go to PodExtra AI's episode page (The Rise of Agentic AI [AI Today Podcast]) to play and view complete AI-processed content: summary, mindmap, topics, takeaways, transcript, keywords and highlights.

Go to PodExtra AI's podcast page (AI Today Podcast: Artificial Intelligence Insights, Experts, and Opinion) to view the AI-processed content of all episodes of this podcast.

View full AI transcripts and summaries of all podcast episodes on the blog: AI Today Podcast: Artificial Intelligence Insights, Experts, and Opinion

Episode: The Rise of Agentic AI [AI Today Podcast]

Author: AI & Data Today
Duration: 00:13:57

Episode Shownotes

Generative AI has significantly influenced content creation, data analysis, and interactive communication, but there is a growing need for intelligent systems to perform tasks with minimal human input. In this episode of AI Today hosts Kathleen Walch and Ron Schmelzer dicuss the concept of Agentic AI, including what it is,

and why it’s becoming popular. Continue reading The Rise of Agentic AI [AI Today Podcast] at Cognilytica.

Full Transcript

00:00:01 Speaker_01
The AI Today podcast, produced by Cognolitica, cuts through the hype and noise to identify what is really happening now in the world of artificial intelligence.

00:00:10 Speaker_01
Learn about emerging AI trends, technologies, and use cases from Cognolitica analysts and guest experts.

00:00:22 Speaker_02
Hello and welcome to the AI Today podcast.

00:00:24 Speaker_00
I'm your host, Kathleen Mulch. And I'm your host, Ron Schmelzer. One of the interesting things about AI is that you find that we always come back to the same themes. It seems like we come back to the same things on some sort of repeating basis.

00:00:37 Speaker_00
I mean, hey, even AI as an idea isn't that new. It goes back to the beginnings of computing, even before computing when the term was coined in 1956.

00:00:48 Speaker_00
So you'll always see comments by people saying, wait a second, this idea that you're talking about now that's so new, weren't we talking about this one wave ago, two waves ago, three waves ago?

00:00:57 Speaker_00
The answer is probably, because there's only so many ways that we can think of machines doing things intelligently. The thing is that, of course, timing is everything.

00:01:05 Speaker_00
Sometimes we talk about an idea, but the technology we have is really not set up to actually do what we want it to do. And then sometimes things just come together at the right place at the right time.

00:01:17 Speaker_00
And that's actually one thing we're going to talk about because we're going to revisit an idea that we had talked about back in 2019 that is now becoming top of conversation in the context of everything else that's happening with AI.

00:01:30 Speaker_02
Exactly. So back in 2019, if you've listened to our podcast or followed any of our research, you know that we introduced this idea of the levels of intelligent process automation. We abbreviated it to IPA for short, intelligent process automation.

00:01:44 Speaker_02
And it really came out of this idea. So first, you know, we had levels zero through five with autonomous vehicles. And we said, hey, how can we also take other processes that we're doing and move it up

00:01:56 Speaker_02
this journey, you know, these different levels, from being fully human to fully autonomous. And this was back in 2019, if you remember, when robotic process automation, RPA, was really hot.

00:02:10 Speaker_02
And we said, okay, well, don't confuse RPA and AI and intelligence, because they're not the same. We talk about that a lot, how automation is not intelligence. And yes, automation does a lot of great things.

00:02:21 Speaker_02
but it's really just repeating repetitive tasks in the same way over and over. So you can think about, you know, maybe copying and pasting from one form to another. We talk about that swivel chair integration.

00:02:34 Speaker_02
So automation has its place and it's useful, but it's not intelligent. What could we do if we added layers of intelligence and levels of intelligence to this. And this was the idea behind our intelligent process automation.

00:02:46 Speaker_02
So we called it that because back in the day, this again was when RPA was really popular, but it's not exactly automation. And so we never were in love with that term, but we wanted the market to understand it.

00:02:59 Speaker_02
Well, now, if you've been following AI, you may have heard of this term, agentic AI. And this really is our idea of

00:03:08 Speaker_02
intelligent process automation and how we can go from the systems that we have now to being really fully autonomous and providing tremendous value.

