cover of episode EP 555: Accessible AI: Practical Strategies for Every Business Leader

EP 555: Accessible AI: Practical Strategies for Every Business Leader

2025/6/26
logo of podcast Everyday AI Podcast – An AI and ChatGPT Podcast

Everyday AI Podcast – An AI and ChatGPT Podcast

AI Deep Dive AI Chapters Transcript
People
A
Aashvarya Srinivasan
J
Jordan Wilson
一位经验丰富的数字策略专家和《Everyday AI》播客的主持人,专注于帮助普通人通过 AI 提升职业生涯。
Topics
Jordan Wilson: 我认为如果公司还在沿用10到15年前的商业策略,然后试图在最后阶段加入一些人工智能元素,效果可能不会太理想。过去,大型企业在采纳新技术时,比如互联网、云计算、个人电脑和移动电话等,往往可以慢慢来。然而,在人工智能领域,这种做法就行不通了。你不能简单地将传统商业策略与少许人工智能技术结合起来,而是需要在企业核心层面建立起一种人工智能思维模式。这意味着企业不仅要思考如何应用人工智能,更要从根本上转变心态,将人工智能融入到企业的DNA中。 Aashvarya Srinivasan: 我认为过去三年最大的变化在于人工智能技术的易用性和可访问性显著提高。三年前,我曾做过关于文本摘要的演讲,当时主要依赖BERT和GPT-2模型,但技术还很初级,只能实现简单的抽取式和概括式摘要。而现在,感觉我们仿佛跨越了十年。技术的成熟度以及大众获取技术的便捷程度都发生了巨大变化。以前,这些模型主要在研究实验室或大型科技公司中用于特定用途,但现在,任何小型企业主都可以轻松使用人工智能。这种转变的关键在于技能差距的缩小,这得益于开源模型的普及。此外,随着框架、供应商和封装技术的不断涌现,终端用户可以更轻松地使用这些技术,从而提高了生产力,改善了企业的底线。

Deep Dive

Shownotes Transcript

Translations:
中文

This is the Everyday AI Show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business, and everyday life.

This podcast is supported by Google. Hi folks, Paige Bailey here from the Google DeepMind DevRel team. For our developers out there, we know there's a constant trade-off between model intelligence, speed, and cost. Gemini 2.5 Flash aims right at that challenge. It's got the speed you expect from Flash, but with upgraded reasoning power. And crucially, we've added controls, like setting thinking budgets, so you can decide how much reasoning to apply, optimizing for latency and costs.

So try out Gemini 2.5 Flash at AISTudio.Google.com and let us know what you build. If your company is using the exact same business strategy it was using 10 or 15 years ago, and then you're just trying to insert some AI at the end,

might not work out too well for you. I think that throughout the course of the last 20, 30 years, enterprise companies could sometimes take their sweet time when it came to tech adoption, when it came to implementing things like the web.

cloud, PC, mobile phones. You can't really do that with AI. You can't take your old school traditional business strategy and just sprinkle a little AI on top. That's not how it works anymore. I think you have to have not just an AI thinking, but you have to really have an AI mindset at the core of your business.

All right, I'm excited to talk about that today and a lot more on Everyday AI. What's going on, y'all? My name is Jordan Wilson and I'm the host. Thank you for tuning in to Everyday AI. This is your daily live stream podcast and free daily newsletter, helping us all not just keep up with what's happening in the world of AI, but how we can all actually get

get ahead to grow our companies and our careers. So it starts by what you learn here on this podcast and live stream, but it is actually, that's just part one. You have to, for part two, go to our website at youreverydayai.com. There, we're going to be recapping in our daily newsletter, some of the best insights from today's conversation. And our guest is amazing. I can't wait to bring her on. But also in that same free daily newsletter, we're going to keep you up to date with

everything else that you need that's happening in the world of AI. All right. So enough from me, enough chitchat. Please help me welcome my guests. There we go. Aashvarya Srinivasan. Aash, thank you so much for joining the Everyday AI Show. Thanks, Jordan, for having me here. And I'm very excited to be a guest on your podcast.

And as somebody who is listening to your podcast every day, I really enjoy it. And I'm hoping that I'm able to add as much value as your other guests.

