Artificial intelligence (AI) agents are currently at the center of discussion in the global technology industry and are attracting billions of dollars in investment.
However, most of their use so far has focused on writing code and developing software, while many executives are still forced to handle heavy administrative workloads themselves or employ full-time staff to manage their calendars, emails and routine tasks.

Now, Israeli startup Catch is emerging from stealth and unveiling an autonomous solution designed to address precisely this headache.
The company announced the completion of a $5 million seed funding round, co-led by Entrée Capital and Pitango, with participation from Seedcamp and Factorial Capital. Catch, which employs six people at its Tel Aviv offices, was founded by Nir Sabato, formerly an investor at Entrée Capital and now the company’s CEO, and Yoav Ramon, who serves as chief technology officer.
An AI Agent That Learns About the Executive Within Minutes
The system developed by Catch offers an AI-based administrative assistant that operates completely autonomously. Unlike existing solutions on the market, the system does not require its users to learn a new tool, write complex prompts or define manual workflows.
Once connected to an organization’s systems, it builds a personal profile of the executive within minutes. It learns who the executive works with, identifies work patterns and preferences for scheduling meetings and travel, and begins acting proactively through familiar communication channels such as text messages, email, Slack or phone.

The system can independently handle meeting scheduling, emails and follow-ups, and even identify gaps in the schedule, such as suggesting a room booking at the executive’s usual hotel after detecting a flight reservation.
Co-founder and CEO Nir Sabato explains that executives do not want to build agents themselves or approve every small action performed by the assistant. He says the market has been waiting for a solution that will perform administrative work correctly and reliably 24 hours a day, and Catch’s system was built specifically to provide that trust and accuracy in real time.
Information Security and Splinter-Agent Architecture
One of the main challenges in deploying AI agents in organizations is maintaining privacy and information security. Catch emphasizes that these aspects were integrated into the product architecture from the initial planning stage, and that the company complies with the SOC 2 standard and undergoes independent external audits.

Yoav Ramon, co-founder and chief technology officer, explains that the system is based on “splinter agents” technology. Under this framework, each agent is exposed only to the information required to perform a single defined task. This approach makes it possible to maintain confidentiality and reduce risks while achieving rapid response times, low costs and high-quality performance. Ramon also noted that the data is anonymized and used to train open-source models without compromising users’ trust.










