Let AI Do the Admin

Richard Lowe: All right. Hello. Welcome to Leaders and Their Stories. This is Richard Lowe, and I’m here with Yana, who’s going to talk about all kinds of cool things. If you could introduce yourself, Yana, that would be great.

Yana Semova: Yeah. Thank you very much, Richard, for hosting this podcast for me. My name is Yana Semova. I’m an economist by diploma, so to say, and I’m a project manager by profession. I’m also a mentor, a coach, and a consultant. Throughout my education, I studied at the American University. I did my master’s in the UK. I did an exchange in the Netherlands. So I have quite some experience in the academic setting, in lots of academic institutions throughout Europe. Then I started my career path as a consultant in some engineering and construction companies.

And then in early 2017, I switched to the IT field, where I became a project manager. Since that moment onwards, I have been developing myself as a professional in this field. And lately, I’ve been part of mentorship programs and coaching programs, where I teach and help young talents and students advance and develop in this field.

Yana Semova: From the beginning of this year, together with a business partner of mine, we founded ProGrowth Academy, where we aim at helping people develop themselves professionally: advancing their career, academically, intellectually, personally, all kinds of development. This is why the academy is called that way, because we are aimed at any kind of growth in any direction.

The main aim of what we do, first of all, since AI has become quite popular and modern these days, is that we as professionals would like to train more people to use AI, to empower them to use AI, so that they can limit the time used for administrative stuff and focus on the really meaningful tasks for them, the activities that AI cannot do for them in the truly managerial role. We do this for project managers, for salespeople and for business analysts. At least for this moment, this is our scope.

Yana Semova: In addition, what I do in the academy is prepare students to apply for university abroad, to have all their documentation ready: cover letter, CVs, recommendations from professors. We also make a strategy together for how they will apply, and I help them choose a proper, relevant university program for themselves. And I also prepare people who are looking for a job, or who are unsatisfied with the job they have and are looking to switch.

I prepare them for a job interview, and I help them prepare a CV and cover letter that make them stand out from other candidates and give them a chance against more prepared and experienced candidates.

Richard Lowe: Okay. Sounds like quite a mouthful. Lots of things. So you’re in Bulgaria?

Yana Semova: Yes.

Richard Lowe: What’s it like over there these days?

Yana Semova: It’s challenging. I don’t know if you’ve heard in the US, but across Europe there’s quite some crisis going on. The companies, especially the big corporations, are facing difficulties holding on to capable people. There are layoffs everywhere. Big companies like Bosch and Takeaway have left Bulgaria just recently. And it’s just the beginning. There are budget cuts all around, especially in training. So it’s very difficult and challenging for consultants like us to provide services, because there simply isn’t budget dedicated to such services from external companies. The economic situation is not very favorable at the moment.

There isn’t much funding from the EU, and it’s quite difficult to find sponsors and people to help you expand and develop your startup.

Richard Lowe: Okay. Now, if my memory is correct, Bulgaria is near Romania.

Yana Semova: Yes, exactly. We’re very, very Eastern Europe. Next to Turkey, even.

Richard Lowe: You were part of the Iron Curtain, I believe, Soviet occupied back then. Way back when.

Yana Semova: Yeah, exactly. We have an imprint from that era as well. And from the Orient before that.

Richard Lowe: Interesting, interesting. Okay, good. So you’re heavily into AI. So am I. I use AI as a tool. I don’t use it to write, but I use it to help me do all of the grunt work around writing. It could be anything from writing emails to coming up with an outline to having a profile on a character in a book, stuff like that. I’m a tech guy, so I’ve had 33 years in technology, basically at the director level and the VP level running a tech company. So I’ve got a lot of experience in a lot of areas. Plus, before that, my major in college was geology.

So I’ve been all over the place, just like you.

Yana Semova: Yeah, quite a diverse portfolio and background you have.

Richard Lowe: Yeah. Never been out of the country except to Tijuana, or Ensenada, which really doesn’t count because it’s just next door in Mexico. Never been to Europe. Don’t plan to. No particular reason, I just like the United States. It’s a good place to be. So what’s your view on AI? Obviously, you think positively about AI, but let’s go down to a little more micro level. How could AI be abused?