Oh, absolutely. Absolutely. So let's talk a little bit about your background because it's impressive. It's like, you know, looking at your, you know, your LinkedIn profile is like looking at, you know, every single big tech company out there. So right now, you know, head of AI developer relations at Fireworks AI. So tell us a little bit kind of what you do there and a little bit about your background in AI.

Yeah, absolutely. I wouldn't say I have had a non-traditional path towards what I'm doing right now. Started off as a software engineer, minored in machine learning, eventually got into doing a master's in data science. So I've been in this machine learning AI space for quite some time, much before like the generative AI large language model hype started and

That's another thing like I love doing traditional machine learning and I still love continuing to do that and it's just fun to see that there is a whole new world of not just people building large language model systems and general BI systems but also the very very growing economy of users who are interfacing with it.

It's just fun that now my mother is more interested to learn more about what I'm doing. Thanks to ChatGPT compared to like what it used to be probably a decade ago. So yeah, it's good that a lot of people are understanding what AI is, what the use cases are. So it's definitely a very fascinating time to be.

And even kind of how I started out the show and as someone that's been in machine learning for a while, how do you think it's even personally changed in terms of business priority? I know for companies that had been data first, companies that have been using artificial intelligence and machine learning for many decades, maybe it hasn't changed as much.

But for every other company, right, it seems like to me, at least someone that talks to a lot of people from the outside, it went from, ah, what's AI, not for us, to all of a sudden it's one of the most important things at the core of their business strategy. So talk a little bit just about how it's changed in the industry as someone that's worked in it since pre-ChatGPT.

So I would say the biggest difference that has happened three years ago till what is today is simply put ease of access and ease of use.

Three years ago, I had delivered a session about text summarization, about how companies do news text summarizations or blogs or larger books. And at that point in time, it was BERT models and GPT-2 model. And it was still very naive in the sense that it was extractive summarization and abstractive summarization in the very, very simplest form. And now in the last three years, it just feels like we have crossed a decade.

already in just like a short span of three years. And it's just the level of maturity of the technology, of course, but also how easily it's accessible to a lot of people has changed. So earlier, the reason why, say, traditional machine learning models or even like the early phases of large language models or like the BERT models and like the early GPT models

They were very narrowly used in certain research laboratories or certain big tech companies for very specific use cases. But now what is the biggest change I would say in the last three years that has happened is how easily this technology is accessible to people and how easily they are able to build up on top of it.

Any small business owner, I was even speaking with a couple of small business owners and how they can start using AI. And it's just how easily it's available to them, right? So if you think about this stack of users and builders and providers,

The earlier stack used to be very, very technical, which is like if you want to use a GPT-2 model, imagine the level of complexities you have to go through to really put it in production. So that itself and like the fact about like the skill gap that you would have as a small business owner or a college grad who has a great idea but doesn't know how to like really use these models in practice to build their product has changed. Now that

it's easy to get access to it thanks to all the open source models. Second, it's just become the use, the usability has increased. So with more and more frameworks and providers and people who have encapsulated it with the terminology of wrappers. So,

it's just available to the end users in a much easier fashion and that's why we are seeing a lot more people using it that's why we're seeing tools like auto ai which i'm using for like translation um or like you know fireflies which is like doing your uh meeting meetings recordings and summarization or like node gpt or all of these toolkits right like that we're using every day it's not just it's not just users uh

from like a technical background, but it is everybody who is trying to like use these technologies to increase their productivity. And at the end of the day, it improves their bottom line, which is increased productivity is faster results, better business value.

So, I mean, one thing that you talked a lot about there, Ash, is just this accessibility, right? And not just for consumers, but also for businesses, right? And how easy it is now, right, to build on top of this amazing technology, right? Between Google and OpenAI and Microsoft and

Claude and everyone else making it so easy for even non-technical people to kind of do some development work, right? But when, like, how does this impact actually what businesses are doing, right? Like, I think that's something, you know, a lot of business owners are still, you know, struggling with. They're like, okay, does this mean we should be, you know, building on top of it if we don't necessarily have a use case, but because it's easy and maybe we can, you know, a

create new lines of revenues, right? But how should, you know, business owners, entrepreneurs, you know, people at big companies, how should they be thinking about their business strategy, maybe differently with this accessibility and the usability, the two things you mentioned?

