Leveraging AI

09 | Supercharge Your Business: Discover All the AI Tools You Need for Immediate Efficiencies 🌟 with Adi Vaxman, CEO of Sheba Consulting

• Isar Meitis, Adi Vaxman • Season 1 • Episode 9

Share episode

Are you ready to harness the power of AI for your business? In today's episode, we dive deep into the world of AI language models, like ChatGPT, and their game-changing impact on businesses across the globe.

Join us as we explore the possibilities AI offers, including increased efficiency, better decision-making, and improved creativity. Our guest, Adi Vaxman,  CEO and founder of Sheba Consulting, shares her experience implementing AI tools to drive efficiencies and success in multiple companies across various industries.

In this episode, we cover tools such as:

🔸 ChatGPT: Generating content and a lot more in seconds
🔸 Numerous.AI - ChatGPT extension for Excel and G-sheets.
🔸 D-ID / Synthesia / Movio - create "talking face" videos in seconds
🔸 Bing/Bard - online AI chatbots
🔸 Fathom.ai - Let AI record, transcribe, summarise, and catalog your meetings in your CRM 
🔸 Engage AI: Automating client outreach and improving engagement
🔸 Magical: Saving time with shortcuts for background prompts
🔸 Koala.sh - a prompt generator for amazing content creation.  (See the detailed process described in the show)
🔸 We also discussed - Limitations of AI: What to watch out for when using these tools

Adi Vaxman is an AI enthusiast, entrepreneur, and founder of Sheba Consulting. Her company provides fractional COO services and tech and AI implementation for businesses. She is leveraging the power of AI to create better experiences for clients while increasing efficiency.  


About Leveraging AI

If you’ve enjoyed or benefited from some of the insights of this episode, leave us a five-star review on your favorite podcast platform, and let us know what you learned, found helpful, or liked most about this show!

Isar Meitis:

Hello and welcome to Leveraging ai, the podcast that. Chairs, practical ethical ways to leverage AI to improve efficiency, grow your business, and advance your career. This is Isar Meitis, your host, and in today's episode we're gonna host Adi Vaxman. Adi is an experienced CEO she's also a tech and AI enthusiast who has implemented multiple AI tools for various use cases in her business and for her clients. And in this episode, we're going to dive to multiple of them. She's gonna share. Probably the most important shopping list you can have right now are the exact tools you want and the exact use cases you can use them for. Right now, most of them are either free or almost free, so you better open your notebook and be ready to take some notes. But before we dive into the episode with Adi, I would like to share some updates of important AI news that happened this past week. This is part of my way to educate you and keep you up to date with what's happening in the worlds of AI, the first piece of news I want to jump into is the news from Google Sundar Pichai announced that DeepMind, one of their research groups has been working on AGI development for a very long time, is joining forces with the Brain team from Google Research to form a new group called Google DeepMind. And the goal of this is to combine all their talents and resources and efforts in acceleration of AI implementation and. Obviously to fight the big success of ChatGPT and open AI together with Microsoft. My take on this is I think they feel for all the right reasons that they maybe have a better technology, which I truly believe they do, but I think they messed up the implementation and they feel the train left the station and now they're playing a very serious catch up game. And I think by combining resources as a huge company as Google, with the amount of data and technology they have, I think they will catch up to open AI in Microsoft. By the way, in parallel to this happening, Google is in the middle of a very interesting period where they already announced in January that they're laying off 12,000 employees worldwide. Some of them are even working on different AI initiatives, so I definitely think that Google is in the process of reorganizing around this new idea of a more centralized strategy to implementation of AI as a way for them to be more competitive and run faster in this race to what kind of AI tools we're gonna use and how many of them are gonna come from Google. A very interesting next step in this race, between Microsoft and Google. Another announcement from Google. That happened, this past week, is that Google has announced that they're going to test out a new model that will allow better AI powered features within their search. And based on internal documents that were leaked in quoted in this article, they're saying that ads will still show up in search results, In this new project that they codename, Maggie will just provide, as they say, better improve search results using AI capabilities. I don't know exactly what this means, but they're saying this will roll out shortly to a relatively small test group, and then they will obviously find the results and decide if they wanna roll it out or not. But they've assigned 160 engineers to this project. Google is making big steps in a lot of investment into finding ways to save the hen that lays their gold eggs, which is search, while still enjoying the benefits of AI and chat bots, et cetera, will be interesting to see where this goes. I. Don't have a clue, obviously, but it's