Proof

How a Kenseda answer is proven

Every answer decomposes into claims, and every claim carries a receipt: the exact metadata — flow element, trigger, save-order step — that makes it true. Here is one finding from our demo org, receipt by receipt.

demo org · Loan_Routing__c

Two automations write the same field. One of them silently loses.

Q: What writes Loan_Routing__c.Score__c on save?
callsWRITES · step 8WRITES · step 14LoanRoutingTriggerapex · after updateLoanRoutingRouterapex · route()Loan_Routing_Score_Overrideflow · after saveScore__cLoan_Routing__c▲ last writer wins —trigger value discarded
LoanRoutingTrigger
An after-update Apex trigger, correctly guarded with a record-scoped Set<Id>. It calls LoanRoutingRouter.route(), which writes Score__c. In the Salesforce save order this runs at step 8.
Loan_Routing_Score_Override
An after-save record-triggered flow with a sane self-terminating entry filter. It also writes Score__c — at step 14, six steps later.
finding
Last writer wins. The flow overwrites the trigger's value on every save — no error, no warning, no debug log. The trigger's routing logic is effectively dead code. Neither automation is individually wrong; the org's combination is. This is exactly the class of behavior no one designed and no one can see.
verified
This isn't a static-analysis guess. A test deployed to a scratch org confirms it empirically: after a bulk update of 200 records, Score__c ends at the flow's value for every record, and the write log reads trigger-then-flow.

The comparison everyone asks for — a general-purpose AI assistant answering these same questions about the same org, side by side with Kenseda — is being recorded against this same demo org and will be published here. We won't paraphrase the incumbent's answers; you'll see them verbatim.

Book an audit