00:03:16 Speaker_00
Yeah, I think that's like really the big insight here is that the problems that RPA was originally trying to solve really had nothing to do with AIs.

00:03:24 Speaker_00
We're saying these were automation things, but not really intelligence things, because what we're trying to do is automate the things that people were doing, working with various different user interface systems. You know, people are

00:03:35 Speaker_00
downloading, you know, people are emailing invoices to an inbox, so I have to extract the PDFs from the inbox, make sure that I'm only taking the invoices, and then I basically have to take those invoices and I have to put them into my ERP system or my CRM system.

00:03:50 Speaker_00
People were spending a lot of time doing it. They had whole jobs where their primary focus was just taking stuff. out of one system or document, putting it into another. Not a very good use of time.

00:04:01 Speaker_00
And sort of before we developed this idea of this user interface automation, this desktop automation that became RPA, we hired people to do it, usually in low-cost countries.

00:04:11 Speaker_00
This whole idea of business process outsourcing, which is like, this is a business process we've got to do. Let's find the cheapest place to do it. Send the documents there. They do it.

00:04:20 Speaker_00
Voila, it's now in my patient management system or it's in my customer management system or my payroll system, whatever it was. And of course, the big sort of innovation was that we could automate these activities using very non-AI things.

00:04:36 Speaker_00
We basically record the screen, that's not AI, or we can work out some sort of flow chart, like a diagram of the flow, and I can connect those flows to various things. That's not AI. That's a process. Or I can write rules.

00:04:49 Speaker_00
These are basically the three approaches that were used, basically, by the three big RPA vendors. Now, AI was happening at about the same time that people were doing that, and people got caught up in this word robot.

00:05:01 Speaker_00
They just implied that, oh, AI is hot, there's something to do with robots or AI, and oh, this user interface automation thing is hot, there's something to do with robots. So it was conflated.

00:05:11 Speaker_00
That was quickly, I think, people realized that when these systems aren't that intelligent, because humans need to set up the processes, humans need to handle all the exceptions, and basically humans need to handle all sorts of decision making that happens along the way.

00:05:25 Speaker_00
So when we came up with these layers and levels of process automation, we said, well, what if we could take the human increasingly out of that loop? What if machines could figure out what the process steps should be?

00:05:37 Speaker_00
What if machines could figure out what happens if there's an exception or if the user interface doesn't matter? What if machines can read the documents and understand what's in the documents?

00:05:46 Speaker_00
What if machines can basically even communicate with other machines? Well, we just happen to be in the right place at the right time now, where we have these large language models, which we've been using for

00:05:58 Speaker_00
Generative AI things where people realized hey wait a second we can use these very same systems but instead of having them just talk to people in.

00:06:06 Speaker_00
You know so you tell me what you want i'll give you what you want back you can say hey system tell me the steps that i need to do to put an invoice into my. CRM system or ERP system. And it comes back and says, we'll do these steps.

00:06:21 Speaker_00
And it's like, well, why don't you go ahead and do those steps? And you could put an instruction that says, if you run into any problems, here's the ways that you could figure out the problem.

00:06:30 Speaker_00
That's when people started thinking, like, wait a second here. We don't need these other forms of automation. We could just make use of the technologies we have now. And that sort of brings up sort of this idea of agentic AI.

00:06:42 Speaker_00
And that word agentic comes from the use of the word agents, where instead of the AI system just generating stuff, it's communicating with agents. So let's get into a little more like, what are the characteristics? What's in an agentic AI system?

00:06:56 Speaker_00
And then we'll take a look at some of the applications. We'll come back to this idea of RPA and kind of maybe why RPA may be a favor.

00:07:04 Speaker_02
Yeah, yeah, yeah. Or people are moving on. So It's called agentic AI. Some people also use the term autonomous AI. But like Ron said, it's agentic because of the word agent.

00:07:17 Speaker_02
And it's this idea that you can have these AI-enabled systems capable of pursuing complex goals with minimal direct supervision. So you don't really need the human in the loop for the system to be performing a variety of different tasks.

