As generative AI enters the workplace, fear of this new technology has crept into the minds of not only the average worker, but CEOs and other high-level leaders. Just how far will AI go? Will it take over everything and can it do the job better than a human can?
Kavita Ganesan, AI for business expert and the author of The Business Case for AI, wants to not only alleviate these fears, but show business owners how to leverage AI to enhance their business strategy and increase productivity. Stimulus Technologies CEO, Nathan Whittacre, discussed AI for business with Ganesan on the company podcast, Stimulus Tech Talk.
Because AI technology is emerging and constantly expanding, many business owners are unsure how it can fit with their business.
What is Generative AI?
In short, generative AI turns existing data into other types of content. ChatGBT for example pulls existing data from the web and turns into content based on prompts from the user. The more people interact with ChatGBT the more it learns. For example a user can tell ChatGBT to write a blog post in the style of Carl Sagan, Ernest Hemingway, or even Dr. Suess. Generative AI can also be used to create videos and graphics.
While ChatGBT is an open source, meaning any available source can be pulled from, there are also closed sources where a company or individual could store their own intellectual property and create content based only on their own material. That material would not be available to others to use unless it was already public.
How Can AI Be Used in Business?
There many business uses for AI as Ganesan discusses in the podcast. AI can be used to repurpose content for marketing and communications, to analyze existing processes, for customer service, for quality assurance, just to name a few applications. Ganesan cites some examples of companies using AI to alert customers of possible fraudulent credit card use or to alert employees of a company to anomalies in a process.
There are many applications where AI can be used. For business owners and company leaders to get the most out of AI they should determine where the biggest needs are, what the investment will be, and what return they can expect. As a consultant, Ganesan is experienced in helping businesses make these determinations and implement the new process.
Will AI Cause Job Losses?
While AI will certainly change the job landscape, Ganesan sees AI as something people will work with and use as a tool. About fear of job loss, she states, "It's going to augment, can augment, your workflow and make you a lot more productive."
She stresses a mindset shift when it comes to the workplace as well as "upskilling" employees so they can be prepared for a changing workplace.
For more on this important topic, be sure to listen to the entire episode and you can read or download the transcript below.
Stimulus Tech Talk Episode 14: Harnessing the Power of AI to Boost Your Business Strategy with Special Guest Kavita Ganesan transcript
Intro 0:00
You're listening to Stimulus Tech Talk. A conversation based podcast created by stimulus technologies covers a range of topics related to business and technology.
Nathan Whittacre 0:09
Hello, welcome to Stimulus Tech Talk. I'm Nathan Whitaker, CEO of Stimulus Technologies and we're very excited to have our guest Kavita Ganesan. Today. She is an AI advisor, strategist, educator and founder of Opinosis Analytics. Kavita works with teams across the organizations to help them integrate AI strategically and get meaningful outcomes from any every initiative.
Nathan Whittacre 0:39
So just a little bit about more about Kavita. She has over 15 years of experience and scale to deliver multiple successful AI initiatives for large companies such as eBay, 3M, GitHub, and McMaster Carr as well as smaller organizations. She has also helped leaders and practitioners around the world through her blog post coaching sessions and open source tools. Kavita also holds degrees from prestigious computer science programs, specifically a master's degree from the University of Southern California and a PhD from the University of Illinois
Nathan Whittacre 1:12
at Urbana Champaign, with a specialization NLP search technologies and machine learning. Kavita has been featured by numerous media outlets, including Forbes, SEO world CMSWire, Verizon, CD Times, Technopedia and Ted magazine. So welcome, Kavita.
Kavita Ganesan 1:31
Thank you for having me, Nathan.
Nathan Whittacre 1:33
You're welcome. So AI has definitely become a big topic in the news recently. And we've had a couple of sessions on AI and I ran into your book a couple of months ago, and was very excited to reach out to you find out a little bit about how businesses are using AI. So it sounds like you've had a lot of experience working with companies on integrating AI into their businesses and their business models. So tell me a little bit about yourself and kind of your background and what you do more than more than the traditional bio.
Kavita Ganesan 2:07
Sure. Um, so my history with AI goes way back to like 2005, I was actually introduced to AI during my master's program. And that was more on the academic research side of things. So I've done a lot of AI research. And then over the years, I got my PhD, and I was planning to become a research scientist or professor. But that's the time data science as a field really started to take off. So if you remember 2011 to 2013, is where the big data
Kavita Ganesan 2:44
revolution started to happen. So instead of becoming an academic, I decided to go and solve a bunch of industry problems. So that's where I got exposed to 3M healthcare, I worked for GitHub. And I saw different types of AI problems, put a lot of models in production. And I learned a lot in those years. And as I was doing that, other companies were approaching me to help them implement AI in their organizations, because it's a very, it was a very new concept at that time.
