Loading summary
Transcript101 lines
- [00:01]
A
I've got some very exciting news. We've been secretly planning four months, and it's finally here. Introducing AI Business World 2026. It's Social Media Examiner's first ever dedicated AI conference. Get trained by top AI marketing practitioners on April 29th and 30th at the Anaheim Convention Center. Learn which tools actually work for your business, master workflows that save you hours daily, and connect with thousands of marketers serious about AI. And here's the best part. It's included at no added cost with your Social Media Marketing World 2026 ticket. Or get standalone early bird tickets for just $597. But that pricing ends really soon. Head to AI BusinessWorld Live right now to get your tickets. Finally, practical AI marketing training that makes you confident and irreplaceable.
- [01:02]
B
Welcome to the AI Explored podcast, helping you put AI to work. And now, here's your host, Michael Stelzner.
- [01:14]
A
Hello, hello, hello. Thank you so much for joining me for the AI Explored podcast brought to you by Social Media Examiner. I'm your host, Michael Stelzner. This is the podcast for marketers, creators and business owners who want to know how to put AI to work. ChatGPT has hidden features that can do things that might surprise you. People say ChatGPT is really only good for written content, but is that true? In today's episode of the AI Explored podcast, we'll explore advanced ChatGPT features that any business can use. Today, my special Guest is an AI strategist who helps B2B small businesses apply AI to all aspects of their businesses. She's the co founder of Cincy AI for Humans. Her newsletter is AI Simplified on substack. Kendra Ramirez, welcome to the show. How you doing today?
- [02:04]
B
Fantastic, Michael. Thank you for having me.
- [02:07]
A
Super awesome to have you. So let's start with how in the world did you get into AI?
- [02:11]
B
Good question. I'm a technology nerd. I'm constantly dabbling in something and in 2018, having had a digital agency for the last 20 years, in 2018, just out of need from our clients, we were doing social media management and they're like, hey, we can't keep up with the direct messages and inbox. What do we do? And so I found some tools back then to build bots. And I always say the good bots, not the annoying ones. And from there I just continued to learn and evolve. We were using copy AI and Jasper and a couple other tools prior to ChatGPT. And so once ChatGPT hit, it was like a rocket ship all of a Sudden, it just shifted everything in our business and for our clients.
- [02:53]
A
Tell us a little bit about what you're doing now.
- [02:55]
B
Yes. So what I'm doing now is a lot of travel, a lot of facilitation, and being able to understand what the heck AI is, where to start, identifying use cases, prioritizing them, doing workflows, and then flywheel again. Right. Continue taking that through all the different departments.
- [03:15]
A
Yeah. And you're also working with local businesses in Cincinnati, right? Like Helen Todd, I believe, who's been on this show. You and her do some work together, is that correct?
- [03:23]
B
Yes, yes. Helen is co host as well, in Scentsy AI.
- [03:28]
A
Excellent. Okay, so let's start with what are some of the biggest misconceptions that you see amongst your clients and amongst the people that you're teaching when it comes to ChatGPT?
- [03:38]
B
Something I hear quite often, and those that are new to ChatGPT is they think it's just simply a writing tool. So when they're going in and trying to prompt or play with it, they're like, write an email for a client. And it does so much more than that. And getting them to understand the difference in using Google versus AI. Right. You can use Google, you can ask questions, you can ask questions of AI, but the magic in AI is actually giving it a task, getting it actually to do work for you, which helps all of us.
- [04:07]
A
Excellent. Any other misconceptions worth talking about?
- [04:10]
B
Yes, I think the biggest other misconception I hear often is, gosh, this thing is so dumb. And it's really about the prompts. Right. Because if I just simply say, write an email for a client, there's not enough context there for it to actually do a fantastic output. So what we want to do is give it context and then say, ask me questions for clarity. And in that way, these systems are super, super smart. They'll start prompting you and say, well, I need to understand this, this and this before I can actually do the task that you gave me.
- [04:41]
A
Love it. And we're going to get into some really advanced features of ChatGPT today. But before we do, if people take advantage of of all the advanced features we're going to be talking about today, and we're going to get into that in a minute. What's the unlock? What's the benefit? What is the upside when people truly take advantage of everything ChatGPT has to offer?
- [05:02]
B
Yes. It was one of those things that was a huge aha. When I'm going through and I'm showing individuals push all the Buttons, like, just push all the buttons. So using the plus sign, you can see all the other modes that ChatGPT offers. And so being able to do that, there's scheduling capability, there's connectors, so I can actually start using my own data and connecting and getting work done. There's the agent mode, all these things that we're going to talk more about today, and then being able to do the deep research. So it's so much more than just writing and we want to go through and make sure that we all understand, encourage each of you to push more buttons.
