I built an engine that refused to issue directives on two real clients — here’s why that’s the point

Launching CDAI tomorrow. Before you upvote or ignore it, here’s what actually happened when I ran it on real data.

Two real businesses. Two different problems.

Client 1 — Google Ads data, ~$2,100 spend, 37 conversions. Engine ingested everything cleanly. Then the health monitor flagged the data as 4 months stale and set directive_safe = FALSE. No directives issued. A SCALE recommendation on stale data is worse than no recommendation at all.

Client 2 — Meta Ads, $2,414 spend, 7 campaigns ingested perfectly. 276 CRM leads — every single one had “Contact Import” as the source field. No campaign attribution. Engine refused to ingest them. directive_safe = FALSE. No directives issued.

Most analytics tools would have filled in the gaps, invented attribution, or issued recommendations with a footnote. CDAI did none of that.

The engine measures 7 cost layers that platforms don’t report — broker payouts, refunds, chargebacks, compliance costs, variable costs — and only issues SCALE/HOLD/CUT/PAUSE/FLAG directives when the data can actually support them.

Full case study with the SQL verification queries is at alloceraintelligence.com. Happy to answer anything before launch tomorrow.