Kavita Ganesan 3:17
So they wanted help on how to integrate AI into the business or come up with AI startups. So I got more and more into consulting roles. And I really liked it. And now I do a mix of consulting, training, and advisory type of work. So I came from the academic side research, and now I'm like, purely AI for business type of person.
Nathan Whittacre 3:42
So it's interesting. I mean, AI is a very developing field. How is it you know, working in a field that's constantly changing and evolving? And, you know, how, I mean, how does that work in your learning? Because there's no like, textbook, you're kind of writing it as you go along. Right?
Kavita Ganesan 4:00
Yeah. So as long as you have the foundational skills, building on that skills is not hard. So the last few years, LLM became a really popular thing. But because I have all this background in AI in NLP, and understand how things work, so building on that knowledge was not difficult. It just, it takes me some time to read some papers, reading through blog post, just really understanding what it is, how it works, and how I can use it with my clients and also like going deep and experimenting with things. So so the foundation is very critical, whether you're on the technical side, or on the non technical side, so having a breadth of understanding of what AI is would be crucial.
Nathan Whittacre 4:55
Ok ay, interesting. So, I would imagine, you know, 10 years ago, 15 years ago the community that was working on AI was pretty small. Not a lot of organizations or researchers that were working on it it. Is it grown quite a bit, or is it still a small tight knit community working on these these problems?
Kavita Ganesan 5:16
So I would say on the recent side, it's fairly still a small community. It's the big tech companies, those universities and those well funded AI research labs. But on the industry side, I've seen significant growth and people wanting to somehow use AI for something. So I even had a client who approached me for AI for storytelling, so he wants to integrate it into his courses. So that growth has been quite significant the last six months, I would say.
Nathan Whittacre 5:51
Interesting. Do you see do you find, you know, as was maybe 15 years ago, you know, that was a big marketing buzzword? Do you find when working with businesses, it's, you know, AI is kind of a buzzword that people want to attach and they're really not doing AI? They just want to attach that label on it? Do you find that?
Kavita Ganesan 6:12
All the time. So that happens more often than I'd like to see. So a lot of people think that they are doing AI, but they're just talking about AI. They're experimenting with things like ChatGPT. But they don't really have, I would say, a plan on how they're going to come up with the AI product, how they're going to evaluate it, and how they're going to release it. So the product plan is not there. They just know that they want to use AI to do a demo. I see that all the time, especially with AI startups, they don't have a product design in place. So yeah, I see that as a problem. But clients who do succeed have a plan on where they will use AI. So I have a client currently, he knows exactly, he needs AI in product recommendation systems. So that's where I'm helping him develop that solution. So his approach is I have a design, I have this area where I need AI, and I'm going to hire an expert to help me with that. So he's set himself up for success. IQ.
Nathan Whittacre 7:21
Okay, perfect. So I Yeah, it's it's an interesting problem where you know, people, you know, want to jump on the buzz, and then they're not. So that's good ideas and tough to implement. So, I mean, you know ChatGPT, obviously, is big in the news. But what other what kind of tools are businesses using, because I'm sure they're not developing an AI infrastructure from scratch, I mean, what what kind of software and tools are companies integrating into their systems,?
Kavita Ganesan 7:52
I would say this a range of products, like Amazon has a lot of machine learning tools like AWS transcribe. Other tools to understand text data. So companies are using those types of prepackaged tools. And some are developing from scratch, because a lot of AI problems don't need ChatGPT, they need simple steps. So all of that still needs to be integrated from scratch, specific to the, to the product needs. And it's not heavy machine learning or anything of that sort. But it's still some form of NLP some form of intelligence in the product. So it's, it's a mix. And the new AI startups are the ones who want to use ChatGPT generative AI in some capacity. And if you see the blog post, or LinkedIn posts, that that you may see on in your feed, that's typically the researchers putting out tools about new generative AI tools, or people just curating tools, but I don't think those are actually being used in production systems as widely as you can.
Nathan Whittacre 9:06
Good. Again, back to ChatGPT. We had a guest on a couple weeks ago, talking about issues of copyright with, you know, things like ChatGPT, and I'm sure it's a you know, big question is, you know, copyright intellectual property issues. And, you know, he mentioned, and I think you mentioned it also that using a private stack or private IP, you know, machine learning versus going to a public database like ChatGPT may circumvent some of those issues with intellectual property and copyright.
