The Future of AI in UX: Interview with Lazarev Agency Co-Founders

Here is the latest interview of the Creators team with guests Kirill Lazarev and Volodymyr Karishev, co-founders of Lazarev Agency, to share their expertise about the intersection of AI and UX, technology regulation, an innovative search engine, AI startup challenges, and the myth that AI steals jobs.

Some insights about our guest’s firm:

Lazarev Agency is a product design agency specializing in AI and boasting over 8 years of expertise in crafting digital designs for startups and more.

They have clinched three Webby Awards and 2x Honoree. Moreover, they’ve been nominated for Agency of the Year and Website of the Year at the CSS Design Awards. In 2022 alone, more than 80 of their projects have garnered international acclaim for their creative excellence.

Let’s initiate our adventure, venturing into the heart of this enlightening discourse to extract its essential insights. Our quest commences right now.

Q. Could you provide a brief introduction about your agency?

A (Kirill). Our user experience agency has been in operation for nearly a decade, and during that time, we’ve had the privilege of winning over 120 design awards. Our clients have also achieved remarkable success, raising more than half a billion dollars as a result of our design work. It’s important to note that we’re currently witnessing a significant shift in the UX landscape. This transformation isn’t solely driven by the advancements in AI technology, but also by the evolving nature of interfaces and user experiences.

What is the best city for AI development?

A (Volodymyr). The story is pretty simple. The reason we decided to make the move to San Francisco was because the AI industry in San Francisco is thriving, and most of the groundbreaking AI innovations are taking place within a very close proximity, within a range of just a few kilometers, which is quite unique in the history of the world. So when we talk about this, we’re not just referring to the entire United States; we’re specifically talking about California. The major players like Entropik, Bard, LLAMA, and OpenAI, are all headquartered in one city, and Silicon Valley. Silicon Valley is now more vibrant and influential than ever before.

So, currently, we see that San Francisco is a key driver of core innovations in AI. If you’re in the United States and you want to make strides in AI, whether it’s related to infrastructure, UX, UI, or developing your own custom LLaMA models, you should consider coming to San Francisco and participating in local events. The depth of knowledge and the insights you gain from the local experts are unparalleled, even when compared to other cities within the United States, not to mention cities in Europe or elsewhere around the world. Right now, San Francisco is at the epicenter of all the groundbreaking developments, and other places are essentially following the trends that originate in this region.

A (Kirill). I believe it’s all about the connections that develop among the individuals who attend the events. If we were to compare the events in New York during Tech Week and a typical event in San Francisco, you’d notice a stark contrast in the caliber of speakers and attendees. In New York, you’ll find more individuals who are theoretically involved in building something and are eager to learn and explore the potential of AI in the future. In contrast, in San Francisco, you encounter people who are actively engaged with organizations like the OpenAI Foundation and the Berkeley AI Research Labs. These individuals attend with not just a casual interest but to apply their knowledge to areas where UI is poised to disrupt the world.

So, without a doubt, if you’re contemplating delving deeper into the realm of AI, there is no other place in the world quite like San Francisco.

Q. How did you come to AI? 

A (Kirill). Historically, the journey has been quite fascinating for us. We’ve been collaborating with an AI startup since 2015, essentially from the time of their inception. Our role has primarily involved assisting them with product design for their core product, which is designed to analyze data for hedge funds through the use of AI. Before venturing into this field, they operated as a stock prediction platform, and we worked on devising user experiences that could predict stock market prices based on data sourced from news and media. Over time, we continued to collaborate with them and eventually helped them transition into a different type of product. Presently, we are actively involved in creating various exciting solutions for them under the name Answer.

The majority of our clients in the past belonged to the realm of companies looking to develop intricate solutions, often within the B2B sector. Our work spanned various domains, ranging from factory automation platforms, project management tools, risk assessment systems, and marketing prediction startups, to automation solutions for advertising. We had a wide array of interests and were constantly pondering what the most captivating and fulfilling endeavors were. However, when we observed the immense potential and the evolving landscape of AI-driven experiences, we made a strategic decision to focus our efforts exclusively on one industry. This decision was rooted in the realization that this field is undergoing substantial changes and introducing novel paradigms, which greatly appealed to us. Our primary motivation is to guide users towards adhering to sound user experience principles and make their interaction with products as seamless as possible. Leveraging AI technology has allowed us to achieve unprecedented levels of success in this endeavor, which is why we chose to steer our course in this direction.