This podcast is supported by Google. Hey, everyone. David here, one of the product leads for Google Gemini. Check out VO3, our state-of-the-art AI video generation model, in the Gemini app, which lets you create high-quality, 8-second videos with native audio generation. Try it with the Google AI Pro plan or get the highest access with the Ultra plan. Sign up at Gemini.google to get started and show us what you create. ♪

That's a great question. And that sort of brings me to the point that I was chatting to you about earlier, right? There is a perception and there is a reality, right? The perception is that with these AI, Gen AI, LLM, like these terms are being used interchangeably. With these tools, there are certain possibilities that you can unlock. There are some really cool things that you can do.

And they come with a bunch of features, right? With features that seems very interesting, seems very fascinating, seems almost impossible if you were talking about it a few years ago. So that's the perception of, you know, like how big of an impact it can have. But when it comes to reality, which is about how businesses are using it, I would say it's the same mindset in the sense that

businesses think about their bottom line, right? At the end of the day, small businesses, large businesses, universities, organizations, whoever is planning to use AI toolkits, it's about getting to their bottom line.

Now, how do you get to that bottom line can be in using different channels, right? So the way that I would recommend people to think about it is right now, as simple as, you know, breaking down your everyday tasks, right? If you have 200 employees in your company, what are these 200 employees spending their time on? What are the parts that are highly critical, which is focus?

really risky and requires critical thinking which requires human decision making and what are the parts which do not which are mundane which are repeatable which can have a for the things which can where we can have a risk averse mindset versus not so those are like some of the internal evaluations any company needs to be doing

And that itself will drive them towards what AI tools to use. I think what happens a lot of times is people are trying to approach it in the opposite direction, which is like, hey, I read about this AI tool. How can I use this in my company? It should be the other way around. What are the things that requires a fix? And then you go about thinking what's the right tool for you. Because at the end of the day, I think what people need to understand is that AI is a tool kit.

It's the same thing like a hammer or a drill or a saw. And every tool has a purpose. It cannot solve every single problem. And that's the exact same thing with the app. It cannot solve every single problem. It has its own challenges. It should not be used in a set of use cases. So if you approach it from the standpoint that, hey, I want to use this tool, where do I fit it? That's going to take you a lot more time to figure it out rather than...

taking the right approach, which is like, here are the problems and what are the right tools for me to fix it? I think what you just said right there is like a perfect way to hammer home this point. It's something I say a lot. So I'm glad that, you know, someone with your background is saying it as well, you know, because I think sometimes people are always looking too far forward and saying, what can we do now that we couldn't do before? But Ash, what you just said right there in this example is saying like saying, what are our 200 employees doing?

spending their time on now, right? Because sometimes I think you have to work backwards because you can't go on the same path you were going if the way those 200 employees work has fundamentally changed, right? - Exactly. - And then another thing you just talked about there is the toolkit. So how can decision makers kind of find the balance between those two things? It's like, if they look backwards and say, "Oh, our 200 employees, now how they're spending their time

it might be a really bad way, right? Because it's very inefficient. But then you look at this, you know, AI is a toolkit. It's an ever evolving toolkit as well. So where do you kind of, you know, find the happy medium between those two things?

So

Starting with, what are the things which will give you a faster return on investment? Now, the return on investment could be different things for different businesses. For some, it would be how is it going to free up my time? How is it going to make me better in my decision making process? How is it going to serve my end customers better? So the metric that you use to define this return on investment could be different for different businesses.

But as I said, always start with like, what's the value proposition that you're going behind? And how can you, what exactly are you trying to solve? And that will guide you towards the right toolkit. Because a lot of times people come back to me and ask, should I be using a Lama model or a Quinn model or this or that or a small or a medium model or like a 400 billion parameter model? Well, it depends on what are you trying to solve.

So it all goes back to the question of what are you trying to solve? And the best part right now is that the literacy or like the access to information has become so much easier that you can do a lot of experimentations by yourself. Things that would require a team of like five to ten people, like as giving an example, right? Something as simple as I want to launch a company similar to Airbnb, right?