00:07:33 Speaker_02
And while many of these currently leverage large language models and foundation models for generating content and analyzing data, the true potential for this agentic AI is really those powerful models that, you know, it's the potential for these powerful models to have the ability to execute

00:07:53 Speaker_02
intricate multi-step tasks with little human supervision and little human intervention.

00:08:01 Speaker_02
So right now with some, you know, robotic process automation that RPA processes, it can do it on its own, but if it gets tripped up, it stops and a human needs to go in and fix the problem and then kick it back off.

00:08:14 Speaker_02
So say that this, you know, it's a multi-step process and it's kicked off overnight and it stops at step three. Well, guess what? It didn't run overnight. and now you wake up and you go into work the next morning and you're like, oh, that's great.

00:08:25 Speaker_02
It's like the one-step process stopped at step three and now I need to wait until tomorrow night to run it again because you can only run it at off hours, which is not incredibly useful.

00:08:37 Speaker_02
So wouldn't it be useful if the system could go in and figure out how to correct that error? Say a form of the field has changed, you know, the position of that field.

00:08:47 Speaker_02
Well, if you're doing that, you know, screen recording and it's changed, well, oops, that's not exactly how it looks. And these are not intelligent systems. So now I can't go in and do it.

00:08:56 Speaker_02
And I really need to be babysitting the system or maybe re-scripting, you know, re-recording. doing all of this different stuff. And yes, it saves time, but also it can be very cumbersome and you have to have people there to manage this.

00:09:08 Speaker_02
So what if we didn't need to do that anymore? And that's really where the excitement and the power comes in with agentic AI.

00:09:16 Speaker_00
Yeah. And so a lot of people who are working with the Gentic AI systems, they're not actually familiar even with RPA.

00:09:23 Speaker_00
So they're just coming at it from, hey, I want to chain together a bunch of AI activities so that we can accomplish some tasks, not just generate a picture or response to a question or create some text or create some code. I mean, all that's

00:09:40 Speaker_00
hugely powerful, right?

00:09:41 Speaker_00
But people want to say, well, let's sort of chain these things together so that one thing can tell me the steps I need to take, the next thing can actually execute those steps, and the third one can respond and take the outputs and then do something with it.

00:09:53 Speaker_00
That was actually originally the idea with things like LangChain, where we would connect together multiple LLM things to accomplish some sort of task. And people said, well, hey, I can use these chains to not just

00:10:04 Speaker_00
handle the image outputs or the text outputs and generate more image, but I can actually take it and maybe say, if you're going to generate some code, why don't you just run that code, and then tell me the results of that code, and then run something else.

00:10:17 Speaker_00
Especially if this code is something like, let's say, SQL database entries, where I could say, hey, here's an invoice, find the name field, now,

00:10:28 Speaker_00
Generate for me the SQL statement that would insert that name into my ERP system And then you say go ahead and run that and if you would encounter a problem Tell me what the errors might be and then what you would use to fix those errors It's kind of interesting and crazy kind of

00:10:45 Speaker_00
how well it would work. I mean, if we're using AI to write code, why can't we use it to just go ahead and run it? Of course, nothing's perfect. AI systems are probabilistic.

00:10:53 Speaker_00
But there's a couple of things here that really make a Genentech AI sort of interesting and different from, say, the way that you might be using generative AI today. So vanilla.

00:11:01 Speaker_00
It's kind of funny recalling a vanilla because we've only been doing it for a couple of years now. But, you know, the straight way of using generative AI right now is we sort of use generative AI as part of a process that we know.

00:11:11 Speaker_00
So we're like, I need to write a blog article. OK, well, what is step one in the blog? I mean, like get a title. And so we know the steps. So we can even ask the AI system, tell us what steps are necessary.

00:11:21 Speaker_00
But basically, we're just kind of having it do those point steps. But these systems can really fill in the blanks quite a bit. I mean, you know, LLMs are capable of handling any natural language processing tasks. So that means they can handle

00:11:35 Speaker_00
user interface stuff, they can handle documents, they can handle anything that is a digital output, basically.

