
IsraelTech
2.7K posts

IsraelTech
@IsraelTech
Unfiltered Israeli tech, startups, & VC news in your 𝕏 feed. Hosted by @YoelTIsrael & @RevitalMyer



A single wrong word in a live caption can put a broadcaster in violation of federal law. The FCC requires 99% caption accuracy on national broadcasts. Generic AI tools land around 90 to 95%. We talked to Yair Amsterdam, CEO of Verbit, about why that last few percent is the entire business. Amsterdam has been in the CEO seat since 2024, after two years running Verbit as COO and President, and before that spent close to a decade running operations for ProQuest and Ex Libris. He's spent most of his career inside companies where the gap between good enough and accurate enough decides who keeps the contract. His says you can't build one model for every conversation. You build a different one for an NBA game, a deposition, and a presidential debate, each prepped in advance with the names and terms that generic engines get wrong. The easy 90% is getting commoditized fast. The hard 9% is where the money is. Hosted by @YoelTIsrael

























Stav Levi Neumark joined Monday com when it was about 15 people. She told them in the interview they were already too big for her. She stayed anyway, built an internal tool called BigBrain that connected company data to go-to-market decisions, watched Monday scale to an IPO, and then left to build Alta. In this episode, Stav compares the current GTM infrastructure problem to on-premise computing before AWS. Every company is building their own room full of servers, manually stitching together data sources, channels, and signals that should be connected. It's expensive, it's slow, and it pulls focus away from the parts that actually require a human. @alta_revenue connects to 50+ data sources and runs three agents across outbound, inbound, upsell, and full-funnel visibility. The product ships with a services layer, because when they ran their tests, people didn't want to self-serve. They wanted someone to tell them what to do with it. She came in expecting customers to know what they needed. They didn't. They knew something was off, they were leaving performance on the table, but they couldn't name what was missing. So she built a company that does the diagnosing too. They closed their first $1M over the course of a year. Then closed another $1M in a single month. Hosted by @YoelTIsrael









Stav Levi Neumark joined Monday com when it was about 15 people. She told them in the interview they were already too big for her. She stayed anyway, built an internal tool called BigBrain that connected company data to go-to-market decisions, watched Monday scale to an IPO, and then left to build Alta. In this episode, Stav compares the current GTM infrastructure problem to on-premise computing before AWS. Every company is building their own room full of servers, manually stitching together data sources, channels, and signals that should be connected. It's expensive, it's slow, and it pulls focus away from the parts that actually require a human. @alta_revenue connects to 50+ data sources and runs three agents across outbound, inbound, upsell, and full-funnel visibility. The product ships with a services layer, because when they ran their tests, people didn't want to self-serve. They wanted someone to tell them what to do with it. She came in expecting customers to know what they needed. They didn't. They knew something was off, they were leaving performance on the table, but they couldn't name what was missing. So she built a company that does the diagnosing too. They closed their first $1M over the course of a year. Then closed another $1M in a single month. Hosted by @YoelTIsrael