Historically, if you were speaking, you'll need like a team of front-end developer, back-end developer, product designer, creative person, database manager, all of these individuals whom I need to even get to the first step of starting my company. But now I think with the AI tools which are accessible to us, we can get to the MVP part very quickly. So you can develop what a prototype would look like and

Getting to that point has become really, really fast. So for you to get to market, for you to build out something and test out something, it's become very easy. So for you to learn a particular tool, there are so many online resources which you can access, learn a particular toolkit and get started. So that's how I would recommend like any company, if you are a 200% company, first evaluate where exactly is your effort going in? Why?

What are the bottlenecks? What is the bottom line for you? Revenue is one thing, but then what are the other proxies for those bottom line? And how can you improve that in a better fashion? Like, can you save people some time? Can you make them work on new products? Can you diversify your business or can you expand your business? How far can you scale your business? So all of those things will direct you in

What sort of tools can help you in achieving those goals? So, you know, when I hear you talk, I can tell that you have kind of this, you know, AI mindset and AI thinking down, right? But that's because you have a background in machine learning, but, you know, not everyone does. So, you know, how can they kind of,

address both, you know, you said, hey, going back, finding the bottlenecks, you know, fast ROI, right? Like, how can they go back and make those decisions or look forward to new opportunities and make those decisions if they don't, you know, have that kind of AI thinking already, if they don't have that AI mindset? Because I'm trying to keep up with the

The hype cycle of AI is impossible, even for someone like me that does it every single day. So how can those maybe non, people that don't have a decade of experience in machine learning, how can they still have that AI thinking or AI mindset? I'm going to give you a non-technical answer to this. I love it. And it goes back to a conversation I was having with my friend because I was talking to my friend who wanted to like...

go over like a diet plan for you know like a healthier diet plan for his day-to-day eating eating stuff and like improving his lifestyle and I think the the part like that I'm getting to in this conversation is like how do you break that inertia of thinking in the standard style that you always do right I cannot really point back to the time when it started for me but then now for every single thing I just go back to an AI tool to help me make my decisions better

How can I put out better videos? How can I improve my writing? How can I improve my email styles? How can I improve my product design documentations? For every single thing that I try to do, I try to consciously make a decision to not let inertia take me in the standard way, the way I used to work.

So giving you a simple example, right? Like if somebody tells you that, hey, I want to lose weight or my goal is to gain weight and I have like XYZ dietary restrictions. So how do I go about it? I have to like go contact a nutritionist. I have to do this. I have to do that. It's an elongated process. And the way that people...

that people think about it is in this particular manner like which is how we have been doing things historically but now with AI tools a simple change is that I can chat with chat GPT and tell it that hey this is my goals this is this is my current weight this is my current workout regime this is what I can eat I cannot eat this is how busy I am so help me develop a day-to-day eating chart the amount of time that it would take for me to find a dietitian and go through the process

is now completely solved by this particular prompt that I've sent to ChatGPT, right? And I'll give you another example, the way that my mom's life changed because of ChatGPT. Again, like this is not a sponsored video or like, you know, I'm not really endorsing any particular tool, but

My mom uses a lot like social media in different formats. She's like watching YouTube videos. She's reading news. She's using Facebook. I mean, she's probably one of the very few people using Facebook. But she's on all of these platforms, right? And she hears a lot of information. She hears different news narrating certain side of the story. And she used to get triggered reading something. And she would be like, oh my God, I read this. And like, you know, it's really disturbing. And or like she would get worried about something.

And now with tools like ChatGPD or Perplexity, you can just go and clarify certain things on the platform. So now every time my mom reads something, she knows that there is a possibility that this news is fake or there is a possibility that it's only being narrated from one side or it's not really the complete side of the story. And she goes and uses these AI tools to get answers for herself. So

Now she has moved and she has her own AI mindset because now she is familiar with all of these tools and she knows how she can use that in her daily life. And that's, again, going back to the fundamental way of how humans work. We need to get

away from our inertia, get away from the autopilot of how we have been historically doing things and just try out something. It might be a learning curve, maybe for a few hours, for a few days, but then that learning curve will really set you up.

