GenAI: Cost Perspectives and Optimization Abilities - An Interview with Piotr Majchrzak, Co-CEO of Boldare

Matt: Piotr, thank you so much for joining us today. Maybe we can start with a very brief introduction of yourself. I’ll give you 20 seconds to tell us who you are, and then we’ll move on to the topic of AI.

Piotr: Hi Matt, thanks for hosting me here. Yeah, it’s always challenging to describe myself, but I think I would describe myself as someone who can translate how to solve customers’ problems and how to assist them with technology. So that’s my primary purpose.

Matt: Perfect! Today we’re focusing on AI, particularly how it can optimize costs and save time. AI has been a hot topic for quite some time. Perhaps it was more prominent before, but I’m curious about your perspective. Do you think companies need to immediately adopt AI, or is it better to wait and see how it evolves? What are your thoughts?

Piotr: Yeah, AI is definitely a hot topic. We might be hearing less about it lately because it’s discussed everywhere. The buzz around this technology is unusually high. If you’re a tech company or aspire to leverage technology, you really need to consider AI. While it might be hyped, it’s hard to predict which technology will revolutionize the industry. At the very least, you should explore AI to understand what it can offer your business.

Matt: And then, as a follow-up, do you think all businesses can benefit from AI, or is its use limited? Does a company need to be very specific or technology-focused, or is the application of AI broader?

Piotr: Yes, that’s not an easy question to answer, right? But I believe that every company is somewhat digital today, right? So, if any aspect of your customer journey involves technology - even something as basic as email or customer service conducted over the internet - you can benefit from AI. I think sooner or later, all businesses will see benefits from artificial intelligence.

Matt: And so, if you say that all businesses can benefit from AI, then a big question from customers, and one that I also have, is: where do I actually start? Let’s say a company recognizes AI as an important topic and wants to engage with it. This initiative might start with a simple decision, like ‘Let’s do something with AI,’ possibly prompted by the CEO or an enthusiastic team member. What comes next? Should you try to implement something on your own, or seek external support? I can imagine this is a common question among many customers, as it is for me personally, and it’s even more complex when considering an entire company.

Piotr: Yes, I think this is the question that not only CEOs are asking themselves, right? How should I start? There are definitely different paths you can take, but my advice would be to first start experimenting with it yourself. For instance, we know tools like ChatGPT and others that act like copilots in sales and software development, but also general tools like ChatGPT are good starting points, right? I think experimenting is a great strategy, at least initially, just to understand this technology. At first, it might be astonishing, but as you delve deeper, you begin to see its limitations and how it can be used to achieve your goals. This experimentation with specific tools is very important. But then I think the big question-or big opportunity-lies in how you can customize it to fit your products or processes. Every company is different, each with its unique selling points (USPs). You can derive this uniqueness from AI if you tailor it to your needs. And today, the support structures are pretty robust.

Matt: Okay, so the basic lesson is that playing around is very beneficial because you discover both the benefits and limitations, right? And I think essentially anyone can experiment with it. But then the next step is, how do you use it for your business? How do you customize it, essentially?

\ Piotr: Yeah, exactly. So, you know, you can find tools for marketing, sales, or other verticals that you can simply take off the shelf. But what really fascinates me, and us at Boldare, is what you can achieve when you integrate these tools with your own data, when you incorporate them as part of a solution or process. That’s when things really become amazing

Matt: Is there a specific example of how Boldare is currently using this approach, or not yet?

Piotr: Yeah, we experiment a lot in this playing field, trying to involve almost the entire company. It’s not just myself, we have various teams exploring different aspects. Obviously, developers are benefiting from it. For instance, we use different bots that interact with us through Slack. These bots provide answers on how to do things, who is responsible for what, and they even summarize information. We’re planning to add more features, like having the bots ask pertinent questions about project statuses or suggesting actions. So, I would say yes, it is helping us maintain software development processes, but it also generates ideas on how to improve various other processes.

Matt: And I know that Boldare is a very technology-driven company, so you clearly know what you’re doing. I assume there’s some trial and error with AI, but not all companies are as technologically focused or knowledgeable. It might be helpful for others to know where to start. They begin by experimenting and then move to customization, but what does that involve? How much does it actually cost? How is it done? I can imagine that many want to experiment and think about customization, but what would be the next step for me as a business owner looking to get started? Do you have any tips or ideas on that?

Piotr: Yes, I think it’s always good to seek out someone more experienced, like attending meetups or consulting companies, and really try to engage with them about how AI can help. When customers come to us with these questions, we don’t just talk - we organize workshops and prepare proof of concepts for them. I highly recommend this approach because you can talk for hours and show examples, but seeing something that operates on your own data or is integrated into your app is quite different. We really notice how it captures our customers’ attention and their excitement grows. Definitely try starting small with someone experienced.

To answer your second question about costs, I believe this is a pivotal aspect of the revolution we are in. The first part is that large language models are delivering spectacular results, almost like magic. But the second part is that this magic is now accessible to clients. Unlike a few years ago, when AI was only available through an API and solutions could take weeks or a month to deliver, it doesn’t cost a huge amount of money anymore. Especially since you can iterate on it. I would say a month of work can already provide tangible benefits and insights for your business.

