Ep 33: The Democratization of AI for Small & Medium Businesses

Upside/Downside is a podcast about value creation, and how the actions we take affect profits and cash flow.  I'm your host, Matt Cooley. 

In this episode, Tommy Huynh joins me to explore the democratization of AI and how SMBs in particular can increasingly take advantage.

From sales battlecard creation to low code/no code customer service chatbots, use cases abound if you know the problems you want to solve. 

Get ready to accelerate your value creation curve on this episode of Upside/Downside!

Thank you for listening and please visit Upside/Downside podcast and enter your email for my FREE list: "10 places to look for higher profits and cash flow right now!".

Matt

Transcript

SPEAKER_00: 0:03

Welcome back, everyone. This is Matt Cooley, host of Upside Downside, where we explore value creation and how the actions we take affect profits and cash flow. And we do all of this with a sense of humor. By day, I'm the head of finance for Ericsson's global network platform API business and a self-professed nerd for value creation and how it impacts companies and everyday people. Joining me today is Tommy Hoyne, a former colleague and friend from way back. Tommy has spent 16 years as a growth leader for various technology companies, and for the last four, he's been obsessing over democratizing AI for mainstream audiences. If you know Tommy, then you know the word obsess is a spot-on descriptor. Welcome, my friend.

SPEAKER_01: 0:46

Hey, Matt. Thanks for having me on the

SPEAKER_00: 0:48

show. Thanks for being here. I appreciate it. And I have to say, I love that concept of democratization because the large language models driving ChatGPT, BART, and so on have required billions in investment. We've all read and understood that at this point. In fact, my panel and I covered AI here on Upside Downside right after ChatGPT had gone viral, and the landscape of startups and all kinds of solutions has only exploded since then. So it really is looking like technology everyone can use, and I like that you're into the democratization side of this. So today we're going to take it to a more practical level and talk about small and medium businesses, or SMBs, which is a group I love exploring because There are so many companies and use cases out there. So why don't we start there? What's changed, Tommy, between the release of ChatGPT, BART, and so on, and now?

SPEAKER_01: 1:44

Yeah, that's a great question. And if I can take a step back and just kind of give you perspective on the growth of ChatGPT or just GPT technology overall. So Netflix, when they first came into the market, it took them three and a half years to reach a million users. Instagram took two and a half months to get that to reach their million users. Jack TPT did that in less than five days. So five days. Wow. Five days. So they came out, they released it November 22nd. And five days later, they got 100 million users. And I'm sorry, a million users. And then by, I think, February 23, which was the last time we got some metrics, they were over 100 users, 100 million users. So you can see the magnitude of the AI technology and GPT specifically.

SPEAKER_00: 2:39

And what's changed? Has it improved? Or just in general terms, since all this happened, it seems like it's better. I play with ChatGPT and Bart all the time asking funky questions, and it does seem like it's quickly improving in terms of output.

SPEAKER_01: 2:57

Yeah, I think just from the usability perspective, the response time, the output as well. So AI has been around since like 1950 with Alan Turing. and his research back then. So it's nothing new, but I think now that you're able to ask the AI basic questions on whether you're doing research on AI itself or if you're doing some type of book research, for example, medical, just understanding your symptoms or just doing market asset development. The specific example is it usually takes me three to four days to build out some type of competitive sales battle card for any of the companies I worked with in the past. Now, I just have to put my thoughts into ChatTBT. It can organize that for me. It can put it into the right flow. Obviously, you have to do significant editing, but at least it provides that flow. It adds additional context to what I'm trying to achieve, whether that's identifying potential competitors, understanding what the market is, evaluation and so on and so forth. And I tied myself just to see what the difference would be. So it took me about three and a half hours, maybe four hours to put that together versus before where it would, you know, I have to Google and find research materials, read PDFs and so on and so forth. And I would take, you know, three to four days. And so you can see that time saving

SPEAKER_00: 4:31

difference. Yeah, that's huge. That's huge. Where do SMBs come into play and And, you know, because that's what we wanted to sort of dig into today. And what use cases from your perspective are looking promising for that segment? And, you know, this is a podcast about value creation. So particularly around improving profits and cash flow, you know, what are those use cases that you're seeing and, you know, are on the horizon for the S&P space?