again, another step in this interesting journey towards a new era where AI and AI chatbots play a bigger and bigger role in probably everything that we do, but definitely in how we interact with the internet and information as a whole. My next update from this past week is Ted released a TED Talk with Greg Brockman, Who is one of the co-founders of OpenAI, the company who created ChatGPT. It's a really great TED Talk, so go and watch it. It's 30 minutes that worth every second. He dives into some of the ways and the technology that they're developing and how they're developing it. But the biggest take for me out of all of that is the idea that we've seen before through the concept of agents, AgentGPT and God Mode and Baby AGI and hugging GPT. All these agent-based models that basically are moving into the next stage, if you want, of these large language models in which the large language model is just an interface to an endless and growing set of tools so you can interact with ChatGPT or any other model, and it will know how to complete much more complex tasks by having access to multiple other tools and then coming back with results after it by itself identified what the process that needs to happen in order to get to the outcome has picked the tools. And as came back with a result, some of them are completely autonomous. Some of them, like the example that Greg Brockman is showing in Ted, allows you to pause it in different steps of the way, give it feedback and approve the next step, or pause it before it does something you were not planning. But this is definitely seems to be the direction that this world is going. It's. Like everything else in this world really fascinating and amazing and really scary. Just think about a project, any project you can imagine and think about ChatGPT or a similar model, being the project manager, you're just telling it what the project needs to be and it'll figure out everything else on its own. But this is now not just coming from small developers applying these concept, but from open AI themselves, which means you have really big groups out there that are making this thing move forward, and hence it will move forward and we will see these thing becoming functional and operational in a way that we can use them probably within months. And the last piece of AI related news I wanna talk about this week is a research paper by Nvidia about text to video capabilities. It's incredible. It's really fascinating. I wrote a LinkedIn post about it, so you can go and look it up if you want the links. It will also gonna be in the show notes of this episode, but go and check it out. They're making some big strides into enabling video out of thin air in 4K resolution that looks more and more realistic, and it's not there yet. It's not perfect. You know you have a fox with its tail on the ground, and movements of animals that doesn't look completely realistic, but at the end of the day, the videos are. Mind blowing. And the fact that they were created with no camera crew, no filming, no lighting, no editing, they're just prompts that generated these videos is out of this world. And it has profound implications on every content creation that we know. And it will happen. Like I said on the other stuff, within a few months, we'll start seeing operational capabilities like that we've already seen before and we've heard announcements before as well. But I think again, now that we have Nvidia, maybe one of the three largest forces in the AI world today saying that this is where they're going, it's coming and it's probably coming faster than we are expecting. So expect to see big new things coming in the text to Video world within the next few months. And now to the amazing episode with Adi Vaxman who's gonna share practical, tactical list of tools and the use cases and how you can use them in your business. Hello and welcome to Leveraging ai, the podcast that shares practical, ethical ways to leverage ai, improve efficiency, grow your business, and advance your career. This is Isar Meitis, your host, and I am truly excited today using AI tools in your business has amazing benefits and companies who will not use AI tools will fall behind and may run out of business. But knowing which tools to use and which use cases to apply them to is a little scary. And they're literally new tools and sometimes hundreds of tools coming out every single week. Hence, what you need in order to get going is somebody who is an amazing practitioner who literally enjoys geeking out on testing these different tools and seeing how they work and which use cases to apply them to. And this is exactly our guest today. So our guest today, Adiva Man is an amazing practitioner of AI tools. She's also the CEO and the founder of Sheba Consulting. And in the last 14 years, she has been helping companies reduce cost, increase productivity, and enable scale, which in the past few months, Has involved more and more AI tools implementation. And so in this episode, it's gonna be, I promise you, after talking to Adi with her level of energy, it's gonna be drinking from a fire hose. So get ready for a serious ride. Get ready to take some notes and we'll dive into multiple tools and the use cases you can use them in your business today for either free or very little money to get much higher efficiencies of different aspects of the business. Adi, I'm really happy to have you here. I know we're gonna have a fun conversation. Welcome to leveraging.