Matt: Yeah, it seems super fast to me, right? To have something that actually works within a month is quite amazing. Of course, we should define what ‘working’ means, but to have a functioning proof of concept that you can use in some way within just a month is impressive.

Piotr: Yeah, exactly. That was quite amazing to me as well. Obviously, you can’t do everything, but you can certainly get a good idea of how it might be applied to your business. Maybe one more thing to consider is that we’re all learning AI, right? It’s really beneficial to see it as a journey, not just a one-time project or idea. Having an iterative approach is important, but staying interested in the subject and keeping up with its rapid growth is equally crucial. The dynamics of this space are incredibly fast-paced; there’s something new every week. You don’t know what will happen next month, so the sooner you engage and understand this new paradigm, the better for both you and the company.”

Matt: It reminds me of the early days of the Internet, which was quickly adopted by young people and transformed many businesses. Just today, I overheard a young child on the subway asking his mom if the voice announcing stations was an AI or a human. It struck me as a sign of how intuitively the next generation is integrating this technology into their lives.

Piotr: Yeah, definitely. It’s an interesting topic, right? How everything can now, or in the future, be artificial. But yes, I would compare it to the iPhone. The revolution sparked by the iPhone resonates with me more. It illustrates why, for many years, Google was the primary site for getting information. However, today, I see that many people are starting to use ChatGPT first to get answers. So Google is threatened by this shift. Maybe providing mere links to information isn’t enough anymore when AI can offer direct context and a better understanding of what users are searching for. Obviously, it’s still just technology, so we can use various sources and rely on our own intelligence. Yet, increasingly, it’s becoming the primary source; Google is no longer the first stop for many of us.

Matt: I’m curious about the financial aspect. Suppose we have a proof of concept that we’ve developed in one or two months and it’s working. How do we measure its success or ROI? I can imagine some companies might pursue this for prestige or to maintain a competitive edge. Do you have any practical examples of how companies or customers evaluate such projects.

Piotr: Yes, the great thing about modern AI is that you can start with a small project. For example, we’ve been helping our customers with content generation. Take e-commerce: you might need different product descriptions targeted at teenagers versus older people. Measuring ROI becomes straightforward by comparing the cost of a copywriter’s hour to the time AI takes to generate content. This way, you can clearly see the ROI on such initiatives, even in small-scale projects.

I always recommend starting with small tasks and scaling up rather than making a large initial investment.

Alternatively, you could explore how AI might optimize customer experiences, like how many hours you can save using AI for tasks such as voice responses, which you mentioned regarding subway announcements. This can be done quite affordably today.

It’s not necessarily about replacing people, but allowing them to focus on tasks that require more intelligence. Remember, while AI can serve as an assistant, it’s not genius-level; it’s less intelligent than humans and serves more as a basic helper.”

Matt: So, if AI isn’t too costly and you keep the project small and quick, you can measure the savings in hours or costs compared to traditional methods. Additionally, it frees up time for people to focus on tasks that require their real skills, increasing productivity. A question to consider: does working with AI make people happier? I can imagine someone who frequently performs manual tasks might be happier focusing on more meaningful work. Have you seen examples of this? Perhaps developers no longer need to handle certain tasks they used to.

Piotr: It’s a good question, but I think it depends on the person. Some are a bit scared, thinking, “I’ve been doing it this way for years.” Others see the potential to do things differently with this technology.

So, yes, I’ve seen happy faces and sometimes even shock. But I’ve also seen people who are a bit afraid, especially of misuse, because this is still just technology. We need to control it and operate it-it won’t work on its own. You have to be a skilled operator. Like any new aspect of life, if used incorrectly, it can lead to bad outcomes. This is what I think the skeptics are most vocal about-that the technology wasn’t used properly.

Matt: Good point. And that brings me to my last question. We all know the dramatic examples of computers gone rogue, like in ‘Terminator’-a movie I loved as a kid. It’s obviously very different in reality. But from a business perspective, should we be cautious? Is it risky to start using AI, or can businesses safely integrate it

Piotr: It’s a big question. Right. I think we still need some time to reach AGI (Artificial General Intelligence), so we’re not there yet. And I’m not the best person to judge when it might happen. Of course, I’m a little scared about where it could lead. But on the other hand, once you start using this technology, despite its many shortcomings, you see how marvelous it is. So, I think we really need some time, if ever, to be truly scared about it. I remain on the optimistic side of technology and AI. I don’t think the Terminator scenario is going to happen anytime soon.

Matt: I’m also optimistic. I don’t think the Terminator scenario will occur, and I believe we have many benefits to enjoy in the meantime. Thank you so much for joining today. We’re out of time. It was a pleasure talking to you about all things AI, the ROI, the financial aspects, the timeframes, and how it operates. I’ve certainly learned a lot, and I think our listeners have as well. Thank you so much again for sharing your insights and information with me, and with the audience as well.

Disclaimer: This interview is a transcription of the session available to watch on YouTube and is part of the ‘Around the Product Development in 25 Minutes’ webinar series. You can watch the previous episode here, or read its transcription here.

Our next event will take place on Monday at 3:00 PM CET with Dawid Żurowski, Product Designer at Boldare. It’s going to be a talk about design systems and the process of product development. See you there!