Why I Didn't Reach for AI Here

The founder instinct right now is to lead with AI, so let me be honest about a project where the strategic call was restraint. Shibolet & Co.'s hub was a digital community for a leading law firm's employees — benefits, courses, notices alongside events and interest-based connections. The temptation is obvious: surface AI-driven recommendations, auto-match people by interest, generate content. But the actual problem wasn't a shortage of intelligence; it was a shortage of warmth and reasons to show up. In a legal environment, trust and discretion are the culture, and a heavy-handed algorithm guessing who you should befriend reads as the opposite of warm. So my AI strategy was a posture, not a feature dump: the smartest move was earning engagement through human-centered design first, and treating any future automation — surfacing relevant events, nudging interest-based connections — as something that enhances a hub people already trust, never a substitute for the editorial and social judgment that makes a community feel human. Founders fund vision; the disciplined version of an AI strategy is knowing where intelligence genuinely serves the user and where it would erode the very trust the product depends on. Here, the strategy was to keep the foundation human and let AI earn its way in later.