Skip to main content

A conversation with Yiftah Frechter on leadership and impact

Yiftah Frechter headshot
Image used with permission by copyright holder

Growing an organization requires more than just a monetary investment but also time and human resources. The costs and risks associated with business growth can feel overwhelming to an entrepreneur, yet the consequences of staying as you are can be worse than stagnating and can result in failure.

Tech entrepreneur Yiftah Frechter has experienced this challenge firsthand in many of his leadership roles, but he learned to use AI and open-source technologies to help lean teams achieve a large-scale impact. While hiring the right talent is essential to build a successful company, many organizations struggle by expanding their teams too quickly and facing the difficult decision of downsizing later. Frechter’s approach to companies like Origin Media highlights that it’s possible to scale while maintaining a streamlined team.

Recommended Videos

Q: You’ve had quite an extensive career. How do you approach innovation and building companies differently than others in the tech space?

Frechter: My approach to innovation has always been centered around efficiency and smart engineering choices. In today’s world, many companies rely on large cloud providers and expensive services, but I believe you can build massive-scale systems differently. By focusing on automation, AI, and Open Source innovation, we’ve built billion-request platforms with lean teams. It’s about using the right tools for the job, not just what’s popular or trendy.

Q: Your philosophy on engineering and scaling seems to challenge conventional wisdom. Can you expand on that?

Frechter: Absolutely. I believe that the cloud should be seen as a tool, not a necessity. Similarly, AI enhances automation, but automation itself has always been the key. I’ve built scalable platforms without the traditional infrastructure constraints by making pragmatic, cost-effective choices. For instance, at Origin Media, we process almost 20 billion requests per day, run real-time matching algorithms, and ensure high-performance video transcoding, all with a lean engineering team. It’s about crafting a highly optimized infrastructure that leverages intelligent automation at every level, and it’s been proven that small teams can outperform much larger organizations if you’re strategic.

Q: Let’s talk about Origin Media. What was the vision behind it, and how did you scale it so efficiently?

Frechter: Origin Media was built to transform the Connected TV (CTV) advertising space. By this, I’m referring to smart TVs and televisions that connect to the internet. Before they were found in every home, my co-founders, Stephen Strong and Fred Godfrey, and I saw the potential for advertisers to connect with audiences using live bidding and content-matching algorithms. However, the challenge became how to scale this idea while keeping costs down. We focused on building a small but talented team, used AI-driven automation, and crafted our infrastructure to handle billions of requests without the usual overhead associated with this task, which allowed us to create an efficient platform that outperformed many larger companies in terms of performance and cost-effectiveness.

Q: You also made an impact at companies like BounceX and Mediamind. What were some factors in those companies’ successes?

Frechter: At BounceX, a performance marketing SaaS company, I oversaw product development and strategy while leading a team of 150 engineers. We scaled conversion services to $100 million in annual revenue. In 2023, BounceX was named one of America’s fastest-growing SaaS companies, which I believe was due to us using AI to enhance customer engagement. At Mediamind, my team’s ideas were what helped the company grow, which resulted in its $414 million acquisition. The core lesson I took away from both companies was that smart use of technology and automation can lead to impressive growth even with smaller teams.

Q: It sounds like the key to your success is being able to scale with small teams and big ideas. Can you share a bit more about your other projects, like Springtide and Rello?

Frechter: Springtide focused on autism care and used technology to make therapies more effective. It raised $22 million in funding before being acquired in 2023. Similarly, with Rello, as mentioned earlier, is a new way of dealing with leasing properties. The program makes use of AI and automation to provide live screening and instant lease signing. This makes the entire experience of vetting a tenant and signing a lease faster, safer, and more efficient. We have specifically tailored this system to New York City’s market and regulations, and we’re already seeing great outcomes in terms of increased conversion rates and advanced fraud protection.

Yiftah Frechter’s ability to scale quickly with a focused engineering team has been a significant factor in the success of Origin Media. By utilizing the right tools and automation, he has demonstrated that small start-ups can compete with larger organizations. This approach has become a core principle of the company’s structure and overall philosophy. While continuing to lead Origin Media, Frechter is also exploring other sectors, including real estate, where he sees the potential to apply similar technology-driven optimization to transform the leasing process. Given his history of innovation, it is likely that he will continue to create impactful solutions with streamlined teams.

Digital Trends partners with external contributors. All contributor content is reviewed by the Digital Trends editorial staff.
Chris Gallagher
Chris Gallagher is a New York native with a business degree from Sacred Heart University, now thriving as a professional…
“AI Alone Cannot Save Disjointed Data Systems,” says Coginiti Founder

Artificial Intelligence has grabbed the attention of every business around the globe, driven by visions of automation and smarter decision-making. This has fueled a race to integrate AI into their operations. And yet, as per Gartner’s latest update, at least 30% of AI initiatives will be abandoned this year.
Why? “It’s not because AI doesn’t work,” states Rick Hall, founder of Coginiti. “It’s because most companies haven’t done the foundational work. They expect AI to clean up their mess, but it can’t. Companies need to clean up their own data before inviting AI in.”
 