Kavita Ganesan 9:40
Yeah, and a lot of companies don't want to deal with that problem, like uploading things that are proprietary to them to a third party service, and costs for startups cost is also a big problem. So each call that you make to the API costs you money, so just testing a and evaluating and then finally integrating everything to customers is the cost are going to really add up. So that's not something they want to tolerate. So the problem is small enough, they just want to build it themselves.
Nathan Whittacre 10:14
Perfect. So, I mean, if it's a business leader, not a, you know, an IT professional, I guess, but if it's a CEO, or CFO that's evaluating a AI project, you know, in the small to midsize business range, you know, what kind of things do they need to know about AI? Is it a quick ROI in the investment? Or is it a long term investment? You know, what are what are some gotchas?
Kavita Ganesan 10:38
Yeah, so the there are two parts to this one is finding the right AI opportunities, because there may be 100 different opportunities in your company. But some of them may have very marginal benefits from using AI. So finding those high impact ones is critical, because that will show you where your competitive advantage is really going to be. So let's say a lot of your problems are in customer service. So then you can think about how I can make customer service, my competitive advantage, and maybe enhance all the workflows in that department with different AI solutions. That makes sense. So that's one area. But for the long term, if you want to be able to repeatedly use AI and deploy AI systems, you have this long term planning also involved, like thinking about your data infrastructure, a lot of companies collect data, but the data stores are sometimes in silos, sometimes you cannot access the data. And all of that become problems, when you're actually trying to like build models, or even fine tune existing models, you need that data source. Some companies may not be collecting data as aggressively, so they need me to think about how to get that in place. So then there's the cultural elements. So you know, a lot of people are fearful of being replaced by AI. And this even happens at the CEO level. So I had a CEO approach me and say, I'm afraid, I'm going to lose my job, I'm going to lose my company because of generative AI.
Kavita Ganesan 12:18
So addressing those types of fear, and how AI fits into the picture, is important, like AI is not going to take away your job, but it's going to augment can augment your workflow and make you a lot more productive. So having these types of conversations is going to be important because it's a mindset shift. And you also need to think about how you upskill your employees. So everybody in the company needs to have a general understanding of AI, some of the engineers need to get those AI skills, because they are the ones who are going to be integrating these solutions. So upskilling existing employees is also a crucial piece. So I would say focusing on all of this is for the long term. But in the short term, you also need to know where the opportunities are. So you can plan you can take these data points and start creating comprehensive plan.
Nathan Whittacre 13:15
Can you I would imagine that, you know, AI is, as you mentioned, a product productivity enhancer, you know, we've seen in the US that productivity increases have flatlined in the last few years as we've come out of COVID. And that's what I've read, you know, the biggest advantage of AI is is increasing productivity of employees. Those that embrace it. Can you give us an example maybe of a company that's implemented an AI strategy that has seen an increase in productivity or an increase in in workforce capability?
Kavita Ganesan 13:50
Yeah, if you take the credit card companies, for example, if you take the fraud detection department, so they use a lot of AI, and they deal with high volumes of transaction, and each transaction needs to be flagged for different occurrences of fraud. And that in itself, just using AI in that workflow significantly reduces the need to hire a lot more employees than they they are really using right now. So that is a very good example of how AI can really increase productivity because you don't have to manually verify each and every transaction. And the AI system flags a transaction. And maybe a human reviews the transaction. And if it's a high confidence flagging, then they don't even need to do that. It's automatic, so the customers then alerted. So that's one area where I've seen good use of AI .Another is in manufacturing defect detection. So companies like Seagate, they use AI to detect defects in silicon wafers. Now, it's very, very tedious to detect defects in silicon wafers, because it's so microscopic. But with the use of AI, specifically computer vision, it's able to detect those little defects and highlight where those are. And then, instead of doing this in a matter of days, it does it in like an hour or so. So imagine the amount of throughput that you're getting from this AI systems. And also, you don't have to stop the production line as often because things are just, it just keeps going on, you don't have to stop, inspect, and then proceed with the manufacturing process.
Nathan Whittacre 15:55
So if a company wanted to start investing in AI, what kind of team or you know how to how do you get started, because that's, I mean, that's the scary part for a business, where to invest and how to invest in it.
