Product of the week
(For a summary, click here.)
John Eischeid: To summarize, give me a quick summary – like five or six lines – about yourself and the work that you do.
Romina Massa: So today I am a digital marketer working remotely and locally specializing in personal branding, specializing in events marketing and in a lot of digital marketing on the back end as well. I’ve been doing some SEO for clients and some ads.
JE: I noticed on your website that you have what you call a human- centered approach. So, how does that differ from other marketing approaches?
RM: If you wanted to talk about which marketing approach does it is it the most similar I guess it would be account-based marketing and I think what the way that it differs from a lot of marketing approaches is the way that it makes account-based marketing the next thing and everything that we should be doing. It’s really targeting one human at a time. And it goes not just in the way you approach content and as a systematic way to talk to one person even if you’re talking to a room full of people. I don’t mean it metaphorically like that. I mean really it is about building one-on-one relationships and the metrics that matter are more your profile views and not your impressions. It’s human- centered marketing because it’s based on human relationships. It’s based on lead nurturing. It actually leads from a place of no cookies, right? We’re moving towards a world of first-party data, right? And so we need to be doing that. We need to be connecting on a personal level. We cannot be connecting the way that we have been before. And so I think that’s what makes human-centered marketing so human- centered.
JE: So one could say it’s an offline approach.
RM: It could be. Yeah, I guess.
JE: Yes. Okay.
RM: Yeah, I like that. It’s interesting.
JE: But you also mentioned that you also work with a lot of data and you said that your approach is data-driven.
RM: Very much.
JE: How do those two combine?
RM: It’s not how they combine, but really data drives everything I do, from my content to my strategies, right. I’m very data driven because I am looking at the back end to make every single decision for myself and for my clients. If I’m posting about something and not about something else. On LinkedIn you get to be very authentic, so if I say something, it’s because I want to say it. But it is also because I know what is working. I know what kind of content is driving engagement. I know the kind of ways that you want to approach it. And I do that because I’m looking at the data. And when I say “data-driven,” I also mean that metaphorically and figuratively and literally because if you look at my content, I actually speak of data as the content itself. When I make a chart, when I make a graph, I’m showing you my own numbers. I make AI graphics of myself looking at my own LinkedIn impressions and it’s a real life graph of my LinkedIn impressions over the last seven days. And so I make content that’s based on data as well. I just don’t know how to frame the value of what I do. It’s thought leadership, but it is educational. That’s where the value is. There’s something that you take with you that you can take and it’s actionable learnings that you take from my content, so I guess that makes it educational in a way. I minored in education. Can’t help myself. But I mean that’s really a little bit the point of it.
JE: And you also have described AI as a tool and a lot of people are afraid that it might actually be a replacement for human work. So how do you draw that distinction? Why do you differ in that assessment?
RM: Prompting. So AI becomes a tool and not a replacement when there is skill in the prompting. If you don’t know what you’re doing by manipulating the AI and managing it and in this prompt engineering skill of telling the AI what to do then – what I see in AI as a tool is the skill. How far can you take it? What amazing things can you do with AI that actually are truly generative, that truly are new because there are so many things that take reference from others. We can have that conversation around how much of AI is copied or not, but in reality, how sophisticated your prompt is going to really show or mirror – if your prompt is a sophisticated prompt that has skill in your prompt engineering, it’s going to create an output that really truly reflects the human behind it. And in that way, it’s not replacing anybody because that output would never have been possible without the human that put in all of that time, all of that effort, all of that skill, all of that practice into the prompt or into that the set of prompts that it took to create something using AI. If you’re looking at AI as an avenue for automation, and yes, it can do what I can do 10 times faster. Yes, of course, it can replace several things, so did computers. I mean there are so many conversations around, “Where is the line between just normal evolution and the pervasive advancement of technology?” The part of it continues to be just normal evolution and progress. The problem with AI is that it’s growing at such a pace and at such an unregulated pace that I see so many ethical pitfalls and I think that’s somewhere else where I like to say my work is human-centered because it’s used ethically. I can only promise how ethically I use the tool and I can say that I use it with honor and responsibility if I can say that what I mean, but I can only speak for myself as a human and I hope and I wish that we all could. I know that it’s not the case, but I think ethics are and are important in the way that we use AI and we should be talking about that.
JE: So that leads well into another question that I had – which was along the lines of, “What do you think AI is very good for? What do you think it is not good for?” And one of those not good uses, for lack of a better word, would be unethical. So, can you just expand on that distinction between what you think it’s not bad for?
RM: What do I think is a good use of my time using AI or not? I could use AI to pop out five blogs in three hours. Or I could make two blogs that are better, that are good, that actually reflect me using AI as well. It’s still a lot to produce in two hours, three blogs or two blogs versus five blogs. You’re still producing a whole lot more than you would if you were writing the blog from scratch, right? But you’re not using it right when you maximize the tool so much that you didn’t even read any of the five blogs that you put out and you published them anyway. And trust me they will and trust me that they’re publishable.