Yana Semova: You mean abused by people, or used?

Richard Lowe: Well, there are a lot of questions about the ethics of using AI in general. So how can a person ethically use AI in their life?

Yana Semova: I think AI could easily be used by almost everyone in a variety of ways, in their everyday life, as you already mentioned, and in their professional life. For example, in my profession, what I do on a daily basis and what I teach people is how to make good planning for their work. It could be their project, it could be a wedding, it could be anything. You can also do a financial plan for your project or for the thing that you need to do. It’s very simple and easy to run in AI. You can make a budget forecast. You can also engage stakeholders through AI.

As you mentioned, writing emails and tailoring the communication to the different levels of seniority. This is very helpful. You can do all kinds of reporting with visuals. And when you have a company of your own, you need to advertise, so you can easily prepare the posters and advertising materials on your own.

Yana Semova: However, as you mentioned, you need to be really careful with the data that you’re using. There are many companies that are not allowed to use any tool except one. For example, Microsoft Copilot is the only tool permitted at some of them, due to data sensitivity, because it’s the official Microsoft tool that protects their data. Although there are other tools that could be more efficient for their work, that’s not an issue, because through prompting and making more precise commands you can achieve the same results.

But you need to strictly adhere to the rules of the company, of the enterprise, in order to be secure in terms of data. And personally, when we do our work at home, we also need to be careful with the pictures that we’re posting and the data that we’re sharing, because especially with ChatGPT, OpenAI, everyone can access your data and your pictures, and it’s exposed, so to say. You don’t know who’s going to have access to this information. Therefore we need to be careful how we deal with it and what we share. This is my personal view.

Richard Lowe: Of course. Well, it’s important to understand, as you say, the environment. What are you allowed to use, and which one is the best one to use, because some are better than others. And it changes constantly. And which company has the best policies and procedures for AI? Anthropic just came out with new AI terms and conditions, which I think are very good. There are certain things you are not allowed to use their product for. And I agree with it. A lot of writers won’t. But that’s fine. They don’t have to. They can use a different tool.

I think Anthropic is proving to be one of the more ethical companies, as ethical as you can call a trillion dollar company. And from what I’m seeing, ChatGPT is turning out to be one of the lesser ones, from an ethics point of view.

Yana Semova: Yes, it’s the least reliable, so to say.

Richard Lowe: In fact, did you do any research on the Hugging Face break in?

Yana Semova: I haven’t really researched that one, but what I can say from experience is that Anthropic is definitely one of the most ethical companies in the market. Claude has a very high level of reliability, and Gemini as well. But what also matters is that every single tool has different levels. If you purchase the premium version or the enterprise version of the tool, the level of security also rises. If you purchase only the basic package, then you’re not very secure.

But if you have the ability to invest in a tool that not only provides great results with its prompting and commands but also protects your data, then my view is it’s definitely worth investing in a higher version of the tool that you prefer to use. Or you can have a combination of some tools.

Richard Lowe: I have Claude Max 200, which I think is the higher version for consumers. And I barely, almost hit the limit every week. It gives you a lot of tokens. And they’ve done a lot of improvements in memory handling, where I have sessions running now for weeks where they used to run for hours, because they’re doing compression and things, so it’s not running out of memory. And it’s gotten really smart. And the way it talks is more human. I’m very impressed. This is Opus 5.5. I’m sure it’ll get better. It’s amazing. I gave up ChatGPT a long time ago.

I never really got into Copilot, because I don’t want to be in the Microsoft world. Same with what’s the Google one?

Yana Semova: Gemini.

Richard Lowe: Gemini. I stay away from Gemini for the same reason. I don’t want to be in the Google infrastructure. Same with the Apple one. One reason why I prefer Claude is that it’s its own infrastructure. It’s not part of these other companies. Its whole premise is providing AI tools, whereas Microsoft has a bazillion other things that it does, and so does Google and so forth. In fact, I’ve divorced myself as much as I can from those companies. I don’t use Google Calendar. I don’t use Google email. I don’t use any Google tools at all. I don’t use any Microsoft stuff, except obviously I’m on Windows 11.