Hall believes AI is still in the early phases of a decade-long transformation, one that will redefine nearly every business operation. But for that transformation to succeed, organizations must first tackle three persistent problems: poor alignment to business goals, bad data, and failed integration.
 
Companies often rush to implement AI without understanding how it will actually improve their business. “They start with the technology, not the outcome,” Hall states. “They can’t articulate what success looks like.” Without a clear hypothesis of what AI will do and the value it will drive, the initiative flounders.
 
Messy, redundant, and poorly labeled data is one of the most common obstacles to AI success. One of Coginiti’s clients, for example, had been using a cloud-based sales system for years. They wanted to deploy AI to optimize their sales pipeline. But their sales opportunity record had ballooned to 250 columns with four different fields just for purchase orders.
“Over time, processes change, people add fields, create nicknames, and now AI has no idea which one to use,” Hall explains. “You’re asking an algorithm to make sense of something even your sales team doesn’t fully understand.” Without clearly named, clean, and integrated data, no AI can deliver value.

Even when an AI model is successfully trained in a pilot program, many businesses struggle to scale it. That’s because they haven’t built clear bridges between their test environments and real-world operations. “You might have something that works in isolation, but integrating it into your existing systems, your workflows, your infrastructure, that’s where everything breaks down,” Hall says.
 
To solve these problems, Coginiti helps organizations build a digital twin, a virtual representation of how the business actually works. A digital twin isn’t just a model or a dashboard; it’s a semantic representation that maps your business processes, data, and systems in a structured, meaningful way.
 
“When you have a digital twin integrated into your system,” Hall says, “you can simulate change. You can say: ‘If we improve this process, we’ll see this benefit.’ It aligns your AI efforts to business value from the beginning.”
 
The digital twin acts as a translator between your messy, physical reality and a logical business model that AI can understand. It defines your data layer clearly and exposes it to AI tools in a way that is consistent and scalable.
 
A digital twin starts with understanding how your business actually runs. What are the key value chains? What are the metrics for improvement? This means data will be well-defined, consistently named, and logically organized. “The semantic model is the heart of a digital twin,” Hall says. “It’s what allows AI to interface with your business meaningfully.”
 
A true digital twin links clean, logical models to real-world data across systems: CRMs, ERPs, call centers, data warehouses, and more. A digital twin further pulls from across departments and platforms, ensuring a single, unified source of truth. With the integration in place, businesses can use AI agents to model changes, test new processes, and develop rapid pilots that can easily be deployed in production.
 
Ultimately, AI is not a magic bullet. It’s a powerful tool, but only as powerful as the foundation it rests on. “Think of AI as a new room in your house. If the blueprint is vague and the bricks are cracked, even the best architect can’t build a stable structure,” Hall compares. “The companies that invest in doing this the right way today will be tomorrow’s leaders.” Coginiti can help you with the transformation. 

Read more
Shaping Enterprise resilience through cybersecurity and software quality

The lines between software quality and cybersecurity are starting to blur in the online world. Companies can’t afford to think of them as separate issues anymore. A bug in the code can just as easily become an open door for hackers, and security gaps often trace back to how the software was built. According to Cybersecurity Ventures, cybercrime is expected to cost the world $10.5 trillion annually by 2025, highlighting the significant risks for companies, governments, and national defense systems. Cybersecurity today involves protecting private data, safeguarding critical infrastructure, maintaining public trust, and supporting the stability of economies. Gopinath Kathiresan is working to bridge these critical areas. As a seasoned quality engineering expert, he’s helping businesses build safer, stronger systems.

The Business Imperative: Quality and Security as Two Sides of the Same Coin
With more than 15 years working at the forefront of software quality and automation, Kathiresan understands that getting software ‘right’ is not just about avoiding bugs; it’s also about maintaining trust. Currently, each mistake incurs hundreds of thousands of dollars in the digital economy. Customers do not just want perfection—they expect it. When something goes wrong, they will not just lose users; they will lose their reputation and earnings as well.

Read more
Sunseeker’s X3 Plus brings high-precision mowing home for a budget-friendly price
robot vacuum face

Sunseeker recently introduced its newest consumer-grade product, the X3 Plus Wireless Robotic Mower, to improve how households handle lawn mowing.

Product release is timely for Sunseeker. Having recently entered the U.S. market and achieved commercial success with its professional X7 series, the company is now focusing again on the residential market. The X3 Plus is tailored to small to medium-sized lawns.

Read more