00:11:42 Speaker_00
And since they're capable of doing that, not only do I not need to use like screen recording or rules-based systems, I can just prompt, I can just talk to it.

00:11:50 Speaker_00
And so we're kind of like talking and conversing with a system that's going out and accomplishing tasks for us, which is really a much more

00:11:58 Speaker_00
sort of powerful idea because in the old days, the old days of RPA, a person would have to have been some sort of RPA developer or some sort of consultant that you'd have to bring in.

00:12:10 Speaker_00
They would have to understand your process and they would have to spend some time building out the process either as you want it or as it exactly happens to be. And then if something changes, you have to keep them employed.

00:12:21 Speaker_00
And there was a statistic from

00:12:23 Speaker_00
One of the vendors that was out there in the space, the RPA vendor, said that for every dollar that's spent on license revenue for an RPA solution, something like $12 to $20 is spent on the services part, configuring the agents.

00:12:39 Speaker_00
Actually, they won't call them agents. Configuring the automations, managing them, running them. Well, what if we could take that

00:12:45 Speaker_00
big chunk and sort of get rid of it and basically say, well, we only really need the human in the loop to handle, you know, the management of the system to make sure things are running.

00:12:54 Speaker_00
And so, yes, computers aren't perfect, so maybe there still are exceptions, but we're not building any hard-coded automations. We're not, you know, hiring people to constantly manage these bots that are out there.

00:13:07 Speaker_00
Instead, we're moving away from bots to agents. I know, maybe it's another way of thinking about it. That's the big thing.

00:13:13 Speaker_00
We're stepping away from the bot perspective, where these are dumb bots just doing things that we tell them to do, to smart agents that use the power of LLMs and some reasoning and all these things where you can define goals and they can tell you the steps and you can work with them and refine the steps and it can do memory and has context and all that sort of stuff.

00:13:33 Speaker_00
you know, it sounds like it's pretty powerful to me. And I think, yeah, there's a lot of ways you could, I'm even thinking now of all the things that I wish- Oh, I know.

00:13:41 Speaker_02
I mean, and also as you were talking, I'm thinking, yeah, you're right. Now, no longer do you need to hire these expensive consultants who are skilled in how to, you know, run this one vendor bot system, this one RPA system.

00:13:56 Speaker_02
You can have people at your organization who understand your processes and understand your organization and our subject matter experts do this without those hard skills that we talk about.

00:14:08 Speaker_02
And that's why it's so important and we really have this emphasis on these soft skills because how do you bring in that creativity? How do you bring in collaboration? How do you bring in critical thinking?

00:14:19 Speaker_02
How do you bring in all these soft skills that we talk about to figure out how best to use these LLMs and this idea of agentic AI so that you really can have powerful, powerful things happen.

00:14:33 Speaker_02
Also, it's important to understand, yeah, for every $1 when you used to spend, OK, $12 to $20, so you're like, wow, OK, so I went from this $200,000 You know thing to like a million dollars or whatever it was and you're like this isn't cheap.

00:14:48 Speaker_02
We always say a guy is not set it and forget it will guess what RPA wasn't set it and forget it. You need to constantly be looking at this and I can't handle that, you know. exceptions, right? It can't handle process exceptions very well.

00:15:00 Speaker_02
But now, because this is in the hands of everybody and it's a lot easier to understand, it can handle these process exceptions. And so, you know, what can you really do and how can you move forward?

00:15:12 Speaker_02
So we've also, you know, shared how with RPA, it was very hot for a number of years, especially pre-pandemic and then during the pandemic, because we had this massive shift of everybody was in an office to everybody working from home.

00:15:27 Speaker_02
And we're like, oh my gosh, what do we do? How do we handle all of this? How do we get things? And so our PA was a big focus.

00:15:34 Speaker_02
But now, as things have settled down again, and people are figuring out how to work from home, how to collaborate, maybe some people are back in the office, or we have these hybrid situations, we've had a few years now under our belt to figure that out.