Kavita Ganesan 16:08
I think that strategy will look different for small businesses versus large enterprises, small business, you have to look into where you're doing a lot of manual work, where your processes can be enhanced. So look for opportunities like that, and look for tools to augment your workflow, it may or may not be AI, it could be just software automation. But if it helps you, then why not look into using it and evaluate it for a period of time. So look for those inefficiencies. But for larger enterprises, there's a lot more planning involved, because the stakes are high. So once you deploy an AI system, it has to create value for the business, otherwise, you're losing money in the long run. So finding those opportunities is very critical. So you want to look for problems, complex problems, that require human like decision making. And it's also high volume. So you want to find really high volume problems, where the use of software automation can produce some sort of tangible benefit, maybe reduction in time to perform a task, or reduction in errors. So something tangible, something tangible and measurable. And also the planning so that this AI type planning is not just you should not just try to approach it because of AI, but it's needed, because at some point, you may find an AI opportunity. And you may want to start looking into using AI systems. So that planning has to start now, especially on the data side. And the cultural side, we have to come to terms with AI is going to be within businesses. So let's get started.
Nathan Whittacre 18:00
Perfect. So you have a book. Want to talk about your book a little bit?
Kavita Ganesan 18:05
Sure. Yeah.
Nathan Whittacre 18:07
So what, you know, why would somebody want to read your book, but you know, how does it help an executive or business owner understand AI?
Kavita Ganesan 18:18
Yeah, so it's tough. So my book comes from a place where I'm trying to bridge the gap between the technical world and the non technical, the CEOs, the directors, who may not have much coding experience. And I've seen that it when there's mismatch expectation on the leadership side, AI systems tend to fail not because the models are failing. It's because the initiative, there's a, there's a misunderstanding of what AI systems supposed to do, and what it's actually doing. So I've seen many AI initiatives going sideways because of that. So what my book does, it starts with the basics. It talks about what is AI from a business perspective, provides use cases that where you could use AI, and why that makes sense for AI. And then it goes into more deeper concepts like, so how do you plan for AI? How do you prepare your organization? What are the five preparation pillars? So I talk about that in my book. And I also talk about how you find these opportunities. So what I spoke about earlier, this complex decision making problems that are high volume, so I provide a step by step framework on how you find those opportunities. How do you rank and how you prioritize? So I wanted to create something that's repeatable approach so that everyone in the organization can use the same approach, and they have a way to collaborate using that approach. And finally, I talked about build by considering cause durations, and also how to measure AI success. And what does that even mean? So does it mean model success? Does it mean, just measuring ROI? So getting that clear in an executives mindset will help them have better expectations of what AI systems are supposed to do for them. So that's, so that's what it covers.
Nathan Whittacre 20:23
Perfect. And how you, I just wrote a book, my books coming out in a couple of months. And yeah, tell me about your process. How was it? How was it writing the book? It's, it's interesting writing a technical book for non technical people, I did the same thing. So for you, yeah. What was that, like?
Kavita Ganesan 20:41
Um, I would say I spent a lot of time taking the knowledge that was already inside of me, and then trying to codify it. And then running through those steps and making sure it works in different cases. So kind of like a test case for programming. So you run it through different test cases. So I, I spent a lot of time developing those frameworks and making sure that they actually, is something that these are things that I do intuitively, without thinking. But for somebody else to repeat it, I need to make sure it makes sense. So the draft took me around six months, the first draft. But editing was took me about a year to do a lot of back and forth and back and forth.
Nathan Whittacre 21:34
I had about the same experience, I thought, Oh, I'd have it done in about three months and a half later, we were ready for publishing. So it's it's an interesting process. But certainly fun. It makes it challenges you.
Kavita Ganesan 21:50
It challenges you It clarifies your thinking. It helps you think about others and how they may be perceiving your message. Yeah.
Nathan Whittacre 22:02
Excellent. And we'll be sure to include a link to your book, for those that are listening to the podcasts that might be interested in picking up a copy from Amazon or another reseller. So excellent. And if somebody you know is interested in AI and wanted to get a hold of you, what's the best way to get in touch?
Kavita Ganesan 22:19
Yeah, you can visit my homepage, kavita-ganesan.com, or my company website, opinosis-analytics.com. So these are the two places where you can get a hold of. And you'll also have some free chapters to my book on those websites.
Nathan Whittacre 22:38
So we really appreciate your time. And hopefully, you know, those that are listening are looking for strategies to improve their businesses. And I think AI is a great way to increase productivity, even if it's solving small issues. Like I love your example of, you know, from a customer service aspect or, you know, analyzing data as they're coming in able to produce those results. So that's, I mean, businesses need that today, especially with it, you know, trying to find good employees and increasing, you know, their productivity without having to increase their workflow force. So it's definitely a great, great strategy to move forward with it. So good. Thank you so much for your time. I appreciate it. And we'll make sure we put those links in the podcast so they can find you. Thanks.
Kavita Ganesan 23:25
Yeah, thank you.
Nathan Whittacre 23:28
Excellent.
