This works best when something concrete is stuck. A prototype that demos well but can't be trusted with real traffic. An automation chain that fails in ways nobody can explain. An agent idea that needs an operating model before it goes anywhere near production. Or a build you'd like a senior second opinion on before funding the next quarter of it. Bring the workflow or product area, your current tools, data sources and model providers, and a sense of what has already failed or feels fragile. You'll leave with a sharper diagnosis, a build-vs-buy recommendation, a first slice with the risks and decision points called out — and an honest answer on whether I'm the right person to help further.