00:15:46 Speaker_02
So those aren't our immediate problems anymore. Well, we've seen a big shift, especially now that these large language models really are becoming so powerful. Prompt engineering is becoming a thing. It's in the hands of many.

00:16:00 Speaker_02
People are really seeing this on a daily basis. And quite frankly, you're using it on a daily basis for both personal and professional.

00:16:08 Speaker_02
we weren't using RPA for any sort of personal, like I wasn't building, you know, RPA bots to help me handle any sort of personal workflow that I have. It was really just in a professional setting.

00:16:20 Speaker_02
Well, now people are feeling more comfortable with using large language models with prompt engineering.

00:16:25 Speaker_02
And so we've started to see a de-emphasis on RPA and specifically some of that, you know, large investment and real focus that we saw in years past.

00:16:35 Speaker_00
Yeah, I think part of it was like there was this big movement, you know, towards low code and no code, this idea that you can give sort of what's called citizen developers, people who are maybe not traditionally in the IT role, the power to do things like build process automation.

00:16:51 Speaker_00
But I would say even for most people, That's not really... It involves changing the way you think in some ways, like, oh, I have to design a... Even though it's just screen recording or... For most people, it still feels very technical, right?

00:17:05 Speaker_00
And I think the really interesting thing about generative AI is that it's told us, it's like, if you just put people in front of a prompt, like something they can type, people will get it right away.

00:17:14 Speaker_00
Now, some people are better at prompting than others. always tips like how to generate photorealistic images on like, you know, mid-journey, and people are always providing tips for that. Yeah, of course.

00:17:25 Speaker_00
I mean, there's always, and we talked about it ourselves, we had our prompt engineering series where we talked about prompt patterns.

00:17:31 Speaker_00
There are ways, and as you start using them, you're like, oh yeah, it does a little better when I do this, and maybe some models are better than other models.

00:17:38 Speaker_00
things like that, and you learn how to iterate, and you learn how to ask it to ask itself questions, and all that stuff's good. But you know what? That's a skill you could teach people. You could teach anybody to do it.

00:17:47 Speaker_00
You could say, play with it, practice it, here are some tips. It's very different to tell people, go play with RPA. You know, it's like, it just feels very different.

00:17:57 Speaker_00
And I think, honestly, the industry was sort of, the whole industry, IT and all the analysts covering the space, so-called hyperautomation, which is just a marketing term,

00:18:08 Speaker_00
I think we're all fooling ourselves thinking that these things were actually really no-code and low-code when they were just a different code.

00:18:15 Speaker_00
It was sort of like, instead of coding, I'm doing something else, but it still kind of feels like I'm using the same part of my brain that I would be using for coding. is the weird thing about AI. I think this is the big word.

00:18:28 Speaker_00
I think you're actually using a very different part of your brain when you're using generative AI. It's not the same part of your brain for doing linear algebra when you're writing a prompt. For a lot of people, I think it feels more natural.

00:18:41 Speaker_00
Just like all soft skills that you're going to come back to it. Yes, you can get better with practice. You can be a better writer. You could be a better reader. You could be a better critical thinker.

00:18:51 Speaker_00
You could be a better communicator, a better collaborator. You can improve your creative techniques. But I feel, this is just a sense that I get, that I feel like it's just more natural for people to learn those things.

00:19:05 Speaker_00
And some people will always be better at certain skills than others than it is to inject this idea of no code So I think that's going to be the real thing, is that I think generative AI sort of really bridges that gap and puts the power.

00:19:18 Speaker_00
Now we can put agents in the power of persons. So you could type in a prompt that says, please put this in my sales system. And if you don't understand, if you're missing data, tell me what data you're missing, and I'll tell you where to find it.

00:19:31 Speaker_00
And then the system will be like, I think I could find this data in this spreadsheet here. Is this correct? You're like, yes. OK, well, then do it. Repeat it. That feels a little more natural to me. It just my impression on the whole space.

00:19:44 Speaker_02
Yeah.

00:19:44 Speaker_02
And I think another thing too, especially when it comes to prompt engineering is you can practice on your own where I don't have like a license for an RPA bot to just, and also I don't have like flows in my personal life that I need to figure out.