RM: Your AI has been trained for a couple of months. If your own engine has written a hundred articles by you, the likelihood that it’s going to pop out five legible and publishable articles that you don’t need to read is there. It can do that. I don’t think anybody should be allowing that to ever happen. We should not be populating the world with unregulated, unhuman content, even though we can. Because of crawlers, because of bots, because of consistency, because of algorithms. I know that the opportunity is there but I think it’s – you’re losing – that’s I don’t know if it’s unethical, but I don’t see that as a good use of AI. I see that as a waste of all of it. And in that sense you are making light of all the skills. You’re taking away the skill of the AI. You’re taking away the skill of the prompt. You’re taking away the skill of the copywriter that never got to write this blog. You’re just doing a lot of wiping out and not putting anything good out in the universe. You’re not leaving anything. So maybe that’s a good way to put What is good AI for? The good use of AI is if you can make a good impact with it– something you can leave behind something of value and you’re not just regurgitating. When it’s truly generative.
JE: Okay. I see what you mean. How do you deal with AI? Well, let me back up a second. So you really just use large language model, and when you say AI, do you really just stick with large language models and generate content that way? Are there other types of AI that you use in terms of data analysis or anything like that? Or are you strictly within the content/LLM sphere?
RM: I’m staying within the LLM area for now, especially when it comes to how you use it because that’s what I use the most. So, I wouldn’t want to go into areas where . . . . [shrugs]
JE: Having said that, how do you deal with AI hallucinations in your line of work? [laughs]
RM: First you get a good laugh, because honestly, sometimes but it’s really funny, because you just have to pay attention. Like, this is just the point that I’m making, right” You can’t take the human out of the equation, because the AI will hallucinate and the hallucinations are dangerous I use AI a lot, a lot, a lot. That means I use AI to apply for jobs. It means I use AI to create content. And it means I use it to write emails. Those are just three things that I do every day. One thing that I do that I take very seriously in terms of the responsibility is creating alt-text, just alternative text.
JE: Are you talking about the alternative text associated with images on websites?
RM: Accessibility, backend, metadata stuff.
JE: So, SEO, that kind of thing.
RM: Yeah. But it’s not just SEO. It’s also accessibility, because anybody without that doesn’t have access to images. Anybody with an image reader is not going to have access to your image unless they can read the description. So you’re providing alt text also for accessibility to make your content more accessible.
RM: I think that is a huge responsibility we have as humans and marketers and I take it seriously, and there is not a single image that I have uploaded anywhere that gives you the opportunity to apply alt text that I don’t apply it. Every single one of my LinkedIn posts, every single one of my blogs, all of my content is accessible. But one time, I was creating alt text for 10 slides and it hallucinated the entire situation, and it was the weirdest thing in the world because – why wouldn’t it do it? I’ve done this hundreds of times. Why would my bot, Joy – I have named her, the irony – why would Joy lose her mind in such a way that she completely forgot what alt text was? And I was in what is a flurry of steps that it takes to publish one article. To publish one article, I have to post it on HubSpot, and then I also want to post it on LinkedIn, and I have to SEO-optimize both backends, and I have meta descriptions and slugs and all of the things on alt text and everything. So it would have been very easy for me to overlook this alt text, and just copy paste it into the slides. I’m so grateful I didn’t, because it wasn’t even funny. I mean, it was describing a completely different situation. And I still don’t get why. My picture was of me standing next to a table at brunch looking at the ocean and the alt text was describing a set table for two at brunch with food and coffee and there were no people but the table was set for two. The situation was different. It was a completely different scenario. And so here I am providing a huge disservice to society, by telling somebody who doesn’t have access to this image that the image is completely different and also creating a whole lot of SEO confusion in terms of what this image is doing. So I’m not helping myself in any way in terms of pushing my own content. It was a complete disaster, but it was 10 slides and so it was easy to overlook.
JE: You mentioned that the bot’s name is Joy. Is there a particular model that Joy is built on?
RM: I’ve committed to Chad GPT. I’ve tested a couple. I’m on a Plus subscription. I’m not going to pay $200 a month, even though I would, but I’ve had others. The moment it launched o3-mini, to be honest and what it’s doing with coding. I’ve taught myself how to code using AI and I’ve been building interactive content modules and quizzes and the noir detective games and things. And so I appreciated the upgrade when they launched o3-mini. I actually saw how much better at coding it got and I was like, “All right, no, I’m sticking with this one.” and overnight the graphics have gotten ridiculous. Just the quality.
JE: Yes, I have seen the recent images that it’s producing. They are impressive.
RM: It’s really impressive.
JE: You also mentioned on your website that you have a crosscultural approach. Do you use AI in any of those ways? maybe for translation or I don’t know research that kind of thing.
RM: My AI is bilingual because my clientele is bilingual. I already have had some Spanish- speaking clients that either prefer for me to speak to them in Spanish, even if their content is in English or need me to make their content in Spanish entirely. And so I use it interchangeably between English and Spanish and Joy gets it.
JE: Okay. I don’t have any further questions. Is there anything else that you would like to add?RM: No. This was fun. It’s actually a weird opportunity to get to talk about the things you believe in.