- [05:40]
A
So let me ask you this. What is the benefit? What is the actual upside? That's really what I want to drill on. Like, hey, why should I push the buttons? That's really what I want to get to the bottom of. Right?
- [05:51]
B
Yeah. So the benefits of using this is time, right? We all are inundated with everything and it's time. So when I can use an agent or doing deep research or using the connectors, right? I can actually save a ton of time and not having to manually go through all the different websites to do a competitive review, or manually pulling some of those tasks or scheduling tasks for it to come back in and let me know about different tasks that can be done and achieved. So time is of the essence for all of us.
- [06:21]
A
Excellent. Okay, we're going to dive deep now into agent mode. So I would love you to first describe what the heck it is and then we can go from there. So just, just help everybody understand at a conceptual level what this feature is.
- [06:37]
B
So agent mode, think of it like a digital project manager. In agent mode, you're actually going to give it an end task. You don't have to give it all the steps. And I want you to do this and then this and then this. You actually just give it the end task, and then it figures out and navigates to fig out how to get to that end task. So that's why agent mode is so powerful.
- [06:57]
A
Okay, so unravel that a little bit more because most people that are using ChatGPT might say, well, ChatGPT helps me get to an end task because I want to write something. But, like, I think this is like way beyond that. Right? So kind of explain a little bit more about, like, what it can actually do, because I do feel like it's like going to blow people's minds. And then let's get into, like, why in the world would I even want it to do this?
- [07:22]
B
Okay, great example. So I was with a client recently and we were going through their ideal client profile. And my client was like, how do I find more people like this? And I said, a perfect example. Let's go use agent mode and ChatGPT. And so we went in in a selected agent mode. We fed it the ideal client profile, and we said, go find people like this. And then within agent mode, it pulled up its own browser and went through and started. You could watch it line by line, start and going and connecting and looking through to find those individuals and come back with that list of other ideal client profiles. And so I told him, I was like, make sure and vet the people. So we got names, clients, email addresses, address, websites. And so that would have taken a ton of time. So being able to point the agent in, what we were asking to do, it went and found all the information and brought it to us.
- [08:21]
A
Interesting. Okay, so what I'm hearing you say is that. And I've never used agent mode, so I'm asking these questions as a rookie, and I bet most of my audience hasn't used it either. But agent Mode inside of ChatGPT effectively opens up a browser window.
- [08:37]
B
Yes.
- [08:38]
A
And can kind of do anything on the Internet. Is that effectively what you're telling me?
- [08:42]
B
Yes, exactly.
- [08:44]
A
Okay, now, there's a lot of questions that people have like, why would I want to do this? And there's even some, like, trust questions. Right. Like, are you giving it access to your computer? I mean, talk to me a little bit about this.
- [08:58]
B
And there are dangers in using the AI browsers, and we do need to be very mindful of the things that we're not giving it, our credit card and different things like that. So think of it more as being able to have the connector. So I gave it my access to my Gmail. I was like, okay, I feel comfortable giving it to my Gmail, not my work email yet. I will eventually. And so I asked it go back through, like, the last three weeks of emails. If there's emails in there that I hadn't respond to, I wanted to find those emails and then put it into an Eisenhower matrix. And so that way I could know what was urgent, what was important, so I know how to navigate that.
- [09:37]
A
Okay, we'll get into the Eisenhower thing in just a minute, But I just want to make sure so we're crystal clear. So when you activate agent mode, like, walk us through, what happens? Do you have to use the app or can you use the browser?
- [09:50]
B
So it just pops up a little window. It's a private browser. And what's interesting is you literally get to watch it go through and do the work. So I'm seeing it saying, I'm at email section 306, reading the emails now, now going through and identifying, you know, which ones are unread, responded and now identifying which ones. Like it's telling me what it's doing in that little browser window.
- [10:14]
A
Okay, so you mentioned connected apps a little bit. Right. So when you're using Agent mode, I also understand that you can connect third party tools like maybe HubSpot or Salesforce or Google Docs and Drive and all that stuff in order for Agent mode to get access to all this information. Is it because you're popping up a browser and logging into your Gmail account, or is it because there's some backend stuff that's happening? Help me understand that a little bit.