A (Volodymyr). You know, the story behind our collaboration with Kirill is quite interesting. While Kirill was involved in the classical Web 2 era with his own company, I was always immersed in the Web 3 realm, working on global ledger technologies alongside our in-house development team. It became apparent to us that the utilization of AI tools was on the rise in the Web 3 space as well. This realization prompted us to embark on a journey of research, and we recognized a tremendous opportunity when we met Kirill to join forces and collaborate on our shared understanding.

I had been working in the deep tech domain, and together, we discerned that AI was set to become a transformative force across various industries. In essence, AI was poised to lead a so-called force revolution, akin to the transitions from steam power to the internet, and subsequently from the internet to mobile technology. We saw AI as the next significant leap, a backbone technology that would find application in virtually every sector.

Some technologies, like AR, VR, blockchain, and a few others, have niche applications in specific industries, reshaping those sectors, but with limited reach. In contrast, you have technologies like the internet and mobile, as well as applications, that have a profound impact on all industries. Just as we no longer say a company is “powered by the internet” today, we foresaw that in a few years, saying that a company leverages AI would be a given. AI tools and AI-driven processes would become integral to every company.

This vision is already materializing, as evidenced by recent news from Bloomberg that PwC is actively embracing AI to reduce costs and enhance efficiency. In the current market landscape, the focus is on maximizing efficiency, as the cost of capital has risen. AI plays a pivotal role in achieving this goal, optimizing companies and making their workforce more effective.

We’ve observed this transformation across various industries, from cinematography, exemplified by companies like Respeacher, a Ukrainian startup actively used by Netflix and Disney, to advertising and logistics firms employing AI components to streamline their warehouse operations. Different industries share the common theme of utilizing AI to boost overall workforce effectiveness. This core distinction sets AI apart from previous technological revolutions we’ve witnessed in recent years.

Q. What is actually user experience for AI?

The essence of user experience in AI lies in the idea that products built upon AI should prioritize hyper-personalization. Embedding hyper-personalization is one of the most straightforward applications of AI in various products. This should be a fundamental consideration when using AI, not only to achieve specific outcomes within a product but also to create hyper-personalized experiences within the product itself.

In the pre-AI era, designers were responsible for shaping how interfaces should appear. They dictated the layout, dashboard design, and purchase flows, all based on their vision of what would work. However, we are now transitioning into a new paradigm called intent-based design. Intent-based design entails adapting the interface according to the user’s behavior. The more personalized the behavior, the more the interface should cater to it. This concept comprises two main elements.

First, for the product, it necessitates a deeper understanding of your users. This, in turn, requires amplifying user research to comprehend what problems your users are attempting to solve within the product. These insights should be incorporated into the core features of your product. Instead of a more random approach, where you experiment to find something that works, you should focus on building features that align with your users’ needs.

The second aspect involves the complete transformation of the interface to create hyper-personalized experiences. Take a chat window as an example. A crucial criterion for AI-based products is the retention rate. What is the ideal retention rate for such a product? Are users meant to spend hours on the platform, or should they be able to resolve their needs within a few interactions? Well-designed experiences can reduce the time needed to reach the point of problem resolution by introducing tweaks that were not feasible before.

For instance, consider rephrasing questions and prompts provided by the user. By doing this, you can accelerate and guide the user through the flow more effectively. Let’s say someone is looking for a coffee place in New York to meet friends. The basic approach would be to provide a list of the top coffee places obtained, for example, from TripAdvisor. A more sophisticated approach involves rephrasing the question to gather more specific information. Do they want a coffee place with a great view, one close to local bars, or one in a historical area? These suggestions help satisfy the user’s needs more precisely and steer them toward what we, as the product, want them to take a look at. This approach can be applied to more complex use cases, making market research for startups more efficient, for example.

These flows change how users interact with a platform and how they use the product. Another fascinating example of AI-driven user experience is creating multi-modal chats for groups of people who collaborate on marketing strategy and execution for a brand. AI can amalgamate different thoughts and opinions to provide a cohesive outcome. These nuances need to be incorporated into various aspects of the product to ensure its effectiveness.

A (Volodymyr). I’d like to add that the experience of interacting with AI interfaces and applications is crucial. It’s one of the central aspects that AI companies are striving to enhance. If you examine OpenAI’s strategy, they are specifically targeting niche use cases. Sam Altman emphasized this approach during a recent meeting in San Francisco, where he encouraged others to build niche use cases because OpenAI has a significant lead in AI infrastructure. OpenAI’s partnerships with companies like Salesforce and PwC, which aren’t traditional tech companies, demonstrate this focus.