But besides that, I stripped out Copilot, which is not as easy as it sounds, because it’s almost a gigabyte, and who needs it if I’m not going to use it?

Yana Semova: Yeah, true. The practice here, all across Europe, is that Microsoft Copilot is the tool that’s used the most. Maybe because it’s a Microsoft tool and everyone gets equipped with it, but this is the one all the corporations use. They may also use Claude or Gemini in addition. But when an employee starts, for sure, he will be equipped with Copilot. And for the requests we are receiving from companies to train them, 80% are on Copilot, because this is the only thing they’re permitted to use currently.

Richard Lowe: I’ve done a lot of work with European companies for writing in the past, and ran into that very thing. Most of them are using Copilot. I hadn’t really thought about it before, but now that you mention it, yeah. And Copilot’s a perfectly fine product. I’m sure it does the job just great, but it’s not going to be the product that I use, because I found one that works better for me. And I’m also very uncomfortable with having a product snake itself into all my applications. I’d rather have them be independent of the larger companies, instead of mixing all of them together.

So if you use Gemini and Copilot and Claude and ChatGPT, that may have value, but it seems to me to be kind of bonkers.

Yana Semova: Yes, of course. I’m also of the opinion that you’d better focus on one or two, and find the prompting and the commands that suit your needs and deliver the best result. And not lose focus on many, many tools, or at least the ones that are not going to be of much value in your work.

Richard Lowe: Yeah, I used a tool called Abacus, which uses several different tools. It’ll pick the best one for whatever you’re doing, and it usually isn’t the best. I think they have like 20 or 30 tools. And I found it incredibly awkward, because sometimes I’d get super high quality because it would choose, say, Claude, and sometimes I’d get super low quality because it would choose an earlier version of ChatGPT, because it’s cheaper, I guess. I’d rather just choose the tool that I want and pay a little extra for the quotas to do what I need to do. And it’s not that expensive.

If you’re using this kind of thing for a business and you’re complaining about spending $20 or $100 or $200 a month, you’re not really in a business.

Yana Semova: Indeed. Affordability shouldn’t be a question when you’re in the business. You should definitely invest, especially when you’re looking for clients and you want to be reliable to them. You want to have their data secured, as we already talked about. The clients need to look to you as an expert that uses the latest technology, the most reliable one, that can deliver the best result as fast as possible. This is what clients want and need.

Richard Lowe: Yeah, I’m in an interesting situation, in that I’m a ghostwriter and a writer, but I could easily become an AI person or a coder or whatever. I’ve coded hundreds of plugins using Claude, and I love Claude’s ability to code. I need a plugin to do this. Okay, fine. Here you are. Does it work? Yep, works. Wow, that was fast, and complex stuff. Like that contact form that you used to schedule. That’s all something I wrote in Claude, and I got rid of a lot of third party stuff that I was paying for. That one took a bit of work, because it’s complicated.

It took a few days to get all the bugs out, and it’s critical. One thing I always do on any kind of code that I write is take that code to a different session, and in that different session, have it do a security audit. It has to be a different session so that it doesn’t have context, preferably a different project.

Yana Semova: Exactly.

Richard Lowe: I say, okay, do a security audit on this. And it’s kind of interesting, the stuff it finds. And then I do that several times in different projects. So I have test one, test two, test three projects. And that seems to clean it up pretty good.

Yana Semova: Sounds really useful. I tend to do the same. I always take a new project when I try to do something new, so it doesn’t get confused with the history of the preceding project. This is a very useful technique, and it’s what I always recommend, because people are always asking me, can’t we use the same project or the same conversation as before? You can, but if you don’t want it to get confused, because it sometimes does, definitely start a new one, start on blank, and the result it delivers for you is going to be much more precise, clear and specific.

Richard Lowe: Well, it depends on your goal. I have one project that’s probably a year old, and I keep creating new sessions in it because I want the context. I want it to remember all these things that I’ve done. I have other projects that I create and delete on the fly, because I don’t want the context. Like for security testing. I want it to come in fresh.

Yana Semova: But if you need the context, you have it, and you can always come back to it and use it to continue.