- [10:44]
B
Yeah, excellent question. So you do have to give it the rights to those. Just like I said, I gave it rights to my Gmail, but it didn't give it to my work email. So all the different connectors, and you're right, there is a connector for HubSpot, so I can turn on that connector and then be able to go in and give it access. So it doesn't have connectors to everything right now. Right. They're building this out because it is fairly new, but within HubSpot, like I could go in, you know, turn on that connector to HubSpot, HubSpot, have it be able to view HubSpot, maybe pull some reporting back to me as well. So I can ask it to go through a very variation of different things and pulling from multiple areas, bring that reporting, you know, back and using it from an agent mode.
- [11:26]
A
Cool. We're going to get back to this Eisenhower Matrix thing in just a minute. A couple more questions. So when you pull up ChatGPT in like Chrome browser or whatever. Right. How do you activate the Agent mode and what, if any, steps do you need to take? At the very beginning when you're on.
- [11:43]
B
The screen, you just hit the plus sign or you do forward slash and that'll activate a dropdown for you to select the different modes you want to be in, like Agent mode or connectors or Study and Learn or Canvas or all the other modes. So one of my little tricks that I always love doing is when I'm especially using an agent, I will go into ChatGPT first, so just five first and say, here's the scenario I'd love to use in agent mode in ChatGPT. And so I just give it enough context and then I say what's the best way to prompt this scenario. And then it will tell me so, like I'll go in and give it maybe this much of content and say, what's the best way to prompt the scenario? And the scenario it comes back, but the prompt is as long as my arm and it includes a whole bunch of other things that I really do want to include. And so then I'll take that prompt and then use it in agent mode. And that way it's in alignment with the, the best system instructions.
- [12:41]
A
Okay, really interesting. Now, just out of curiosity, we know Google ecosystem integrates, so that opens up people's minds to presumably sheets and docs and calendar and email and stuff. Are there any other interesting connected quote unquote apps or could it be literally any website out there once you're in agent mode, like, help people understand that a little bit.
- [13:07]
B
So when you're in agent mode, again, I'm not, I'm just giving it an end task. I'm not really telling it where to go. So it could go through and look through many, many, you know, hundreds and hundreds of websites and then come back with the information.
- [13:21]
A
So does that mean, hypothetically, I could say, hey, I want you agent mode to go look at my insights from Instagram and Facebook. And then it would go to Facebook or whatever this tool's called and it would prompt you to log in and then it would know how to do all that kind of stuff. Is that kind of what we're talking about?
- [13:39]
B
Yes, exactly. So it will definitely prompt you to log in because it won't be able to do that. And you do just have to be really mindful of showing any API keys or again, credit card information. So when you're using an agent, make sure you've got human in the loop. Ask it to like, hey, when you get to this point, I need you to come back and ask me for credit card information or me for information.
- [14:01]
A
Okay, so now let's go ahead and go with this example that you were talking about with email and walk us through it very carefully because I think people are going to be fascinated by this. I believe you said look through my emails. I don't remember the rest of it, but share that and then walk us through kind of everything. How you develop that, please.
- [14:21]
B
Yeah, yeah. So I'm always very fearful of like dropping a ball or going through. And so I thought, oh, what a wonderful way to go and use the agent. And so I just turned on. So you go in and you have to turn on the connector first. And basically it's just asking you, hey, do you give permission to see your Gmail account? You can give it your calendar, you can give it your Outlook, you can do SharePoint, Google Drive. There's multiple connectors in there. Again, I wasn't ready to give it to my work email. I did give it to my Gmail because I primarily use that for e newsletters or different things that I sign up for.
- [14:56]
A
Well, and side note, some companies might have policies where they don't allow these connectors and you're gonna have to go to the admin to get permission. Right, like, because, for example, I know inside of Social Media examiner we have very strict policies and you might have to get exceptions to that if you're going to actually use your company stuff. But it sounds like a good place to experiment is on your personal stuff to begin with. I love that. So you connected your personal Gmail account to chatgpt Generic. Right. And then what happened?
- [15:23]
B
Yeah, so in the connectors and then I went in and again I asked just what I call meta prompting. I asked it what's the best way to run this scenario? You know, prompting this scenario and just gave it my thoughts of, hey, I want you to go back through the last three weeks of emails and identify emails I haven't responded to and then take those emails that you identify and then put them in the Eisenhower matrix. And then the Eisenhower matrix for those of you that don't is the four quadrants of, you know, urgency and how work gets done. And so within, you know, several minutes, I think it took maybe seven minutes, it came back and gave me, you know, the emails that I hadn't responded to. And it was in the categories of that, that matrix for me to make sure that there was urgent things that I needed to respond to.
- [16:09]
A
Okay, a couple clarifying questions. The web browser, which effectively is the agent mode, was that activated and then logged into your Gmail account so you could see it searching through all your emails? Yes. Oh, interesting.