Adi Vaxman:

Thank you, Sarah, for having me. this topic is really exciting. As you said, I have always been a geek and now I'm gonna expose on the geeky side of me to the world. But yeah, really geeking out over these tools in the last six or seven months more than before. And, cool. Let's do it.

Isar Meitis:

Awesome. First question. So you. We all use some of these tools. Grammarly is an AI tool, and Descript is an AI tool. there was AI tools before ChatGPT dropped on us on a again, shiny day here in Florida, in late November of last year. But that really changed everything. How did you guys start looking at these tools? Because I think most CEOs, that's the biggest problem. okay, how do I even get started in picking the right aspects of the business or in choosing the right tools that can really benefit

Adi Vaxman:

my business? So I think it's a really simple answer. The first piece of that answer is that me, my partner, and everybody on my team are all technology geeks, right? We love using technologies. We love testing different systems. We spend hours and hours on a weekly basis, and this is part of our entire teams like goals. Testing new tools, looking at monitoring what's going on and testing new tools. So this has always been the case. It's not new. How do we get involved with AI?Before, I would say before November, we've been using some AI tools, like everybody's been using your grammars and some Jasper and some, I'm a photographer and do a lot of image manipulations. Some of like Topaz Labs tools that utilize ai. And I'm sure a lot of you have been using AI tools without even knowing, with filters on your phones and stuff like that. If you have teenagers at home, which I do, especially female teenagers at home, there's a lot of AI tools that you are using without even knowing. It started like that and then, in November and early November, rumors started. Bubbling about ChatGPT about open ai and I have friends at Open ai. I've been following the company for a long time, obviously since, even before Elon Mask separated with it. So it wasn't huge news, but there were talk about this tool getting released to the public. So we've been following, we've been following the news and the day that the tool got released to the public, we got on the waiting list immediately and I think got a, two or three days later, it was pretty quick. That's before everybody started jumping on that moving train. And I was simply, Flabbergast, like it was just like, it was just unbelievable. I've used Jasper before for a long time, and it just wasn't there. It didn't really save me much time. I used it for very specific things. It wasn't as great, used Grammarly, but again, for very specific things. And then comes ChatGPT. And I'm a writer, always been a writer. And I was just amazed and I was shocked. And that week's our, that week's team meeting, our team meeting was spent entirely, and instead of an hour, it was two. Was spent entirely on ChatGPT demo and showing it to the team and trying to figure out how to interact with it and testing it out. The next thing that happened, I don't remember when it was, but shortly after, was the Microsoft announcement about Bing and we got on that, beta as well pretty quickly and started trying to hack slash destroy slash joke around with Bing or Sydney or whatever you wanna call it at the time, and managed to get some pretty crazy things. And, I'm a part of a community. In general, some of my clients are like that as well. And so we started all talking about, oh, did you see this? Started following the news, following Twitter, following, Reddit forms and whatnot about people doing different things with this tool. then Ben came out to the public and. First, we met, we all managed to break it really quickly. Then they limited it to, I think it was the first thing was six, or five or six interactions. Then it went to eight, then it went to 15. so all throughout, I would say, December through February, I feel like 90% of what we've done, we've put aside a lot of the testings of other systems that we usually test and we focused on. Just trying to figure out AI tools in general. So I wanna

Isar Meitis:

pause you for a second to do a quick summary and generalization of what you said. The first thing you said is getting in early, so it's still early for those who are like, oh my God, everybody's in and I'm not, no, it's still very early in that game. So it's not too late to get in. Now the second thing is, Let the geeks in your company play with it. So you said, okay, you are just like me. We're both geeks. We like these kind of things. We, you spend hours through the night, probably like me, just, oh my God, there's a new toy. I gotta check it out. Yes. And but maybe you're not like that, but there's people in your organization who are. Give them the freedom. And more than that, you said it became a part of company procedure, like this is what we gonna talk about in the next meeting. So you allow people the time to go and investigate and play with tools and come back with their recommendations and ideas, and then do some brainstorming. So now you involved a few more people. I assume you have some KPIs around that as well, on what tools you need to check this week versus next week and what the output needs to be for specific meetings. And then the last thing you said is, be involved in some kind of a community. You can listen to podcasts like this one. You can join different communities online. There's multiple of them, or just hang out with the right people. But this will give you a broader flow of information of what's going on and keeping updated. And this thing is moving faster than anything else before. So this is a summary so far. What about the other tools? Let's just dive in. I think that's gonna be okay. I see you wanna say something, then we'll dive right in. Before

Adi Vaxman:

we dive right in, I wanna say two things. 1) In our world, this train is moving so fast and I feel like it's been a lifetime. In the outside world, that's not the case. Correct? I talk to people every single day that have absolutely no idea, no clue, and never heard of it. And I am so shocked because it's been my life for a good, almost six months now. Like it's been a long time and it's been the focus of almost everything that we do. So that's one thing. The second thing I wanna say is that it wasn't necessarily an easy process even in my team. I have writers on my team. it wasn't an easy thing for them to accept. It wasn't an easy thing for them to adapt. and I, and there are a lot of scary ramifications as well to what's happening in the world with ai of course. But we are at a place now where even the people that were really deadly against it, Are using those tools on a daily basis and understanding two things. One, if we're not on that train, we're gonna get left behind. And two, use it to your advantage. Find the good things in it because there are plenty of those. And the third last thing before we dive in, talk about the geeks on our team. Our CTO has been going to sleep at six in the morning and waking up at eight because he's been spending full nights testing tools like Mid Journey, Dall-e, and whatever a different APIs that he's been implementing for clients. So talk about geeking out. Yes.

Isar Meitis:

Awesome. So let's really start going tools one by one in whatever order you want. Either stuff that you've implemented for clients or stuff you're using internally. And really, let's start with the things that you find most productive, the stuff that you think that saves you the most amount of time. The use cases that you're like, oh my God, this was a game changer and took phase hours whatever you wanna start with.

Adi Vaxman:

Okay, perfect. So I think the very obvious one is ChatGPT, that's the one that everybody's talking about, but ChatGPT has its advantages and has its limitations and a lot of other tools are built on Top of it. So we definitely use ChatGPT in my company and in our client companies that we've implemented, complete workflows and use cases to use a combination of ChatGPT, Bing and Bard. For the reason that ChatGPT has ChatGPT-4, which is the most advanced, currently available language model out there. Combined with Bing and Bard's ability to access the internet and be trained on data that is post September, 2021 which is where ChatGPT stops. And so the combination of these tools allows us to do different things. We use some tools for research, some tools for comparing other tools. so for example, and I'll give you a great example, of something that I'm using. I'll start with the first, productivity tool that I'm using for myself and for my team, and have recommended to clients. And you even told me that you tried it today. it's called Fathom ai. It's a free zoom plugin that is an AI note taker and summarizer. So it does a lot of things. It transcribes Zoom conversations. And it, it does different things to me, the biggest advantage of it is the summary. So I used to take a lot of notes during conversations and now I don't do that anymore because I get an AI summary created for me with the right context of the conversation really well by Fathom. And I've tried different tools before. I've had Otter before, I've had other tool tools before that just didn't. Just were not up to part to my expectations. And this one is there are a few other tools that we're testing in the same kind of reality. And one is called, I think TLDR or something like that. I'll check the name. TLDR. Yeah. and the other is called BlueSnap. And there's one called Jamie. So they're a combination of tools that are doing these things. So that's a tool. It's which is surprising.

