How an entrepreneur from Uzbekistan sold his company to Yandex and created an AI startup in the US

Akmal Paiziev talks about studying at Stanford, selling Express24 and his startup Numeo.ai in an interview.

Akmal Paiziev, Menlo Park, co-founder and CEO, LinkedIn

How it all started

I have been involved in technology entrepreneurship since 2005. I am originally from Uzbekistan and moved to the US in 2024. Before that, I managed to build several companies in Uzbekistan. The first company founded in 2005 was an e-commerce project Yarmarka.uz online store for electronics and home appliances. A year later, we launched Newmax Technology, a company that was engaged in creating auto systems for navigation and electronic maps. We are still the only company that provides digital maps with API used in our projects like MyTaxi.

In 2006 we launched Newmax Technology, which developed GPS navigation and fleet management systems for cars. In 2012, I co-founded Workly, a cloud-based time tracking system, and in 2015 we launched Express24, a food and grocery delivery service. In 2024, we sold Express24 to Yandex.

Last year I moved to the US to study at the Stanford Executive Program. During my studies, the last stages of negotiations with Yandex were underway, and the deal was almost closed. This got me thinking about the next step and creating something bigger, global, in the US. I decided to dedicate myself to the AI segment because it is a new opportunity with huge potential. AI is already changing everything and I wanted to be part of that wave.

Inspired by my work in the enterprise segment, I started developing ideas to create digital employees who will work alongside real-life colleagues. I believe that in the future, humans and digital employees will interact and work in the same office, even if those are not physically present. Working from Palo Alto, I and my coworkers in different cities meet with clients online. I believe that digital employees will become an integral part of business. I want to build a company that specializes in creating digital employees.

Launching a startup

Initially, when I was at Stanford, I wanted to create digital employees for franchise chains, as I understood their challenges based on my experience working with franchisees at Express24. However, after three months, the cofounders and I came to the conclusion that the logistics sector offered much more opportunity. Logistics in the US is a $1 trillion market, it is fragmented and dynamic, unlike the restaurant business, which is low margin and not as open to technology. Beyond that, there is no absolute leader in logistics.

Today, we are building digital employees for the logistics business, such as digital dispatchers, HR agents, recruiters, and compliance agents who help carriers in the US manage their operations efficiently.

Team

Our company has grown significantly and currently consists of 12 highly skilled professionals. We have hired strong engineers, which has allowed us to make significant progress.

Target Audience

Our key partners are the largest Load board in the US, DAT, valued at $30 billion, with whom we collaborate on the development of AI agents. We interact with DAT on an almost daily basis. We also work with TruckStop, the second largest Load board, as well as Uber Freight, partnering with three major market players.

We are currently expanding our partner network and have agreements with companies such as J.B. Hunt and C. H. Robinson. We also currently have over 30 customers testing our system in their companies.

Incubation programs

When I came to the States, we were going through Silkroad Innovation Hub’s Alchemist X accelerator program in conjunction with Alchemist Accelerator. That was an important milestone for us.

In terms of investment, while at Stanford, I informed my classmates about my intention to create a new startup in the US. As a result, several classmates supported the idea and invested $100 000 as initial support. In addition, my partners and I also invested our own money in the development of the project, and we currently continue to maintain its equity funding using a bootstrapping model.

Challenges

The hardest part in this market is establishing relationships with the big players, being a new company or startup. Without integration and the right context, it is difficult to create a product that will be in demand. AI agents are not just technology, but a whole process of integrating into the industry and creating a quality user experience. In order to offer the best UX, you need to have a strong team of engineers and designers, as the task is much more complex than just writing code.

Difficulties

The hardest part in this market is establishing relationships with the big players, being a new company or a startup. Without integration and the right context, it is difficult to create a product that will be in demand. AI agents are not just technology, but a whole process of integrating into the industry and creating a quality user experience. In order to offer the best UX, you need to have a strong team of engineers and designers, as the task is much more complex than just writing code. Working with the user is important here, because, as the ChatGPT example showed, new approaches to the interface, such as text chat, can dramatically change the perception of technology.

The second challenge is integration with large corporations. In the trucking industry, for example, there are many layers of protection, and large companies are often skeptical of new players and slow to sign agreements. For example, it took us more than four months to sign an agreement with one of the big players in cargo security. Opening the doors of large corporations takes time and effort, and the process of communicating with high levels of management can be long and complicated.

Stanford’s networking helps me in this process. Networking with former and current students, as well as professionals in different fields, becomes an important resource that speeds up the process of finding the necessary contacts and connections. Developing networking is a key component of success in our business.

Achievements

Launching a product and having customers willing to pay is a significant achievement for our company. Even at the stage when the product was still in development, our customers were already paying us between $1000 and $3000. This confirms a significant problem in the industry that we are committed to solving. The market is becoming increasingly difficult, especially with the imposition of tariffs by the US, resulting in lower freight and imports. As a result, those companies that can optimize their costs and improve efficiency will remain competitive in the future.

Our business focuses specifically on efficiency and optimization. The average company with 300 vehicles in the US generates about $100 million in revenue per year while remaining inefficient. Every 1% improvement brings those companies an additional million. We believe that our AI agents can significantly improve the efficiency of these companies, with benefits that can exceed $10 million per year.

Our clients understand the importance of this optimization and are willing to pay for innovation that can make their business more efficient. Importantly, we find the pain of the market and provide solutions that actually work. We have a successful track record of building businesses, the resources and financial ability to realize ambitious projects, and that is why clients want to work with us. We currently have over 40 clients who are supporting our project in the development phase and actively sponsoring it.

Plans

We have ambitious targets for the end of the year. We are currently seeing strong competition in the field of brokerage automation, while we are focusing on automating transportation operations. Our plans include processing thousands of orders per day using AI agents. In monetary terms, we are targeting processing volumes of $200 000 per day, which will reach $5-6 million per month.

Our goal is to create a system where AI agents book orders independently, providing a high level of automation and efficiency.