- [16:22]
B
Yeah, it popped up a window and like I said, I couldn't see it literally clicking through Gmail. It was more of it was telling me what line it was like now reading, you know, emails 50 through 150. Now doing this like it was telling me what it was doing and it was like, now identifying the emails. Okay, now prioritizing those emails for the matrix. Okay, now finishing the matrix. So it just was giving me line by line what it was working on. And it was fast. Went through hundreds of emails very fast.
- [16:54]
A
So now the real question is like, how do you continue to interact with the agent, like once it's identified the really urgent things that you've not responded to, can you live interact with the agent or how does that work exactly?
- [17:06]
B
I was just doing that in test mode because that was literally the day after it launched. And so I haven't gone further than that. You can get into NCP and have more developers, you know, connectors to be able to go back to and now have more of an agent response. I'm just not there yet.
- [17:26]
A
So. Okay, to that other example with your ideal customer profile, what transpired there? Was it just searching the Internet or what exactly happened there on that one?
- [17:33]
B
Yeah, for it to pull all the like names and email addresses and websites and you know, anything that you know, it went through. We didn't tell it what sources. Right. We just said here's, you know, the ideal client profile, here's what we're looking for. And then that when it just started searching until it found them information and then we were able to verify the information as well because that was my biggest concern. It was just going to pull back a bunch of junk or they weren't relative to who we were actually trying to speak to. So we just told it to go find 20, just a sample size and so we could verify that and then gave it, you know, more. So the sales team was very thrilled because they were legit individuals and even in the sample size there were two people in there that they had been on the hunt for but didn't have enough data on.
- [18:21]
A
Do you feel like this could be like exported as a CSV file or what's the potential of, you know, because like to make the matrix or you know, or to make like a list. Like how do you get that back into the other system? Where do you see all these interconnections going?
- [18:34]
B
What's beautiful about ChatGPT, you can always ask it to provide a format, a downloadable format of Word Excel. Right. So you can ask for a downloadable format. So you can definitely do that.
- [18:46]
A
Love it. Okay, so what's your thoughts on like where this whole agent mode thing is going inside of ChatGPT? Like, like if you had to imagine months or a year from now, like where do you see it going?
- [18:57]
B
It's more of things that I think about is probably more consumer related now just because of just different things that I've seen individuals using it for. So it would know that I love to fly Delta. It knows that I have a Marriott or Honor Hilton Honors. It knows like where I like to shop and my Sizing. So I could go into the agent and say, I am going to a wedding. I need you to book a flight. Here are the dates. Find the appropriate flight for that. I need you to also book the hotel. I need a dress, wedding, guest dress, you know, where I like to shop, you know, my sizing. And so I'm just again, giving it the end goal. And it's going to go through and search and find the best deals, best flights, hotels, and then come back to me and say, okay, I have found this flight. Does this work? And then I would, you know, go in and put the credit card in, not giving it to the agent, just go in and do that. So that's really just that assistant in your pocket, right, to go through and navigate some of those things where I'm not spending time going through because the AI knows me and it knows my preferences. And so for all of us here, from a marketing standpoint, right now, it's understanding we have to market to these AIs. So it's selecting our clothing, our hotels, our flights. So being able to navigate that. So a big shift that we're all experiencing.
