GPT Party 3.0. Business Optimization with AI: Rethinking Operational Efficiency GPT Party 3.0, the largest Russian-language networking event dedicated to artificial intelligence, took place in Silicon Valley on March 9-10. More than 450 people gathered at Plug and Play to meet leading experts, entrepreneurs and investors, discuss the latest trends in artificial intelligence and gain practical knowledge. The panel discussion “Optimizing Business with AI: Rethinking Operational Efficiency” discussed methods and strategies for implementing artificial intelligence in business processes, and shared personal experience of using AI in business. The speakers of the panel discussions were: Nick Spirin, Mikita Mikado, Max Sapozhnikov, Dmitry Karpov, Peter Mebert; the moderator of the discussion and the author of questions was Boris Kisko. What modern tools would you recommend to small and medium-sized businesses, as well as representatives of large companies? Petr Mebert: “Let’s look at this issue from a new angle. Existing methods are likely not the best choice for aspiring entrepreneurs. The English language has about one million words, not counting jargon, terms and abbreviations. Let’s assume that the average person has a vocabulary of several tens of thousands of words. This means that communication with a wide audience covers only a small part of the language, leaving out the specialized lexicon of narrow specialists. Therefore, for entrepreneurs, the most effective solutions are customizable ones, deeply integrated into their business processes. Why? Because large companies already control and develop their own language models designed to automate processes and increase efficiency. This means that large chunks of the market are likely to be captured by them. For entrepreneurs, the key is to find narrow niches where you can create a unique offer. In addition, by 2030, it is expected that about 80% of investments will be focused on creating specialized models to solve specific problems.” Dmitry Karpov: “Pay attention to customer search. Customer search today is a purely technical task thanks to advanced mechanisms. For example, do you want to find customers for flower stalls with positive reviews or are you interested in managers of a certain gender working in a certain region? All this has become possible thanks to algorithms. Just specify the parameters and start working.” Nikita Mikado: “To be honest, I can’t answer this question exactly. However, I can share how artificial intelligence was implemented in our company. We implemented a machine learning model for identifying interested customers, which was first used in the sales department, and then successfully implemented in the support service. I see artificial intelligence as the Internet: it gradually changes various industries and businesses. At different times, different industries adopt it differently. For example, in the field of customer service, its use has already become standard. Therefore, where and how to apply artificial intelligence depends on the specifics of your business. However, in any case, the area of customer support can be a great start.” Nikolay Spirin: “I will add to the answers above. Artificial intelligence is a universal technology that can be applied to almost any business process. However, it is important to understand that the success of its implementation is determined by specific criteria. Let’s consider the following framework: when we automate operations using AI, we can get one of three outcomes. The first is full automation, when AI performs operations without human intervention. The second is the role of a copy-palette, where AI helps make intermediate decisions, and a person is responsible for the final decision. And the third is cases when AI turns out to be practically useless due to the complexity of the task. It is important to note that over time, the distribution between these categories will change. As for specific cases of predictive AI application, its foundations were laid centuries ago, when statistical analysis methods were developed. These methods are used in many areas, such as credit scoring and property valuation. An interesting development in recent years is generative AI, in particular, large language models and generative models for creating images. This is a new opportunity to create unique content. For example, chatbots based on language models can become your business partner to discuss ideas and improve business processes. Such a dialogue can help you expand your thinking and improve business decisions.” Answers to questions from the audience My name is Olga, our business is focused on construction and real estate development. I am currently compiling a list of areas where I plan to implement artificial intelligence, and I would like to discuss it with you. I would be interested to hear your opinion on this topic. Max Sapozhnikov: “I would like to share with you one useful tool that I actively use when working with clients and colleagues within Google – these are AI meeting trackers. There are many of them available now, there are built-in ones in Google Meet and Zoom. One of my favorites is Sembly AI, which perfectly supports English and Russian languages and provides call recordings. Naturally, before recording a call, you need to make sure that your interlocutor agrees to this. In my opinion, this is an effective way to save time and effort on meeting summaries and notes along the way, since artificial intelligence takes on this task.” Dmitry Karpov: “I believe that it is important for business owners or SEOs to spend less time monitoring employee activity and partner onboarding efficiency. Delegating operational control to artificial intelligence can be a solution. Checking employee compliance with instructions is a simple task for AI with access to the systems. You say you have thousands of partners. Perhaps not so many employees manage these relationships. Data analysis can help understand how interactions with partners are going and how successful their onboarding is. There are many tools with access to your integration systems that allow you to compare real data with expected results at any scale. The main thing is to understand how to get data from your operational systems for analysis.” Can you share the tools that help you personally in your work? Dmitry Karpov: “Superhuman is a mail agent, quite expensive. However, it is a tool that saves a lot of time: it creates a summary of all correspondence, manages projects through a system integrated with the mail client. I was especially impressed by how easily I was able to find and view correspondence from three years ago. The AI provided me with a summary, indicating what exactly this correspondence was about, where we stopped, and what needs to be done next.” Petr Meler: “I think that the network is more important than the fish. You can find many tools on the Internet that really work. However, do not forget that despite the digital world, there is also a physical world. For example, for owners of traditional businesses, such as restaurants, it is important to remember that bread will always be baked in an oven. And that AI is not a magic wand. There is a lot of hype around AI now: there is a lot of noise around new technologies, but not all tools are useful. Just be a little more careful, hackers will probably appear in the very near future who will imitate tools. So be wise and careful, do not use every new tool that appears.” What is your experience with using AI in customer support? Is there any positive or negative experience? Mikita Mikado: “In early 2022, we implemented a couple of bots to answer customer queries. Since then, the number of clients has almost doubled, and the support team budget has grown by 11%. This was due to salary increases, but there was no growth in the number of employees. Although a complete transition to artificial intelligence for communication with clients is not planned in the foreseeable future, the first version of responses to customer queries is already generated by artificial intelligence. If the user is not satisfied with the answer, a live operator intervenes, who communicates via chat or email. We use Zendesk to integrate bots based on artificial intelligence.” How can you become a guide to AI for a conservative business? Nick Spirin: “I think it’s important to start with simple steps and demonstrate their effectiveness. You shouldn’t invest in huge platforms right away, which will be difficult and expensive to justify later. If you work in a large organization, start with small prototypes and gradually develop. The second useful idea that applies to all AI projects is the formation of a portfolio. It is important to approach strategically and take into account the risk and possible unsuccessful outcome. If you have a portfolio of projects of varying complexity and focus, you can create a diversified portfolio, where some projects will be successful, and others will be less successful. When it comes to a large organization and the need to change the company’s processes as a whole, it is important to have a top manager who can make difficult decisions and accelerate the project, overcoming barriers. Having such a person will allow you to overcome obstacles much faster.”