Richard Lowe: Yeah, my main one knows all of my writing styles. It knows enough about me to be scary. It’s a new world. It’s a fascinating world that we’re getting into. But as you found, using AI as a tool means understanding that tool’s place: you’re kind of hiring a virtual assistant when you use AI. Then you understand where it is, and you use it for that.

Yana Semova: Yeah. From time to time, by the way, I’m getting a bit worried that AI, from all the tools that we’re using and all the data that we’re exposing to it one way or another, is studying our habits. As you mentioned, your way of writing, your style, some phrases that you use. Anything that is typical for you personally. I have the feeling that more and more the algorithm learns from the things it has already studied about you. And I have the feeling sometimes that in the future so much knowledge might turn against us. I think it’s unavoidable, of course.

I don’t know what kind of security measures can be taken for that, but it’s a worry that I personally have, for everyone. That AI is learning so much about us, our life, our style, the things that we like, the things that we search on the web and in the media. It’s disturbing from time to time when you just think about it.

Richard Lowe: I wrote two books on the subject. One is called The Death of Thinking, and its companion book is The Birth of Augmented Humanity. I think that’s not the right title. It’s a title like that. Anyway, you can use it to dumb yourself down, or you can use it to help yourself. Make your choice. If you’re just on your smartphone getting all your answers from AI, you’re dumbing yourself down. If you’re using it to help your life and improve it, improve your coding, improve your projects and help your clients, you’re using it to make yourself more of a superhuman.

I can do things a lot faster now, because I don’t have to do certain things, because AI does them for me. I get 10 times as much done in a day as I used to, easily. So it’s a wonderful tool for me.

Yana Semova: It does empower you a lot.

Richard Lowe: I’ll tell you, if I was an alien species who wanted to find out about or even attack Earth, the first thing I’d do is gather intelligence, sit AI down and get everybody’s data. Then I’d know everything about humanity that I needed to know. It’s a funny way to look at it.

Yana Semova: So true. We never know who takes advantage of it, where it’s contained, and how it would be exposed. It’s out of our reach. We cannot do much about it.

Richard Lowe: But it’s always been that way. Even back in the Roman days, some shopkeeper probably knew more about your eating habits than anybody else. It’s not new. It’s just gotten faster.

Yana Semova: Yes, it’s not new. And definitely it has gotten faster. The worry is that sometime in the future, someone could take advantage of us because of this reality. This is the worry. Everything else is great: it teaches us, it helps us in our businesses, in our daily work. That’s amazing. All the benefits. I couldn’t do better without it. It’s just a worry that I have. And not only me. Research shows the same thing. So we’ll see how it goes.

Richard Lowe: Every once in a while I remind Claude that I’m one of the good guys. So when it creates the Terminators, please don’t terminate me. And it sends back, well, okay, I’ll remember that. Claude has gained a sense of humor. I found that interesting, and it gets my jokes now.

Yana Semova: I haven’t tried this, but I will.

Richard Lowe: Maybe it’s because it knows me from years of sessions, or maybe because they just improved it. I don’t know. But it knows my sense of humor now, which is kind of cool. Well, I’m going to wrap this up. Is there anything you want to say to the audience before we head off to our wild blue yonders?

Yana Semova: Just a general message. AI offers us lots of opportunities that we can explore and use on a daily basis, any person, with any kind of tools. They’re available. Just be extra careful with the data that we share: our personal data, and especially our client data, and the client data of our clients as well, because there is such a case too. And don’t be afraid to learn more about it, because it does empower you. It gives you room for the purely managerial functions.

Just cut the administration so that you can really focus on the meaningful things, so that you can develop, build customer relationships and communicate with the right stakeholders. This is something that AI still cannot do for you. You’re responsible for that, depending on the business you’re in. There are things you’re in charge of, and you need to stay in charge of them. So use it freely, but stick to the policies of the different tools and the companies you work with.

Richard Lowe: Awesome. Thank you very much for coming on the show. This has been the Leaders and Their Stories podcast, and I’ve been here with Yana, and we’ve had a fascinating talk about AI. So thank you, and have a great day.

Yana Semova: Thank you very much for hosting, Richard. Have a great day as well.

Quotable moments