- [20:22]
A
I want you to read something a business owner told us after Social media Marketing World 2025. She said, quote, the AI teaching was mind blowing. I realized if I hadn't attended, there's a good chance my business would have been obsolete within a year. I'm now so far ahead of my competitors, unquote, said agency owner Melanie Miller. That hit me hard because that's exactly why AI feels so urgent. And right now, it's not hype. It's a real competitive advantage. So here's what we did. We created AI Business World 2026, our first dedicated AI conference. An entire conference focused completely on practical AI marketing training. It takes place April 29th and 30th at the Anaheim Convention center in California. Walking distance to Disneyland. You get 20 sessions to choose from with AI practitioners who will show you exactly which tools work, how to build workflows that save hours every day, and how to implement AI without losing your authentic voice. You'll leave knowing which AI tools to use for each marketing task. No more guessing. No more tool paralysis. Just clarity and confidence. If you have a Social Media Marketing World 2026 ticket, this is included free. Otherwise, early bird standalone Tickets are only $597. Don't get left behind. Get your tickets at AIbusinessworld live. Finally, AI training that actually makes you irreplaceable. What makes this different than the other features we're about to dig into, which I and you have both used, and we'll reveal in a minute here, is that this isn't just going out and finding information, it's actually doing something personalized with the information. And that's kind of the unlock, right? The idea of potentially maybe updating an Excel spreadsheet or Google Sheet or possibly even entering data into a database or something like that. That's where all of a sudden it gets really fascinating because everybody understands the power of these tools to gather information, to analyze information, but the next step is to actually do something with it that doesn't require human interaction. And that's the key part of it, right? Agent is acting as if it is a staff member or an assistant working for you. And pretty soon I can see and I know other bigger entities are working on concepts like this, special agents for specialized tasks. Your personal agent might be your personal assistant. Your marketing agent might be your in pocket marketer that was trained on all the marketing things. Your sales agent, your outbound sales agent might be the one that kind of reaches out on your behalf. I mean like you can see this is where it's heading and it's kind of mind blowing and scary at the exact same thought. But it's super exciting to see all these like, like things that have been independent features kind of all becoming, for lack of better words, automated. And I love what you said about like, hey, use chat GPT first and just a general thread to say, look, I plan to use agent mode. This is what my outcome, my desire is. Help me write a really nice prompt that will increase the likelihood the agent gets what I want right out of the gate. That's kind of a big unlock. Okay, so we've spent a lot of time on agents. Let's move over to deep research, which some people are familiar with, but it can do so much more than what a lot of people like might understand. So why don't you describe what the heck deep research is, why it's important, and then maybe we can talk about a little bit about how you're using it.
- [23:52]
B
I love deep research. So instead of us manually going through, maybe we're doing a competitive analysis or market research and we're having to manually go hunt, search, look through hundreds of websites. It's just very manual and it's a huge task. And so having to do that, being able to use deep research, pointing it in a direction of what the task is and then it's going and doing the research, right? So being able to navigate that very quickly, like thinking of market research. So a client of mine went through and they're in 52 countries. And he said, look, we need some market research in 2030 in this particular area. We really want to just focus on east coast for now before going to the other markets and we want to do some validation. I was like, have you asked AI? And they're like, no, how would we even ask AI? And I showed him deep research. I said, well, let's run some research right now. So we pulled up deep research and again I did my meta prompting. I said, hey, here's all the things that we're looking for this market research in 2030 in this industry, this part of the country, very specific. And then said, what's the best way to prompt the scenario? And the prompt was very long. I put that into deep research. And then in, I think it was about 12 minutes, it went through 173 sources, read 63 pages, and then produce a full market research for them. And it was about 13 pages long. And so all in 12 minutes in this deep research. So I told him, I said, hey, I want you to spend some time with this, validate it, talk to your leadership team about it. And now I want to know like how close or how off was it it? And it gives you all the sources. So you know, exactly. It wasn't hallucinating. It gives you the sources. And so he very happily called me later that day and he said, you won't believe this. He said one, it was very validating because there were things in there that we had already kind of been planning towards and there were two considerations that they had not been planning for. So it was just huge. And so I asked him, I was like, what would you pay for research like that? And he's like, probably around $30,000 is normally what we would have paid for a research like that. And we were able to do it in 12 minutes.
- [26:07]
A
So this is intriguing. So first of all, in your opinion, what differentiates deep research from agent mode? Like, just help people wrap their brain around this a little bit.
- [26:15]
B
It is very complex, right? So agent mode can go out and, and seek again in task, go out and seek all the different things. Research mode, you're really thinking more of being able to quote the sources. Think of it more of like putting your MIT hat on. It's going to be more academic, more research based, giving you the sources. It's not acting on your behalf, it's just doing that particular task.
- [26:43]
A
Let's ideate on some different kinds of ways. And I'll share one. And if any other ideas come to your head, that would be helpful. So we are developing a software product in house and one of the things that I wanted to do was. You ever been to a app where they have these cool little, like, arrows that point at the various steps that you need to do, like in the onboarding process. So I asked Chat GPT to first of all help me figure out what the heck that thing is even called. And then I asked it to do deep research on the various tools that are out there. And what's really cool is it did an extremely thorough, comprehensive analysis of like a bazillion different tools. It stuck it into a matrix. It knew kind of what I was looking for. And it said, here's the pros and cons for a small, you know, startup software company versus the more sophisticated. And all that kind of stuff was ridiculously valuable to me. And it sent me down a rabbit trail where I was able to, to find something that I would have probably spent endless hours on Google search and not even found what I was looking for. I'm just curious, have you had similar kind of experiences with deep research? And if so, I'd love to hear you share a story.