Isar Meitis:

I wanna say something else about Fathom. I used these kind of tools as well before. The two really incredible things about Fathom three. One is that it's connected to the the crm, it's also connected to the video, so any point in time that it highlights, you can click the link and it shows you that segment within the actual conversation. If you want the full context of something and in real time while the conversation is running, you can make your own bookmarks on things you wanna remember or things you wanna highlight for yourself as they're happening in the call, and they will show up as bookmarks after. Really incredible tool. So I strongly agree that this is an awesome recommendation. like I said, it also connects with the CRM, so you have the notes in the right place in the c r m doing absolutely nothing.

Adi Vaxman:

Yeah, it's amazing. Everybody looking at the, at our CRM is able to click the action item from the summary of the call, go straight to the video, see the right area that this thing is being discussed, and move on with their lives, do what they need to do, really productive. But my point was, is that when I was researching better things than Otter, which I've tried in the past and just wasn't great for me. I asked Bing to give me what's out there in the market, what's out there that does these things, and then give me a comparison and then tell me the install base and then tell me the pricing and create a table that shows me all the features and whether the feature that I was specifically interested in, which is the summary feature rather than the transcription because who has the time to read 50 pages of transcription. And so Bing did most of this work for me, and at the end of the day, that kind of reduced the number of tools that I have to test to three or four, and we just started testing. so that's a really good use case when I use a combination. For ChatGPT I use two plugins that are very, they're Chrome extensions that are very useful. One is called ShareGPT. So ShareGPT is a chrome extension that lets you share an entire conversation, creates a URL with the entire conversation of the prompts and the responses and everything. You don't have to copy and paste. Very useful, especially when you're having ChatGPT write code for you. So I am not a developer. I studied computer science 32 years ago, that was a very long time ago. and I'm not a developer. I know a little bit about structure of code, but that's pretty much it. And every time I had to create a script or a code, it would take me hours of research on Google and on GitHub and on Reddit or whatever, trying to figure out how to do what I need to do. And, ChatGPT helps me write the code that I need to write, that I need to write. And so whenever I need to share that code with someone to troubleshoot, to implement to whatever, I just use, ShareGPT. And it's all there, no copying and pasting, with the format, with the. great little tool ShareGPT. Second one for chat

Isar Meitis:

G pt. By the way, another cool thing about ShareGPT is that you can publicly share stuff and there's lots and lots and lots of publicly shared ones, and you can learn how other people using ChatGPT, what kind of conversations they have with prompts they use, how they progress through the conversation. So also as a learning tool, it's an amazing tool.

Adi Vaxman:

So it's no big surprise, right? That ChatGPT is only as good as the prompts that you use with it. So really, prompting is really the art form around ChatGPT. And so there have been a lot of tools that popped up with pre-engineered prompts, that are being sold or offered to the public. Some of them are really. Great at it, and I'll talk about one of them, shortly. But the second Chrome extension that I wanted to share is called Summarize. It's also a ChatGPT Chrome extension. You need to have a ChatGPT account, so it's using your ChatGPT account. And so whatever webpage that you're on, you can click summarize and it'll give you a summary of that webpage. It could, it works on documents as well if the document accesses, has the right access. And so you can get really great AI summaries of articles, documents,, that research paper that you shared not too long ago and I had no time to read the whole thing I used summarize, and it did a really good job in telling me what I actually needed to read. so that's another cool tool. Awesome. Let's move on to some other tools. One tool that we use and have implemented for clients is called Numerous ai. Ooh, I love that one. Yeah, let's talk about that. It's a Google sheet, spreadsheet AI automation tool that allows you to use ChatGPT on a Google spreadsheet. And so I can tell you the type of things I do with it. First of all, people on my team that don't have incredible ex Excel or Google Sheet skills that need to create formulas or analyze a lot of data can use it very easily just by creating prompts, analyze the data in row B and put the combination of the data from ROW BC and e into row L, whatever. So that's one thing that we use it for. Another thing that we use it for is to create sort of mail merge type things. So we have a list of, people that we wanna send holiday greetings to, depending on what holidays they celebrate. So we have a list in Excel, some Celebrate Passover, talk about clients or talk about whoever, some celebrate Easter, some celebrate whatever. and we have ChatGPT, right? Very personalized messages to these people so that we can mail, merge and send out the emails. same

Isar Meitis:

for birthdays. It's brilliant. I never would've thought of that in this case, but I think it's really brilliant. Yeah, I wanna explain to people what it does. It does two things extremely well. One is it literally knows all the formulas of Excel or Sheets, and it has it for both. So you can either write it within, Microsoft Excel or in Google Sheets, so it knows all the formulas. So instead of knowing all the formulas, you can basically tell it what you want the outcome to be. I want the minimal value of this. I want you to find the first name of all the people in those strings. I want, like whatever the case may be, it knows h what formula to apply. So you don't need to know any formulas, you just tell it what you want the outcome to be and it knows how to do that. And it's nothing short of magical. But the other thing it's. It's the AI engine behind it is GPT. So if you wanted to write stuff, you can just say, Hey, you want, I want in this cell, you need to write whatever it is. You wanna write like a regular prompt, and you will get the outcome in that cell. So now you can use it, like you said, for multiple use cases, as a reference for other stuff in other cells. Brilliant.

