Catch

**Product and problem.** Catch is building an AI executive assistant for the large amount of coordination work that sits between a decision and its completion. The public product description covers inbox triage, email drafting, meeting preparation, scheduling, reminders, follow-ups, travel and restaurant bookings, and real phone calls. The important distinction is operational: Catch presents itself as able to carry out a task across the applications involved, not merely suggest a response in one application. A user can communicate through WhatsApp, email, Slack, iMessage, or a phone call, which is intended to reduce the friction of learning another productivity interface. This targets executives, founders, sales leaders, and other professionals whose scarce resource is attention but whose work is still mediated by assistants, calendars, inboxes, booking sites, and CRM records. The concrete problem is fragmented administrative execution: small tasks are individually easy but collectively create a continuous queue of context switching, delayed follow-up, and missed coordination.

**Core technology and operation.** Public material describes an agentic workflow that combines a user’s email, calendar, messaging channels, and, where connected, CRM context. Catch can infer the requested outcome, gather the relevant context, perform a sequence of actions, and return for approval or escalation when the situation requires judgment. Its sales-assistant description says the system learns from the user’s calendar and inbox, prior corrections, and operating preferences; it also describes a conservative progression in which Catch checks before acting early and becomes more autonomous as trust is established. That is a meaningful product architecture even though the company has not publicly disclosed its foundation-model providers, model-training strategy, retrieval design, tool-permission system, or evaluation results. The technical challenge is not simply language generation. It is reliable state management across changing calendars and conversations, authorization boundaries for external actions, disambiguation of people and commitments, and recovery when a booking, call, or message does not behave as expected. No public source supports a claim that Catch owns a proprietary foundation model or a novel model-training breakthrough.

**Market and go-to-market.** The initial market is high-value knowledge work in which a successful task can save an executive or commercial team meaningful time: meeting coordination, lead follow-up, travel logistics, scheduling across time zones, and preparation for important conversations. The multi-channel experience gives Catch a potential distribution advantage because it can enter through the user’s existing communication habits rather than asking every customer to adopt a new dashboard. The company appears to be pursuing a direct product-led and relationship-led motion toward executives and teams, with sales-oriented use cases providing a natural wedge into companies. Dealroom lists a monthly product price, but pricing, packaging, enterprise controls, and conversion metrics should be treated as public-profile information rather than audited company disclosures. There are no publicly named enterprise customers in the sources reviewed, so the commercial case currently rests on the breadth of the workflow, reported early usage, and the founders’ ability to turn an assistant service into a trusted software product.

**Traction, funding, and validation.** Entrée Capital announced that Catch emerged from stealth with a $5 million seed round co-led by Entrée and Pitango, with Seedcamp and Factorial Capital participating. The same investor account reported early activity of more than 12,000 meetings, 20,000 tasks, and 200,000 emails, while a founder LinkedIn post discussed more than 200 companies and thousands of bookings and meetings. These are useful signals of product usage, but they are company or investor-reported figures rather than independently audited customer, retention, or revenue data. Dealroom corroborates the September 2026 seed-stage financing and lists Entrée, Pitango, Seedcamp, and Factorial among the investors. The round is material enough to fund product hardening, integrations, and distribution, while the stealth-to-public transition provides third-party validation that the company has cleared an initial investor diligence process. It does not yet establish durable retention, gross margins, safety performance, or an enterprise procurement track record.

**Founders and team.** Catch was founded by Nir Sabato and Yoav Ramon. Catch’s official author page identifies Sabato as co-founder and CEO and describes the company’s executive-assistant product. Entrée’s funding announcement identifies Ramon as a former CTO of Nym Health and the first employee at HiAuto, giving the founding team a combination of startup operating experience and technical leadership. Public founder material also says the team spoke with more than 100 executive assistants while shaping the product. That discovery process is relevant because the product depends on understanding the difference between a plausible draft and an acceptable completed action. The public record does not provide a reliable employee count, detailed engineering roster, security leadership, or evidence of prior experience operating a high-assurance infrastructure product at scale. Those omissions matter because Catch’s promise depends as much on permissions, incident response, and customer trust as on the quality of its conversational interface.

**Competitive dynamics.** Catch competes with AI-native assistant platforms such as Lindy, scheduling and time-management products such as Motion and Reclaim, and premium communication products such as Superhuman that increasingly add AI workflow features. It also competes with human executive assistants, internal operations staff, and incumbent travel or expense platforms when the task involves bookings or coordination. Catch’s claimed edge is the combination of broad action execution, a familiar multi-channel interface, and a learning path from approval-heavy assistance toward more autonomous behavior. That combination can create a strong user experience if it consistently closes loops; it can also create a severe trust penalty when it sends the wrong message, books the wrong itinerary, or acts on stale context. No public evidence reviewed demonstrates a proprietary data moat, exclusive integration, differentiated model economics, or a defensible safety benchmark. The company therefore needs to show that execution quality and accumulated user preferences produce retention that general-purpose assistants and bundled productivity suites cannot easily reproduce.

**Defense, security, and resilience relevance.** Catch is not currently a demonstrated defense or security company. Its core product is a commercial executive and administrative assistant, its public materials do not name defense or critical-infrastructure customers, and no public source reviewed establishes a government deployment, security certification, hardened operating mode, or fielded autonomous capability. For that reason this record sets dual_use to false. There is nevertheless a limited strategic adjacency: a trustworthy action-taking layer could help coordinate logistics, schedules, procurement follow-up, incident communications, or high-tempo administrative workflows in organizations where personnel are overloaded. That adjacency is not enough to claim defense relevance, because those environments require stronger identity isolation, auditability, data residency, offline or degraded-mode behavior, classification controls, and demonstrable authorization guarantees than the public Catch materials describe. Its relevance to this database is therefore as an Israeli AI-infrastructure and trusted-agent observation, not as evidence of a current dual-use deployment.

**Stage, trajectory, and diligence risks.** Catch is an early-stage company: public sources place its launch in 2025 and its $5 million seed financing in September 2026, with encouraging but largely self-reported early usage. The next milestones should be measurable improvements in retained active users, successful task completion, safe automation rates, enterprise controls, and revenue quality rather than additional feature breadth alone. Key diligence risks are (1) action errors and liability when an agent changes an external commitment; (2) privacy, security, and permission failures across inboxes, calendars, chats, CRM systems, and phone calls; (3) dependence on third-party models, APIs, and communication platforms; (4) difficult unit economics if human review or expensive tool calls remain necessary; (5) rapid feature bundling by Microsoft, Google, Salesforce, OpenAI, and scheduling incumbents; and (6) weak retention if users enjoy demonstrations but do not delegate consequential work. The company has a credible seed-stage starting point, but its long-term value depends on proving that trust, reliability, and workflow depth compound faster than the market’s incumbent distribution advantages.