I love that. Just not letting the inertia take you the standard way. I think that's just a really great way to think about general, like traditional business strategy, right? It's repetition, it's automation. It's just your brain's on autopilot sometimes from not AI ways of thinking. I'm curious, you already kind of gave a couple

You know, simple examples, you know, through your own life, through your mom's life, you know, but even how how has it changed, you know, how you work, right? Like, you know, just even your, you know, business strategy mindset, you know, having, you know, generative AI even versus, you know, traditional AI or ML.

So I posted about this a while ago that growing up, I wanted to be an artist. I wasn't wasn't very I mean, like I didn't even know about engineering when I was probably like four or five years old. I used to love art. And obviously, I have not been in touch with art for like quite some time. And I had this wild idea of teaching people about AI topics using comics.

So my first version of comic took me probably eight to nine hours to build out from ideation to finding the right tools to do it, to put the storyline together, to put the imagery together and build it out on Canva and then reproduce it. The second time I did it, it took me around five to six hours, slightly less than the first time. But then now I'm going to release another comic.

comic around quantum computing. So it's teaching people about quantum computing using AI comics. And it took me less than 30 minutes to build it up. Wow. So that's the level of productivity I'm talking about. And I don't want to address this because I have heard this from a lot of people that, hey, why are we trying to like use so much of the AI tools? And like, is it like a huge threat against people of employment? What I would say is,

We have had these sort of like thoughts about threat during the industrial revolutions historically. But a simple example is what I explained, right?

Earlier, I used to do 10 things in a day. But now with AI tools, the time that it takes for me to do those 10 things has drastically reduced. Does that mean I'm unemployed for the rest of the time? No, I figured out more things to do in my life. So that's the exact way I would think about not just humans, but also businesses.

If you have a 200% business, rather than thinking about, hey, I'm going to use AI to cut down on what people are doing and fire 50% of the staff, think about how you can use the rest 50% of the staff, upskill them and grow your business.

I think that's a great way to put it, right? And I love the example of the comics going from eight to nine hours to six to now just in minutes. Yeah, just what's possible is changing so quickly with AI. And I think that does really impact everyone's business strategy. So Ash, we've talked a lot

in today's conversation in a short period of time. A lot of super helpful insights. But as we wrap up, what would you say is the one most important piece of advice that you have for business leaders to really just adapt an AI mindset or AI thinking and apply it to their business strategy? I would say try AI tools overwhelmingly as a person rather than thinking about yourself as a business owner.

Start using whatever comes to you. Like whatever you read about on LinkedIn, on newspaper, on X, on threads, on Instagram, wherever. If you come across a tool and if it remotely fits whatever you're doing, try it out. The only way that you can understand the pros, cons, like the features, capabilities of any tool is when you try it out. And as soon as you give it a try, the entire overwhelming feeling goes away.

So that's the best way to like get started. The more you try, the more you get hooked, the more you understand the value and the more you, the more you also get a idea of how to filter good from bad.

how to filter hype from reality. So that also helps a lot. So just give these tools a try. Love it. Great advice. I think for a lot of people that are struggling to make sense and keep up with all the developments that are happening, Ash, I think this was a great conversation. So thank you so much for taking time out of your day to join us on the Everyday AI Show. We really appreciate it.

Thanks, Jordan. All right. And as a reminder, we covered a lot there. There's so many quotables, so much good advice. If you missed it, don't worry about it. We're going to be recapping it all in our daily newsletter. So if you haven't already, please go to youreverydayai.com for more on this show, more on what you need to grow your business, grow your company and career. Thank you for tuning in. Hope to see you back tomorrow and every day for more Everyday AI. Thanks, y'all.

This podcast is supported by Google. Hi folks, Paige Bailey here from the Google DeepMind DevRel team. For our developers out there, we know there's a constant trade-off between model intelligence, speed, and cost. Gemini 2.5 Flash aims right at that challenge. It's got the speed you expect from Flash, but with upgraded reasoning power. And crucially, we've added controls, like setting thinking budgets, so you can decide how much reasoning to apply, optimizing for latency and costs.

So try out Gemini 2.5 Flash at aistudio.google.com and let us know what you built. That's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going. For a little more AI magic, visit youreverydayai.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.