Adi Vaxman:

So this is an example. We recently launched a referral program for our company and we had a list of people, past clients, current clients, other people that we network with, attorneys that we work with, different kind of contacts, and we wanted to send them all emails explaining what the referral program is. Giving them some background and asking them to join. And so each of them comes from a very different angle, right? A client already knows what we do. an attorney that we work with usually knows what we do, but some of these other people don't necessarily know what we do. They just know that we use their tools. And so depending on, we create a category, it's a client, it's a networking, it's a vendor, it's a whatever. And according to that, we had numerous AI create messaging for these different people. Based on how much they know about

Isar Meitis:

our company. So again, I want to take this just to, I'm just going crazy now with the idea you gave me, when we look at the marketing funnel, we talk about the know, like, trust. We talk about people getting them to know you then to like you, then to trust you. You can use it for so many different things, what you just said, because you can put people in categories, even pull existing categories from the crm. Then build messaging with that tool using something like, like this within Google Sheets for practically free. It's amazing. Yes.

Adi Vaxman:

It uses tokens like, because ChatGPT API uses tokens. It's not free, but it's not expensive. It's almost free. And yeah. And so the prompt for ChatGPT includes, a client already knows what we do. A vendor doesn't necessarily know what we do, but we have relationships with these people, so you can't just send the same message to everybody. We have different relationships with every kind of type of person there. You want the message to be very personal. So that's an example. Another example of something that we've implemented for a client, and I'm gonna go over the backend tools a little bit later, but I'm kind of focused on sort of the tools that every user. Regular Joe can use and doesn't need to know any backend to have any backend skills. so there is a tool called Synthesia, another one called MoVI, another one called D-ID that do pretty much the same thing. And we use these tools to create things like explainer videos. I'll explain what they are. Internal training videos, how to feature demonstrations, create social media engagement. One of the things that we've done recently with a client is that we used, Synthesia to create an AI generated video that does employee onboarding, so the employee onboarding new employees. This is a remote company. They hire employees remotely. New employees receive this video on their first day. The video has a checklist of all the different things that they have to complete. They have to complete their I nine, they have to do the registration for this, that with screen shares showing all the things that they need to do and how to do them with a timeline. So they can jump to okay, you're finished. This, you can click here to go to the next section. how do you set up your laptop? What are the tools that you need to set up with links to the things that you need to set up? And there is a person, a nice lady. Walking them through this. What we used to do before is we used to do an onboarding call with an HR person, that walks them through all of this. Now we give them this video. This video took us, probably about four or five hours to produce. It wasn't terrible. And this company's hiring a lot, so every time a new employee comes on board, they only meet the HR person after they're done walking through everything on that video. And then it really reduces the amount of time, that we need to answer silly questions about, oh, where do I click?

Isar Meitis:

I want to connect a few of the dots that we talked about because I think for people who are a little more advanced or just to open people's minds to what's currently possible. So if we're talking about the last few tools that we talked about between numerous.ai and a tool like Synthesia or D I D, that can create videos of people who look pretty real. I mean, you can see it's not a person, but it's very personalized. It still looks like somebody talking to you versus just reading a document. Combine the two together and you can get a more personalized video. So depending on the department they're coming from, depending on their experience, depending, did they come from a company in the same industry? Depending, whatever their background is, you can give them a more personalized video that may not need to explain everything to them that will save them the more time that you can create using something like numerous.ai in a Google Sheets, that will then be the text that d i d or Synthesia, virtual person can read to that person. So what I'm telling you right now is available. Now, this is not like futuristic something, this is available right now and these tools are really remarkable in a, like I said, personalization of stuff because they give you the flexibility. You don't have to rerecord the video cuz the video is literally generated based on text and the other is just the speed that you can produce things and how many of these you can produce without going to recording and editing and oh my God, they said something wrong, or the lighting was bad, or the microphone wasn't working well, all of that goes away. So you can get really solid videos with almost no effort that are still personalized way more than you can do in any other video. So amazing example.

Adi Vaxman:

Yeah, so that's definitely one of the tools that we've been implementing with clients and specifically in the HR world, where we do a lot of internal training and, just training in general. that's been really mind blowing to see. And there's a bit of a learning curve. It's not completely seamless yet. But I think it's gonna get there really in a matter of weeks or month. the next tools, or a bunch of tools I wanna talk about are image generation and manipulation in general. So I'm sure everybody heard about, Dall-e, mid journey like engines that can create image images on the fly. But I wanna give you some ideas about what we're doing with these things. first of all, if you didn't know, Shutterstock, for example, have an implementation of Dall-e within Shutterstock that allows you to, as part of your stock images plan with them to create completely new images. These tools are still a little bit restricted in what they can do. So they can't handle text on top of images yet, but I'm sure that's gonna, it's funny, just text of all things they should be handling pretty easily.