SPEAKER_01: 5:02

Yeah. I mean, the marketing is first. I'm having coming from the marketing and go to market back I think creating assets, especially when you have a limited workforce or you just don't have that resource of hiring a vendor or an agency or just hiring

SPEAKER_00: 5:18

an additional. I'm sorry. Are you complaining that marketing doesn't get the funds that it deserves? As a finance guy, I think this is like a lifelong conversation I've had with marketing colleagues.

SPEAKER_01: 5:29

What do you mean? We never get enough funds. We're asked to bring in 3x, 4x funnel with the pennies that finance has provided. No, I I think we're starting to see a shift in that mindset. And specifically, like the assets I just mentioned, like a sales battle card or sales pitch decks, things like that, where typically you'll have a dedicated marketing manager of some sort to be able to create those assets and have an editor to proofread it and go through that chain. So with ChatGPT or BARD or any other new tools that are coming out today, That can help augment some of the savings from time as well as resource and human perspective. And then other use cases that we're starting to see is customer service. So yes, chatbot 1.0 has been around for, I think, five to seven years. You can go on an e-commerce website, talk to the chatbot, but it's very event-driven, very decision tree-based responses. question, it gives an answer. I ask another question, it gives an answer. It's very prescriptive, right? So you're not going to be able to go outside that box and say, hey, I just want to find out when my order's going to ship. Then, you know, it'll ask, what's the order number? And you give the wrong order number. It's like, no, we're not going to respond alone. Don't give the wrong

SPEAKER_00: 6:56

order number.

SPEAKER_01: 6:58

Or, you know, let's say you misspelled a product. Like, I want to buy this ladder. You know, I know how to describe it, but, you know, then the chatbot will just be cut down. Like, I'm sorry, but I'm not able to understand what you're asking. In those instances, you're going to piss off the user before they are able to become a customer. That's an overall poor experience that AI is helping to solve today.

SPEAKER_00: 7:23

How accessible is that? Are chatbots to SMBs, for example? That strikes me as something of an opportunity now. We have chatbots at my company and I certainly have been you know, following it, at least at a high level, it requires quite a bit of time to implement, you know, at least the 1.0s. So how does this change the playing field for SMBs in terms of chatbots?

SPEAKER_01: 7:55

Yeah, it's still, you know, the bigger companies like the Microsofts and OpenAI, you know, its partnership to allow chat CBT to be, you know, openly accessible, Google's part, Dynthropic, and obviously it's still very much um at that bigger company level. But we're starting to see smaller companies making it available like Chasper or Notion, where it's a little bit more affordable. They have a built-out UI that anyone can just register, create an account, and be able to create content or manage projects and so on. So we're starting to see a lot more of those tools. The last time I think I looked at one of the reports, there are roughly 200 new AI tools released weekly. So that was three, four months ago. We're starting to see a lot more of that. But I think the other thing you have to keep in mind is a lot of them are still hodgepodge. They're just building on top of the OpenAI API. And so there's still a level of development that has to be created for that value, that UI level value that companies are still trying to identify. But if you're just trying to use the tool today just to create content or some type of you know, asking it questions on, you know, like your financial reports, for example. So I think with the new multimodal open AI, you can upload a document. It can recognize, you know, I mean, obviously you want to do it behind your firewall, of course, just for privacy reasons. But, you know, you can start to ask the AI, you know, what's my P&L look like? You know, how much did Tommy spend last month on his wonderful marketing campaign that's going to bring absolutely 10x to the Ericsson.

SPEAKER_00: 9:44

10x. 10x OpEx or 10x net income. Just kidding. Just kidding. Yeah. Let's not get

SPEAKER_01: 9:50

into the weeds. It's 10x. Yeah, let's keep it. It's going on my NVR slide. No, but I think from a high level perspective, those tools are available now. But like I said, you know, like the better tools for smaller teams, you know, like marketing teams that can use like the Jaspers and Oceans that's been around for a couple of years are starting to come into the picture where SMBs can start taking advantage of that to day without having to do some type of setup or creating their own open AI infrastructure behind the cloud.