00:19:58 Speaker_02
Right. I don't need, I'm not like. you know, entering data from one system to another on a regular basis that I want to script this and put a bot in place to do this.

00:20:07 Speaker_02
So to come up with those use cases and those examples are quite difficult, but I have many use cases where I want to write a prompt to help me with something or where I, you know, take a picture and have

00:20:19 Speaker_02
You know, maybe chat gbd explained to me what's in the picture? We've i've been doing that more in my personal life and it's been a lot of fun to play around with it, you know, like you have a A soda bottle that has writing on it.

00:20:30 Speaker_02
That's not english say. Okay. Can you translate this for me or what is this? I've been i've been doing it a lot lately and actually it's funny because we talk about how maybe search is going away and I was out one day with a friend and we were at a

00:20:42 Speaker_02
outside of an ice cream shop and there was a flavor we didn't know the answer to, so I took a picture and I put it into ChatGPT and had it tell me what it was and explain to me what that flavor was while she searched.

00:20:54 Speaker_02
And I'm like, I feel like that behavior is going to start eventually going away.

00:20:59 Speaker_02
And it was interesting, you know, she's still searching, but I first opened up the, you know, took a picture and put it into ChatGPT so that it could provide me, you know, more details, honestly. And so it was pretty interesting.

00:21:11 Speaker_02
And I think that that's where this comes into play, right? People are just more comfortable because we say that there really is very low failure when it comes to prompt engineering. Just try, right? Just try. But I can't just try with RPA.

00:21:26 Speaker_02
First off, I can't just try in my personal life because I don't even know what I'm doing there, you know, like the flow. And then I also can't try because I don't really have access to these systems. where I have chat GPT on my phone.

00:21:38 Speaker_02
I just took a picture while I was out and around in one morning and said, what is this? Let me see what flavor this is. I can't do that with other technologies.

00:21:44 Speaker_02
And I think that that's also why, you know, yes, as Ron said, it brings up different parts of your brain. And also it just feels different. It doesn't feel like I'm coding because I'm not. It doesn't feel like I'm using math because I'm not.

00:21:55 Speaker_02
I'm just using natural language. And that really is the power behind this. So there's other examples of how people are using agentic AI systems, and it still is in its infancy.

00:22:07 Speaker_02
So I think companies and people are continuing to explore the wide range of applications that it's going to have, which is exciting because, again, this comes into the soft skill of creativity, right?

00:22:19 Speaker_02
How can I apply this and really push those boundaries? And another thing that's nice about prompt engineering, which we always say and really are encouraging people to do, is share your prompts.

00:22:29 Speaker_02
share how you're using it, share how you creatively are doing this so other people can have access to it. But different areas that we've already been seeing is using agentic AI to help with IT support and help desk ticketing issues.

00:22:43 Speaker_02
So think also what people really can do is just think about how they've been using RPA and then how they can apply it. So we know IT support

00:22:50 Speaker_02
Is a big area where that you know it's ripe for improvements and so people get help desk ticketing issues especially with handling some of those repetitive inquiries like password resets i mean you know oh my lord how many people have to constantly reset their passwords or maybe software patches.

00:23:09 Speaker_02
trouble ticketing, and also problem diagnosis, we've been seeing agentic AI being used there.

00:23:15 Speaker_02
We're also seeing it being used with human resources, especially with some of those areas that really are great for this, like handling routine payroll, or employee onboarding, or benefits management.

00:23:28 Speaker_02
And HR teams are also using Agentic AI to handle some of the wide range of workflows and administrative tasks that they have, such as handling inbound inquiries on expense reports and travel, and also providing guidance on benefits and employee-related issues.

00:23:43 Speaker_00
Yep. So, I mean, as I said, there's lots of applications here where it's sort of like the whole conversational interface happens to be kind of the way that you're already interacting anyways.

00:23:54 Speaker_00
When someone's filing a trouble ticket because they're having a trouble support issue, it starts with a conversation, and it ends with a conversation, and a whole bunch of things happen in between.