Many AI companies, including Bard, LLaMA, Entropik, and giants like NVIDIA, are not just pursuing AI for AI’s sake. They are looking to develop niche-focused AI applications that serve specific domains. In this context, UX and UI design play a pivotal role because they shape the actual user experience. The current stage of the AI industry has investors asking a crucial question: “Who are your customers?” This question can only be answered effectively with a well-crafted UX/UI design. This is where our focus lies — assisting LLM-based projects in finding their niche and connecting with users while emphasizing that AI is not just about technology; it’s about the overall user experience.

Q. Is the AI industry currently dealing with the challenge of addressing the fears of the general public who will be using AI?

A (Volodymyr). That’s a great point. I heard a compelling perspective during my time in Silicon Valley, and it resonates with me deeply. The idea is that AI will never steal your job; the person who uses AI effectively will. AI is merely a tool, an instrument. Just like Excel didn’t eliminate accountants but made them more efficient, AI is meant to enhance human capabilities rather than replace them.

When Excel was first introduced, there were concerns that it would replace accountants by automating tasks with formulas. However, here we are in 2023, and accountants are not only thriving but are using Excel to become more effective. The key takeaway is that AI has the potential to significantly empower the top 20% of professionals, making them even more productive in their respective fields. It may indeed eliminate jobs in the lower 20%, often referred to as “monkey jobs” – routine, repetitive tasks that can be easily automated.

The primary focus of AI automation is on these types of jobs, which are ripe for transformation. However, the overarching goal is to augment the capabilities of high-tier professionals, helping them excel in their areas of expertise. In essence, AI is a tool that has the potential to make us all more effective in our work. So, yes, that’s the essence of it.

Q. Is AI regulation necessary?

A (Kirill). We’re in a space where we’re essentially rewriting the rules and facing new challenges. Many of the laws and guidelines that we’ve invented by ourselves, we actually cancel right now.

For instance, one core principle was not to grant AI access to the internet, and another was to prevent AI from causing harm to people. The concern has been that if AI can harm people, it could potentially seize control and make humans subservient. However, some sexually explicit LLMs, including those in BDSM domains, are already pushing the boundaries of these laws. This highlights the fact that we are breaking the rules we have set ourselves, and it’s a stark reminder of the complexities we face.

Looking to the future, Volodymyr’s perspective is valid – it’s not about fearing AI but taking thoughtful action and being conscious of what data we upload and share. We should be particularly cautious about the integration of AI in large-scale financial institutions, as seen with companies like BlackRock, which can significantly influence or predict market values.

The future of AI raises the concern that major players could engage in market manipulation and use AI components to vie for control. Proper regulations are essential to ensure oversight in this rapidly changing landscape. One interesting idea is the notion that the biggest risk emerges when one AI can collaborate with another AI in a language beyond our comprehension. Regulations should provide the means to shut down such systems when necessary.

It’s a bit like the “Wild West” right now, with incredible opportunities for those who are curious and well-informed. However, it also comes with its share of edge cases and potential pitfalls. The best course of action is to stay informed, collaborate with individuals who possess deep knowledge in this field, and use AI for the greater good while advocating for responsible AI development and usage.

A (Volodymyr). These are indeed turbulent times, and AI is poised to play a significant role in shaping the coming years. We are currently in a phase that will determine the pace of the entire century ahead. The challenge with regulating AI is that if one country sets limitations, it doesn’t necessarily mean that other countries will follow suit. This is a reality we are witnessing.

For example, the militarization of AI is already happening, and it’s driven by the quest for effectiveness. It’s a force that is difficult to halt once it’s set in motion. Unfortunately, we are likely to witness years where the established rules of the previous world are at least challenged, if not broken.

As Ukrainians, we’ve experienced these challenges on our own soil, underscoring the global impact of these changes. AI will significantly influence various industries, and while the hope is for international consensus on AI regulations within 20 to 25 years, the reality is that such a consensus is unlikely to materialize in the current environment. This means that there will be limited global regulations on AI.

Moreover, I agree that we should appreciate the fact that Silicon Valley is at the forefront of AI innovation. While there are concerns, it’s important to recognize that it’s not the worst place for AI development. Many people with high moral standards are contributing to the industry, and figures like Sam Altman are among those leading the way. So, there are reasons to be cautiously optimistic about the development and impact of AI, even in the midst of these uncertainties.

Q. Can you explain your worl on  reinventing the search engine and how it operates?

When we think about the search itself, and when we think about the information that appears on the screens of the users when they’re actually using the search, there is a lot of the components out there.