SPEAKER_00: 10:24

Yeah. All right. Very interesting. So we're definitely still in a hype curve, right? I mean, you alluded to it. 200 new tools released weekly. Everybody's sort of piling on in this space. And Darwin will take care of who comes out of it successfully on the other side. But we are still in this hype curve. I believe, anyway. And you and I know from our careers in technologies that companies may adopt something new. And particularly if you're an early adopter, you may not quite realize the returns that the people selling you the tools promised. How should SMBs and their finance business partners help manage those types of risks? Should they be early adopters? And if they choose to do that, how do you mitigate some of that? some of these risks. So you're not ending up with something that doesn't work, quite frankly.

SPEAKER_01: 11:20

Yeah, I think you made a great point about training a lot of large language models takes billions of dollars, as we've seen with all the companies are raising tons of money today to do that. And just like any new tool or solution, I think having, setting one, setting the right expectations of what's the outcome, what's the pain point that the company's trying to solve for is probably always the first thing that I always allude to is like, what are you trying to solve? What are you trying to fix? Are you trying to increase your marketing team's efficiency? You know, because you're not going to have a full team like the larger organizations will have. So are you trying to do more for less without burning out your team? Right. So that's the first thing to look at is what are the expectations? And then second is probably the ROI. When we adopt this tool, what are we expecting to get out of it? And then can that be scaled across organization where other teams can use it and have that one company perspective or mindset where one tool can be used across the board so you're not working in silos and you're not requiring different departments to have to use different tools just to talk to each other. I think that's another thing. And then I think the third is just rolling it out in phases. Just like with software development, you want to roll it out. The first one, test it, see how it's working well. Is there a team? And I think that's another thing that I'll add into there is the training and support, right? A lot of people or a lot of organizations will adopt a tool, but their users are not trained properly to use those tools, right?

SPEAKER_00: 13:07

Give us your prediction where SMBs will be two to three years out with AI, and particularly, you know, what you think the impact will be on profits and cash flow.

SPEAKER_01: 13:19

Yeah, I think, as I mentioned, we're getting 200 new tools weekly and large, large models are going to be released almost on a biweekly basis. So two or three years will, you know, maybe Skynet will take over and we'll see robots walking down the street. No, I'm just kidding. Maybe not. Maybe not. I was just at a small restaurant downtown where you would never expect those waiter robots, but one just roll on by So it's already happening a lot faster than we're expecting. But I think in two to three years, we're going to start to see a ton of more tools and there are going to be more use cases around it. So medical health care is one of the big focus. Climate change is another. And so that can be scaled down to the smaller companies to be able to utilize all of these new tools. And just the way for how quickly AI tools are coming onto market. I think it, I'm cautious to say, you know, we're going to get all kinds of great tools and whatnot, but I think we will. It's just, what's the level of penetration that we're going to have? You know, the internet, for example, right? When it first came out, I mean, a lot of the executives like Justin Huang from NVIDIA is referencing the AI adoption to when the internet was created back in the 90s. And so we're going to see a lot more use cases. We're going to see a lot more tools and we're going to start to, you know, and that marketizing AI at SMB level, I think we're going to start to see it across the board to all types of business functions, whether it's customer service, most likely to sales, to, you know, back office operations.

SPEAKER_00: 15:09

Okay. Well, I mean, it does sound promising and there's a lot going on. So companies should, SMBs in particular, should be able to extract value out of this. And in my opinion, it doesn't need to be a long-term return either. With this many tools coming out and the growth curve happening so quickly, it feels like these tools should be accretive to the bottom line fairly quickly after implementing them. So that's pretty neat. Tommy, I want to thank you, sir, for sharing your insights and passion about AI today and SMBs in particular.

SPEAKER_01: 15:52

Yeah, it was my pleasure. I just want to add one more thing. Low-code and no-code platforms are going to be the new thing now. So that will significantly shorten the adoption curve as well as we started to see how small businesses can actually take advantage of that to build out their own AI assistant without having to hire an expensive development firm or developers as well if they have that type of technical capabilities. I'll throw that out there. just to get to the universe to, you know, again, to my whole democratizing AI for

SPEAKER_00: 16:27

the everyday business. That's excellent. So that we're not all, you know, having to fight over the same software engineers, you know, ad nauseum, like we have for decades now. Oh, that's excellent. All right. Good point about low code and no code. Okay. To our listeners, I appreciate your support. And I do want to mention, as some of you already know, I do occasionally accept short consulting assignments and all revenues are donated to charities see the link in the episode notes if you're interested in learning more and remember everyone there's a ton of value out there waiting to be created so sharpen your pencils and we will see you soon thank you