- [27:55]
B
Yes. Yeah. And I love that tools being able to do that, and I love that it spits back a matrix, you know, based off of, you know, here's the things that you are considering from a buying standpoint so you can do the same thing from a vendor. Maybe you're trying to vet a vendor. Maybe you're an RFP and there's five vendors. Now you could go in and do deep research, pull those RFPs in and.
- [28:16]
A
Say, hey, like, here's the five vendors.
- [28:18]
B
Okay, now, now go do some research on these five vendors. You know, here's, you know, the things that are important to us. Here's, you know, the rfp. Like, go look at these five vendors and then again, do the research and then put it into a matrix. Being able. Almost like a rubrics, right, of how you go about making your decision as well.
- [28:38]
A
Yeah, I like that. Another deep research project I was doing was I wanted data to affirm a hypothesis that I had made. So I believed that marketers were fearful of adopting AI because it could replace their job. So I asked it to just go out there and find every kind of study it could possibly find to either prove or disprove that hypothesis. And that was really helpful. So, like, I think that anything where we would go do Google research and look at and compare and contrast things could be fodder for Deep Research, but it doesn't have to be that. It could even be something, I don't know, more complex. Like what if you had. I don't know, I'm just, I'm ideating with you right now, but have you heard of any other unusual applications for Deep research beyond the ones we've just shared?
- [29:30]
B
Just competitive, you know, analysis, kind of a SWOT analysis.
- [29:34]
A
Well, that's a good idea. You could say, hey, here is my company, here's my product. You could say, go find small upstarts that are employing artificial intelligence that I might want to have my eyes on. Right. Because maybe I'm a traditional business and I would imagine it could do that as well. There's really no limit to what it could do. From your experience, do you find that it's fairly accurate or that it needs a lot of like cross checking the final information that comes out of it?
- [30:06]
B
Yeah, I find Deep research does a much better job than just regular, you know, chatgpt, but Deep research, because it always is quoting the sources of where it found the information and then being able to click on the links. Because I find ChatGPT sometimes will give you links and you click on them and you get a 404. Right. The page doesn't even exist.
- [30:23]
A
Right.
- [30:23]
B
I find it does a lot a better job in that area with Deep research.
- [30:27]
A
Do you think it's possible to start with Deep research and then go to Chat GPT agent mode and potentially have it go even further? Or do you feel like deep research is kind of the way to go? I'm just curious if these two, two things could work together.
- [30:41]
B
Yeah, I don't see any reason why you couldn't. I like that idea. I don't see any. You could definitely do that. Take the research and then go activating agent mode and say, okay, now I want you to go through and click on all these sources. Right. You could, you could totally do that and say, I want you to go validate all these sources.
- [30:58]
A
Here's another crazy idea. Hope people don't mind these crazy ideas, but let's say that you have a product or service in your industry and you start with ChatGPT and you ask it to identify who it believes are your top 10 competitors. Right. And to identify their unique value proposition and key messaging that they use. And it would be able to go out and do all that, I would imagine, right?
- [31:23]
B
Yes, for sure.
- [31:24]
A
Then what you could do is you could take this over to agent mode and you could say, hey, agent mode here's some information about my number one competitor. What I want you to do is take a look at my messaging up against their messaging and identify opportunities to improve or alter this information. I don't know, I'm just thinking creatively there. Do you think something like that would work?
- [31:50]
B
I love it. No, I love it. I absolutely love it. I think that's very smart to be able to do that. Like I just have always had the mentality of, you know, the old Simon Sinek when he says in his book Infinite, right, that Apple doesn't care what Microsoft's doing, right. So, and, but Microsoft cares what Apple's doing. So I've just always been the don't really worry about what competitors are doing because I have my, my space. So I love being able to focus in on that. But I love that idea. I love that because there are a lot of people that they do want to do that deep research and understand those key findings and differentiation.
- [32:28]
A
Okay, we're going to move on in just a minute to another feature, but one thing I wanted to ask you about is actual research. Have you ever uploaded data from like Excel CSV file or like a PDF and asked it to run analysis? Because if you haven't, I have and I have some insights to share because there's two kinds of research. There's go out and find what's out there and then there's analyze this special research that I have access to. Have you ever done any analysis inside of ChatGPT?
- [32:58]
B
Yeah, definitely analysis, but I would love to hear your take on it.