Starting with hyper-personalization, the search engine already possesses a deep understanding of who you are and what you’re looking for. It can anticipate the type of search results you need. However, the future of search may involve more visual elements. For instance, the search engine could present information in the form of statistics, graphs, or visualizations directly on the search page. It might extract relevant data from specific articles or pages and display it visually, making it more accessible and user-friendly.

Additionally, search engines could provide quotes or snippets from trusted sources like Wikipedia to offer context and information that directly pertains to your query.

Another intriguing aspect is the potential for more interactive conversations with the search engine. Instead of merely refining queries with nuanced and complete sentences, what if the search engine could engage in a brief chat to better understand your needs and further refine the search results? This interactive approach could streamline the search process and help users find what they’re looking for more efficiently.

The key theme here is the focus on hyper-personalization and delivering a highly customized experience for users. Users don’t want to see ads or content that is irrelevant to them. They want search results, ads, and recommendations that are 100% personalized and directly applicable to their needs and interests. The more customized and personalized the search experience becomes, the better and more efficient it will be for users.

Q. You’re not allowed to say which project, it’s in DNA, right?

A (Kirill). Yes. It was backed by one of the biggest  browser in the industry. And we were exploring how to change the search experience in general.

Q. Why do startups, including AI ones, fail due to poor UX? Can you give some examples of collapsed projects who didn’t think through the UX well?

A (Kirill). Product-market fit is, without a doubt, a critical determinant of a startup’s success. If your product doesn’t address a genuine market need or resonate with users, it’s challenging to gain traction. However, user experience plays a pivotal role in achieving and maintaining that product-market fit.

User experience encompasses not just the interface but the entire journey a user takes with your product. It’s the bridge that connects your product to the market. A great user experience can make it easier for users to discover the value your product offers and become loyal customers. It can also help you differentiate your offering in a crowded marketplace.

User research is a valuable tool for startups to understand their target audience’s needs, pain points, and preferences. It provides the insights needed to tailor the product to the specific requirements of different user segments. For example, tailoring the value proposition and user experience based on regional differences is a perfect example of this approach.

Many startups, not just in the AI space, indeed focus too much on the technology itself rather than the user’s problem they are solving. They become enamored with the capabilities of the technology and lose sight of the fact that users ultimately care about how the product addresses their unique needs and challenges.

In the AI sector, where the underlying technology can be complex, it’s even more crucial to make the user experience as seamless and problem-solving as possible. Users don’t care about the underlying AI model; they care about how it helps them in their specific use case.

Lastly, talking about the rush to create wrappers around AI models, like GPT-4. This highlights the importance of focusing on solving real problems and providing a valuable user experience rather than simply showcasing the technology. Ultimately, the success of an AI startup is determined by how effectively it translates technology into user value.

Q. Any further advice for building AI startups in today’s niche?

A (Volodymyr). I’d like to share a brief quote that reflects our perspective on what we observe in San Francisco and offer it as a suggestion to our current audience. Folks, I encourage you to contemplate the application of AI in a specific field. Please refrain from engaging in AI endeavors purely for the sake of the technology, however enticing it may seem. As I mentioned earlier, trying to catch the Entropik, Bard, and LLaMA could prove quite challenging. However, many companies are genuinely interested in acquiring your services if you have well-defined use cases. And this viewpoint is not solely my own; it aligns with the beliefs of the Head of Ecosystem at NVIDIA, who possesses a wealth of knowledge surpassing mine. She has stated that NVIDIA has established dedicated divisions for various specific niches and actively seeks startups that cater to these niches with relevant use cases.

Moreover, it’s essential to recognize that certain niches are more formidable to enter than others. For instance, if you aspire to implement AI in the biotech sector, it’s crucial to understand that you’ll encounter a myriad of regulatory challenges, and the FDA may not necessarily be your ally. Indeed, there are niches where AI implementation is relatively more straightforward due to fewer regulatory hurdles, while other niches are considerably more challenging. It’s imperative to evaluate not only the potential of an exciting AI use case but also the intricacies of the niche itself. Some niches may be larger, others smaller, some easier to penetrate, and others heavily regulated, such as biotech, genetics technology, and any field connected to personal biodata.

Many startups are exploring these areas, but it’s worth noting that we’ve already witnessed projects encountering setbacks in highly regulated domains due to the labyrinthine bureaucracy they must navigate. In contrast, major investors understand that giants like Pfizer are capable of investing substantial resources in these areas and possess a deep understanding of regulatory compliance, including FDA requirements.

In summary, I would advise you to carefully select a niche and contemplate how AI can improve this niche. This is my suggestion personally.

The entire video is available for viewing below 👇