Nizar Abi Zaher.

Nizar Abi Zaher was four weeks away from starting law school when he changed direction. He had spent the preceding months teaching himself engineering — not as a backup plan, but because the pull toward building had become stronger than the pull toward practising law. He chose not to go. That decision set the terms for everything that followed. His co-founder Adam Wazzan had always been a builder in the literal sense — hardware, software, anything that required making something that did not exist before. The two of them built Orchid, a personal AI assistant that became a Y Combinator Spring 2025 company, spending nine months solving for what nobody advertises about AI products: the messaging infrastructure, the memory, the tools, the hosting, and the endless ways things break after the demo works perfectly. They got it working. Then the companies started asking how.

The nine months building Orchid were not spent on the interesting parts. The AI models were already exceptional. The APIs were there. What required nine months was everything between a working demo and a reliable product: messaging infrastructure, memory management, tool integration, hosting, and the compounding failures that appear only when real users arrive. 'Nobody starts a company because they're passionate about debugging webhooks at 2 a.m.,' Nizar says. That gap — between what AI makes possible and what shipping actually requires — is exactly what Keiki is built to close. The infrastructure tax that Orchid cost them is the tax Keiki is designed to eliminate for every founder who comes after.

Keiki shipped on September 1, 2026. It lets companies build and deploy AI agents on iMessage, WhatsApp, Slack, SMS, Telegram, email, and other channels their customers already use — handling the infrastructure, so the founder can focus on the product and the customer. The team is five people: four focused on growth, one on engineering. That ratio is intentional, not accidental. The vision Nizar and Adam are working toward is the one that the biggest platform builders have already proven: WordPress made publishing accessible, Shopify made launching an online store accessible, and Keiki wants to do the same for agents. 'We believe every business will have an agent, just as businesses came to have websites,' Nizar says. 'Your next purchase might start with "hey."'

The clearest thing Nizar has said about building Keiki is about where value is shifting. 'We believe AI is making coding less of a bottleneck. Understanding customers, communicating clearly, and getting people to care are becoming even more valuable.' A 4:1 growth-to-engineering ratio at a five-person AI infrastructure company is not an oversight — it is a deliberate bet on that belief. The other lesson is structural: Orchid cost nine months to build because that infrastructure did not exist. Keiki's entire proposition is that the next founder should be able to start with their idea rather than spend the first year rebuilding the same foundation. The time between the idea and the product is where most momentum dies. Closing that gap is both the business and the point.