Isar Meitis:

So midjourney literally just announced like the next new version, next few versions of what they're gonna do. And one of the things they're gonna solve for is text. So this one is huge by the way. Another one that is related to our previous conversation from ShareGPT is you'll be able to reverse the process, you'll be able to upload an image and it will tell you the prompt you will need to write in to create the image you just uploaded so you can create similar images using similar prompts. So a lot of cool things are coming to Midjourney, including thank God text.

Adi Vaxman:

So yes, midJourney, Dall-e, Bing that he is using Dall-e. So for those of you who just wanna play around, Bing does a pretty decent job of creating images using Dall-e. It's completely free. So if you have the Bing app on your phone or Microsoft Edge on your computer with the Bing Chat, you can create images by just any prompt. Create for me an image of a green butterfly, with elephant ears and I don't know, really long Antler. Those are things that we use on a regular basis. What do we use them for? We use them to create imagery for training, for presentations, for blog posts, for clients, and for ourselves as well. Things like that. So that's been very useful and they're still not amazing at creating logos and More business-like designs, but I suspect it's gonna get there very quickly. Mid Journey five does some pretty incredible things. It even goes all the way. And as a avid photographer, I can say, you can tell what camera to use, what lens to use, what aperture, to use, what's like all of these things, shutter speed, all of these things and create a real, real looking image. And it's pretty amazing. So we use a lot of these to, instead of stock photography now, we used to use a lot of stock photography for clients for all different reasons, and a lot of it is being generated by AI now other AI tools such as Gigapixel, which is part of the Topaz Labs, suite that helps manipulate. So there's a sharpen, and upscale like tools that allow you to take images that you have and fix issues with them. And also old images, really old pictures, that are very blurry, that are badly scanned, can be made like they were taken, this week. So that's, those are pretty amazing things in general. I think the next thing that I want to talk about is one of our client's, use cases that I find pretty shocking. Can't take credit for it because it was completely created by the client. We've just been involved from the operational standpoint because we provide fractional COO services to the client. So our client is a major web publisher that monetizes content websites. We've been with them for about two and a half years now, I wanna say and they've employed a lot of freelance writers all around the globe to write content and had some very elaborate, very specific guidelines and workflows and automations, and they've done remarkable work with automating a lot of their workflows to work with these writers to produce valid and accurate and good content for their many, many, many, many content websites. And given the AI kind of revolution in the last few months, client decided to change the roles of the writers and no longer use writers as writers, but they use them as what they call AI Wranglers. So they're really working with AI tools to create the content, and they developed an entire workflow, a pretty impressive workflow with training, that our clients shared on her blog., it's yeys.com. I think Y e y s.com is the blog if anybody's interested in how she's doing it. So she shared it on the blog using an AI tool that's based on ChatGPT called Koala Writer. So it's, the website is koala.sh k o a l a.sh. And it's basically a set of pre-engineered GPT prompts, GPT4 prompts that are engineered for specific types of content. And it could be a blog post, it could be a long form article, it could be a lot of different things. And out of all the tools that we created, that we tested out and we've used, including ChatGPT. It actually does a really good job in creating the outline, helping you write the content. So it's pretty amazing what they've developed there and they just developed and an entire long form content creation workflow across all of their content websites using this Kuala tool. It's used successfully across many of their websites, one of which is that's, I think that was their first one that made'em so successful. It's called the Cat Site, the catsite.com. People love cats and I'm a cat lover and this is how we met actually. But, but definitely check it out. She has under yays.com/koala writer dash review is where she does the review of all of these.

Isar Meitis:

I wanna touch on this because this is a big topic, that a lot of people are struggling with or see as a huge potential, but not knowing exactly how to take the potential and implement it in their company without just creating a lot of vanilla content and the key to all of this is, you touched on two points. One is a process, like it's not just, oh, I'm gonna give it a prompt and I'm gonna use it. No, the content has created, because then what you're creating is a lot of vanilla content. So if you want to create unique, Thought leadership kind of content. What you wanna do is you first of all wanna do research on the topics you wanna share, which you can use other tools to do. By the way she's using, I read her post, it's amazing. She's using Koala Chat, which is another tool within the thing that has access to the internet to do the research on the topics she wants to touch on. But there's other tools that helps you to do SEO kind of research. Using AI specific tools. Yes. Then the other thing that you touched upon is prompting is everything when it comes to AI tools and what Koala writer does very well is it guides you through the different components of the prompt so you don't have to have it. You can write a half a page prompt on your own, but when somebody helps you through, okay, what do you want the topic to be about? What do you want the. Voice to be like, what do you want? The in sentiment, the sentiment to be? What, content, resources you want to use as a baseline for this. What ideas you want to incorporate, what you don't want to talk about. So if somebody guides you through the different components of the prompt, there's a very good chance you'll get a very good prompt and hence a very good outcome. And then he goes back to, The human part of this where somebody eventually has to read this, make sure it makes sense, do some fact checking, and do some final editing before you can actually post this. So yes, it saves a hell of a lot of time and it saves a few steps. But if you wanna save everything, you will end up with lousy content. Yes, if you're willing to do some of the heavy lifting, then you can end up with amazing content with relatively little.

