One per scroll. Each one a real reason. None of them "AI-powered."
The schema is yours. The identifiers are yours. The exports are yours. The day you decide to leave, you walk out with the graph in a portable format and a query language anyone can read. We do not own the answer to "who is this customer." You do.
No black-box ML matching. Every time two records become one, the merge writes its own paper trail — which records, which fields, which rule, which confidence. Your privacy lead can ask "why" and we can answer in one query. That's not a feature; it's the price of admission for serious buyers.
Every modern identifier is a first-class citizen of this graph — universal IDs, hashed emails, mobile ad IDs, device fingerprints, the cookieless side. Not bolted on. Not a separate product line. The same graph that resolves your CRM resolves your bidstream. Most CDPs solve half the problem and partner for the rest.
The graph runs on an open-source engine you can read the source of. The query language is industry-standard. The schema is documented and portable. Your replacement vendor — if there ever is one — can ingest your graph in the format we shipped it in. No proprietary export, no migration tax.
No per-event pricing. No per-MTU pricing. No per-merge pricing. No surprise overage at the end of the quarter when your campaign actually worked. The price is the price; it scales with the value you take out of the graph, not with the volume that flows through the meter.