- [33:02]
A
Well, okay, so here's the deal. I have a new AI study that we're dropping in October which probably will be out by the time everybody listens to this, called the Marketing AI Industry Report. And one of the questions I asked in the report was what's the biggest problem you face today when it comes to AI in your work? And it was open ended. So as a guy that does a lot of research, because I've been doing studies kind of forever, I wanted to see how ChatGPT could take open ended data like this. So what I said is I went into Chat GPT and I said, you are a research analyst who has a specialty in messaging. Your job is to do a qualitative analysis of the attached information and to identify themes that you see inside the message. Now in the past ChatGPT could not do this. Okay, Claude was much better at this. But now with ChatGPT5 it's actually pretty darn good. So what it was able to do was it was able to read a export of a CSV file that had like 700 rows in it and it was able to identify not just what it thought were the top seven themes, but to pull little snippets of messaging from it that I that can now use in my marketing materials for, for all the AI focused stuff. So this is something that a lot of people probably don't realize is that chat GPT5 is actually really good at data analysis. So now when you start to connect all the dots together, and I hope everybody appreciates this rabbit trail we're going down. But we've got the agent mode that's going out there and acting on your behalf and maybe gathering up information for you, maybe even based on your own own databases. And then you've got deep research which is gathering more public facing data and then you actually have chat GPT5's ability to actually do data analysis. Like this is magic sauce kind of stuff. Don't you agree?
- [35:00]
B
Totally agree. Well said.
- [35:02]
A
Okay, the next. Do you remember what the next thing is we're going to focus on or do you want me to tell you what?
- [35:06]
B
Okay, I do.
- [35:07]
A
Okay, let's do. What is it? Let's hear it.
- [35:09]
B
Vision mode.
- [35:11]
A
Some people don't even know what that means, so go ahead and explain it in detail please.
- [35:14]
B
Yes. So AI's ability to see, learn and do. Right. It can see. And so I can't wait for when desktop can actually see for ChatGPT. You can actually use it for. Think of it kind of like an assistant or help desk. Right. You've got it right in your pocket. You can go in and pull up ChatGPT in the app, hit advanced mode and then you have a camera and then you have the vision capability. And so with that vision capability, I can hold up my camera to anything on my screen and say talk me through how to navigate this page or talk me through how to solve this situation. So it's like your own little help desk. How many times have we been navigating something in a new marketing system and we're trying to navigate it and we want to be able to have that assistant with us. So being able to have that vision capability literally can see your screen. And so that's why I keep saying I can't wait for it to actually be on desktop. It's coming. I think Michael, you and I had found that it's going to roll out Mac first, before Windows.
- [36:23]
A
Yeah, I happen to have the app and it said it could do it, but then it can't do it. So let's describe really precisely how to use this. First, you need to have ChatGPT on your phone, and maybe you can just kind of walk through how people can activate it if they want to actually try it out while they're listening to. To this podcast. Because they're probably out and about with their phone right now anyways.
- [36:40]
B
Oh, yes, yes. So pulling out your phone, going to ChatGPT and then clicking active mode. Right. So it'll be where the little chat window is, not the microphone. Right. The little icon next to that, that's being able to do the active advanced mode.
- [36:55]
A
It looks like a little sound wave kind of symbol. Right.
- [36:58]
B
Thank you. I escaping the words of what it looks like. That is exactly right. A sound wave.
- [37:02]
A
Yeah, I'm pulling it up right now. It looks like a circle with, like, a little vertical lines inside of it, which is indicating, like, sound. So when tap on that, explain what happens next.
- [37:11]
B
Yeah, so when you tap on it, it'll give you a little video and. Or just, you know, the. The microphone. So with the video, when you open that up, you just simply state, so if I had chatgpt up on my screen, I've got my phone, I'm like, hey, help me navigate. You know, I want to build a GPT. Talk me through the steps of how to build a GPT. And then it literally will say, I see your screen, I see that you're on ChatGPT. Go and click on Explore GPTs. Okay, now you've done that. Now on this screen, you want to click on, you know, Create in the top right hand corner, click Create. And then it just keeps going through step by step by step. And like, all the years I've been in marketing, this would be so, so helpful in trying to, you know, shortcut and navigate, you know, things that were new in learning. That's not just rhythm anymore. So think of it as just literally a help desk in your pocket. It.