Adi Vaxman:

So to me this is really important. This is a very important point because in my team, in the very beginning when we started using these tools, there was a whole heated debate about, what do you mean it's gonna replace writers? I'm telling you already, it does not replace writers. Yeah. It cannot replace writers. Not at this point anyway. Who knows what the future is gonna bring. But when I'm looking at what my client is doing, it's content development. It's not just put in a topic and it creates an article. It doesn't work that way. So if you think about the way that you used to write before these tools existed, and I'm a writer and I used to write, I write for Forbes and I write for Entrepreneur Magazine. And before I used these tools, I used to create an outline. For what I'm writing, I used to put in the outline all of the ideas that I want to cover, and then I used to look at that outline and see, does it make sense? Does it flow for a person that's not in my head? Does it really flow? Where's the beginning? What's the middle? What's the end? Create a big, create an opening paragraph. Create a closing or summary paragraph. Create a thesis paragraph. and whatever is in the middle, right? So you had to go through all these steps to create your content to begin with. What this helps you with is, one, it guides you through the process to create the steps. but secondly, it also helps you develop your points and maintain a uniform voice. Throughout whatever it is that you're writing, so it doesn't look like you copied and pasted a whole different bunch of things. But me personally, when I use this to help me write, I use it to create separate components and pieces of what I'm writing completely separately from one another. I work on each paragraph separately. I work on my outline before titles, things like that. It's really helpful to me. In creating titles like I am not very good in thinking up, like really imaginative, creative, titles that will ca catch people's attention. It's not who I am. I, I write very factual content and so ChatpGPT or even Koala are doing a great job in suggesting titles based on the content that I've written. So things like that.

Isar Meitis:

Great. I think all very solid recommendations.

Adi Vaxman:

Move on to another tool. Another tool that we've been recently testing really recently, just a few days, is called Engage ai. So it's an AI tool to help social media engagement, help you reply and create comments and to chats and conversations. And, I think it's pretty cute. Really helpful again for someone who is not very, I'm a pretty to the fact, matter of fact person, and I'm not good at the fluff. I'm just really not. And unfortunately, the fluff is part of our day-to-day interactions with people that don't know us. And without that fluff, we may seem rude. And I'm really not trying to be rude. It's just I'm not a fluff personality. And so helping me. A little bit of fluff. That's not too much. That's not gonna make me, sick because I really don't like it. Is really helpful to me. And when I use a lot of these tools and I actually say to chat, G p t, this makes me sick. It's, it's too, it's too whatever. It's, too salesy, it's too pushy, it's too whatever. make it more me, make it more. Matter of fact, make it less, whatever it is. So these are things that I use a lot and I use it this way with Engage AI as well.

Isar Meitis:

By the way, I recorded another episode with somebody and it will come out somewhere, either before or after this one, I'm not sure on how to help GPT learn your own voice and write like you. So if you're listening to this podcast, look for the other one, then you'll be able to find a detail process on how to do that. That'll

Adi Vaxman:

be great. And of course, the one thing that we're all waiting for is the ability to train a GPT model. On our own voice, our own personality, our own environment, our own data set. Our CTO has been working days and nights trying to understand and get to know all the tools that are out there so that the moment that this will be available to the public, we will be able to do it. And then it will really be a game changer for business clients complete game changer. Because currently what we do is that we have a lot of, and you told me about a tool that you should probably mentioned here, but we created a library of preexisting background prompts so that whenever we start a new conversation about a client, about our own company, about whatever the topic is, we have a whole list of background prompts that we can already copy and paste into whatever tool that we're using in order to develop the conversation. And you taught me about a tool that can help me automate that, so that's awesome.

Isar Meitis:

Yeah, so the tool that I'm using, I've used it way before ai, it's a Chrome extension called Magical that literally allows you to create, shortcuts that you can save whole paragraphs or a few paragraphs and give them whatever shortcut, 1, 2, 3, 4 characters that you can remember for stuff that you use all the time. And now it's also very useful in prompting because if there's a segment of the prompt, you say again and again, just save it to magical. Use whatever shortcut you want, you type it in, it pops up that whole segment into your prompt and you just saved yourself going to the document, opening the drive, finding the thing, opening it, looking for the right segment, copying, pasting, all of that goes away. Three letters. It pops up on your browser and it's really helpful when you're reusing stuff anywhere, but also

Adi Vaxman:

imprompt. Very cool. Very cool. I installed it and I can't wait to start teaching it. All my background prompts. Let's talk about a few client use cases that we use on the, on our CTO services side, we have a client who's a media marketplace for wedding vendors, specifically photographers and videographers. And they have a tech platform to help people match the people who are getting married with the different wedding vendors and be able to see all the stuff that they've done. So on that end, they created on the backend they created a process with ai, with ChatGPT API to automatically generate the con, the user facing content for the detailed descriptions of the wedding videos and photographs that are being showcased on the site. On tens of thousands of pages on the platform, so to, to help the search experience. So right now, whereas before a person was writing these descriptions of the videos and of the images and trying to get all the keywords in so that someone can find what they really are looking for. Now it's done automatically and being generated by ChatGPT API. They also implemented a video analysis tool for grabbing samples of the video frames, and passing them through AI tools to extract information like objects, things that are in the videos or in the images, the properties that are in the videos or in the images, specific brands, specific bands. Specific moment. So if a couple that's getting married is looking for a wedding photographer with a specific style that likes to photograph specific objects. Now the tools, the AI tools are able to analyze the data in those videos and create those keywords for the search. Amazing. We also created a for, we have a client that's a PR automation platform. And, they have a platform for monitoring news and pitching. Pitching writers and pitching thought leadership to different, news outlets and media. So we created our c t o, created an elaborate web and print news repository that they're using AI for in order to classify the relevance, the importance, and the relative ranking and general sentiment of the different news article to, to inform the search experience on the platform. So that's another use of backend implementation of AI.

