Every "connect your AI to your data" pitch treats the connection as a detail you delegate downward. It isn't a detail — it's the entire risk surface. The AI engine is commoditizing; what endures is the governed access layer beneath it. This is a CIO's honest guide to the three doors into enterprise data — native connectors, third-party MCP, and web scraping — what each costs, where each breaks, who's accountable when it does, and why owning that layer deliberately is the real decision.

The way you link contacts to companies in your CDP seems like a technical detail — but it shapes everything from segmentation to account-based marketing. Most teams only discover they got it wrong when they are already scaling. This guide breaks down the two approaches, when each makes sense, and what enterprise-grade data modelling actually looks like in practice.

For most mid-market businesses, a Customer Data Platform promises unified customer data but delivers a very expensive place where data goes to sit. The real question is not which CDP to buy — it is whether you need a pure CDP at all, or a platform that can both unify and act on your data. This guide compares Composable CDP, Traditional CDP, and Good Bards across cost, team requirements, and execution capability.