- [38:03]
A
So just a couple thoughts on this. First of all, when you do open up the app on your phone, you typically will see this big blue circle show up right where ChatGPT will start talking to you. You'll see the camera in the bottom left. You can have it in selfie mode, or you can flip the camera to the other side, which is really important, just like any other thing. Now, I really like this application you talked about about, like, effectively just having it be there in your hand, showing it what's on your screen and saying, what the heck am I doing wrong? And it will actually try to fix things. And a short little story. I had a mobile mechanic in my driveway. My sunroof on my Lexus had a little piece of plastic that fell off. And then we could not figure out what the heck it was. So I opened up vision mode and ChatGPT said, oh, it looks like it's blah, blah, blah, blah, blah. And based on what you're showing me, I think his jaw draw dropped. He couldn't even believe it. And it helped him because he couldn't figure out where it goes. It helped him figure out that this piece of rubber or whatever needed to go on this one rail in this one way, and he was able to solve it. And he. He was shocked. He's like, I am getting this whenever I'm out servicing my customers. But what I really love about what you're saying is, like, the fact that it can see your screen. I would imagine almost anything you could ask it. Like, you could say, hey, look at this. These charts. What do you see something I don't see? I mean, like. Like, what kind of stuff have you or other people done with this? Or have you just barely tapped the surface of it? I'm just a little curious.
- [39:31]
B
Yeah, and you're right. It's whatever's on the screen. Or maybe you keep getting an error. Maybe you're doing web development and you're like, I keep getting an error in this area. You know, show me. Talk me through how to resolve the error. And I've even used it just like you were talking about for a vision. There was something wonky going on with our washer. It was like, you know, kind of shaking. And our stepdaughter had put way too many clothes in. She was visiting and knocked it off. And so I pulled up ChatGPT and Vision Mode started talking me through how to unhook the back of it and look for an area, look for a specific spring, see if it's connected or did it come dislodged. But just again, having that in my back pocket to navigate that, and I was easily able. The spring had just kind of unhooked itself and hooked it right back and put everything back together.
- [40:18]
A
That's really fascinating. Now, some of you are like, how in the world am I going to hold my phone up while I'm interacting with my computer? I was asking Kendra this when we were prepping for this. If you have one of those little tripods, that would be really handy. It's almost as if ChatGPT is like, almost becoming like the little brains of a Robot, Right. And it's almost like the robot right now lives inside of our phone or it lives on our desktop. But I can see a day where all these things that, like you said, when this comes to desktop and you can actually share, share your entire screen. This is going to be a big unlock for a lot of people. Don't you agree?
- [40:53]
B
I totally agree. And it's coming. It's just a matter of time and I'm expecting it and, you know, anytime. Because Google AI Studio already has it. Like, you can go to Google AI Studio right now and it can see your screen and stuff, you know, walk you through everything. So it already exists.
- [41:10]
A
So when you think about this agent mode right now, it is operating through a little web browser. But very soon, I can imagine it would see your entire screen screen. And I would imagine OpenAI is going to connect all the dots together. The audio, the scene of your screen, all these kind of things are going to come together. But they're doing it in obvious steps where they know people are really concerned about security. But they're also smart enough to know that, hey, pretty soon people are going to be willing to give up some of these security concerns because of the massive advantages that they're going to have. I can see a day, Kendra, where maybe you just have an extra monitor and you've got your AI agent on that monitor and you're on the other monitor and it's just talking to you like, hey, come over here real quick. I got to show you something. Right, you see, that's what.
- [41:54]
B
That's where you're just co working. Yeah, I love it.
- [41:59]
A
I know. We've just tapped the surface of all the great things that you've got going on. Why don't you tell everybody where they can discover you on the socials and if they want to connect with you outside of social media, where. Where do you want to send them?
- [42:10]
B
Oh, thank you. I'm really easy to find on LinkedIn at Just Kendra Ramirez and I also on my website@kendra ramirez.com and I do have something for your guest. I did build a page that has my comprehensive guide on everything that we talked about today. And that is just Kendra ramirez.com SME.
- [42:30]
A
Kendra, thank you so much for sharing your insights with us today.
- [42:33]
B
My pleasure, Kendra.
- [42:35]
A
If you missed anything, we took all the notes for you over@socialmediaexaminer.com a76. Be sure to follow this show on your favorite podcasting app. And if you've been a listener for a little while, we would love a review and if you'd be willing to share with your friends, that would be awesome. I'm most active on Facebook, LinkedIn and X do. Check out our other shows, the Social Media Marketing Podcast and the Social Media Marketing Talk Show. This brings us to the end of the AI Explored Podcast. I am your host, Michael Stelzman. I'll be back with you next week. I hope you make the best out of your day and may AI help you become more successful.
- [43:13]
B
The AI Explored Podcast is a production of Social Media Examiner.
- [43:21]
A
Don't miss AI Business World 2026, our first ever dedicated AI conference taking place April 29th and 30th in Anaheim, California. Learn from top AI practitioners about which tools work and workflows that will literally save you hours daily. It's included free with Social Media Marketing World 2026 tickets, or it's only $597 as a standalone ticket. Secure your spot at AIbusinessWorld live before pricing increases.