Isar Meitis:

And I want to tell something to people who are terrified of everything that you're saying from, oh my God, how can we implement this perspective if we don't have a CTO kind of person? Some of these can be done. So the API is an api. You need an API key, which is not a technical thing. You literally go and you sign up and you get a key. It's a string of letters and digits, and you can use that API Keys in different tools that are not code related. So the Zapier of the world, you can just say, oh, I want to connect through this. And now you can use the API to query existing databases that you have by exposing them through a third party tool that does not require development. So if this still scales you, then maybe it's not for you yet. But all I'm saying, You don't necessarily need to have a computer programmer in order to enjoy the ChatGPT API and similar other APIs within your workflows because there's many other third party tools that allows you to connect the two together without writing any code.

Adi Vaxman:

And you can always ask ChatGPT how to do it because Chad, you'll tell

Isar Meitis:

you that's true. That's true as well.

Adi Vaxman:

You can ask for foolproof instructions for someone who doesn't know how to code and it'll give you those instructions. Including the code. Including the code, including where to implement the code, including, what to open an account on in order to implement the code. Yes. So yes. so these are some of the use cases that we've used. There's another use case that's the most obvious use case that I think all of our clients are currently using, which is to use ChatGPT and or Koala. and similar tools and in my life, a combination of Bing, Bard and ChatGPT to create marketing content. Any kind of marketing content, including company collateral. I've used ChatGPT to help me create legal agreements to help me revise policies to research changes in law. We deal with a lot of HR law, so research changes in law to give me specific quotes from specific pieces of the law so that I can use them in, whether it's negotiations or, back and forth with clients or we client attorneys. So a lot of research, really simplifies the time and, simplifies the process of doing research and save me a ton of time on looking for the right specific marker in a sp in a huge document that I need to make whatever case I'm trying to make.

Isar Meitis:

Wow, okay. Adi is incredible. I think we covered, like I said, like drinking from a fire hose. We covered many, many, many different tools and many different use cases that both you using internally and your clients are using. I think the biggest takes from all of this is that there's literally tools to do. AI tools out there to do almost anything you can imagine today. And it's all about researching that. Like as I said, you can use AI tools to help you in the research, including ending up with a detail table on the benefits and the differences and the pricing and the links and everything. Just tell it what you want in the table and it'll create it for you. And the most important Is to align it with the procedures and the skills that you have within your company that it's actually become useful because otherwise it's okay. So I got the tool, but it's not really being utilized in any way that helps our process. So I think starting with understanding. Where these tools can help existing processes and which skills in the company are aligned with using these tools. That will be probably my step number one. But a lot of the tools that ADI just covered are literally from everyone because they're just extensions on Chrome that can do these magical things out of the blue without you having to learn, new stuff. Adi, if people wanna connect with you, follow you, work with you, work with your agency, what's the best way to connect?

Adi Vaxman:

So we are on LinkedIn. of course our website is Shebaconsulting.com. S H e B a consulting.com. I'm on LinkedIn. The company's on LinkedIn World Solve Facebook and on Twitter. all of our contact information, we even have AI chat on our website, implemented. And one last thing that I just wanna mention before we move on. Everyone should also be aware of the limitations, right? So these tools are incredible and there are so many benefits that I can't even count. that's why we're using them on a daily basis. But they do have limitations. and a lot of people are not aware of the limitations, Do your research, be aware of the limitations. check your sources. Even if ChatGPT or Bing or Bard spits out a list of sources. Click the links, make sure they are a accurate, make sure that the quotes are accurate. Because sometimes these tools, the highly intelligent, artificial intelligence that they are, they make things up. And when they make things up, it sounds very credible. It sound very true, but sometimes it's not. So be aware of the limitations. Just, wanted to say this so that we don't have a whole lot of people, out there and taking it for a fact. And I also think that kind of last words is that institutions that are still rejecting the use of these tools and they are out there, schools, universities, the media. They're gonna fall behind, they're gonna have to adapt these tools because we can't stop progress. And as much as some people are trying, we can't and we won't stop it. And so I think the best way is to use it to our advantage. I

Isar Meitis:

agree. I think there are huge risks, but I think the only way to minimize the risk is to have as deep as an understanding as possible and what it's actually doing and capable and the way it's moving, so you can better prepare yourself, your family, your company, for whatever it's coming. Adi, this was fantastic. Thank you so much.

Adi Vaxman:

Thank you very much for having me.

People on this episode

Head Shot

Isar Meitis

Host