00:24:04 Speaker_00
If you think about it that way, then the range of what these large language model-based systems can do when they have agents they can work with, these autonomous systems, actually is going to keep growing and growing and growing.

00:24:17 Speaker_00
The times when we're going to see traditional RPA being used is when it's very much straight repetition over and over and over again.

00:24:24 Speaker_00
For example, especially the back office tasks, the so-called unattended agents that maybe every night they go in and they do some processing and they generate reports and do that sort of stuff.

00:24:35 Speaker_00
Yeah, I mean, you don't need to LLM conversational your way into it, especially if it's always extracting data from these sources and generating reports. But even there, I think we're going to see

00:24:45 Speaker_02
Yeah, I was gonna say, maybe you don't necessarily have to, but as people feel more comfortable with it, and they're already using these, right?

00:24:50 Speaker_02
I mean, like now, why are we gonna have all these different licenses for all these different types of technology? Maybe it's more expensive. It can be more brittle. We've talked about that. Eventually, why will you want to?

00:25:02 Speaker_02
I think people do what they're comfortable with, right? not everybody's comfortable with even that, you know, kind of low code environment. And so, yeah, why not? I already have a license for this.

00:25:14 Speaker_02
So we're really excited to see where things are going, as always, because this is AI Today.

00:25:22 Speaker_00
Yeah. So I think just in general, I mean, keeping out of it, we're going to take a look at it. We're going to be playing with it ourselves. I mean, one of the biggest things I'm looking to do is even create more video from our podcast.

00:25:31 Speaker_00
We have a lot of video right now. The podcast is mostly audio, but people are really tuning into small snippets. And well, LinkedIn is basically turning into TikTok these days. So it's really favoring content.

00:25:43 Speaker_00
So the process of generating the videos is very much can be a manual process where you can But the thing is, hey, this actually could be really good for a Gentic AI. I could say, find a good one-minute audio or video clip.

00:25:54 Speaker_00
Now go ahead and produce the video like this and put the captions on it. And then at the end, put this little graphic or something. I'm going to be playing with it, because if that works, I'm just going to have it do it.

00:26:07 Speaker_00
And then soon after every episode, we'll have it do it and produce it.

00:26:10 Speaker_02
Yes, stay tuned for our podcast videos. With the help of AI, that's augmented intelligence at its finest. Exactly. So with that, we hope you've learned something on this podcast today and you've really enjoyed it.

00:26:23 Speaker_02
If you haven't done so already, subscribe to AI Today so you can get notified of all of our upcoming episodes. We are in the middle of our Soft Skills series, so definitely stay subscribed for that.

00:26:33 Speaker_02
We also have some additional podcast interviews with guests that we're really excited about. And so if you haven't done so already, make sure to rate our podcast. We love to hear from our listeners. You can always reach out to us. Find us on LinkedIn.

00:26:46 Speaker_02
And we also have a LinkedIn newsletter as well that we encourage you to subscribe to. And I'll link to that in the show notes where we talk about, you know, different topics.

00:26:53 Speaker_02
Sometimes we elaborate on some of the topics we talk about on the podcast and other times we share different thought leadership. So definitely subscribe there as well. Like this episode and want to hear more?

00:27:03 Speaker_02
With hundreds of episodes and over 3 million downloads, check out more AI Today podcasts at aitoday.live.

00:27:10 Speaker_02
Make sure to subscribe to AI Today if you haven't already on Apple Podcasts, Spotify, Stitcher, Google, Amazon, or your favorite podcast platform. Want to dive deeper and get resources to drive your AI efforts further?

00:27:23 Speaker_02
We've put together a carefully curated collection of resources and tools. handcrafted for you, our listeners, to expand your knowledge, dive deeper into the world of AI, and provide you with the essential resources you need.

00:27:36 Speaker_02
Check it out at aitoday.live slash list. This sound recording and its contents are copyright by Cognolitica. All rights reserved. Music by Matsu Gravas. As always, thanks for listening to AI Today, and we'll catch you at the next podcast.