iCustomer

Are scope, spend, and authority enforced while the agent runs?

Policy gates are stated to constrain decisions at run time on named dimensions: "Budget and audience shifts inside policy gates you set", and agents are described as acting autonomously only inside guardrails the customer approves up front. That is a real claim of enforced scope and spend limits, which is more than most entities in this index. Not a 3: the enforcement mechanism, the ceilings themselves, and the behaviour when a gate is hit are all unpublished - enforcement is asserted, not demonstrated.

Is there a tamper-evident record of what it actually did?

Every action is stated to leave an auditable trace that the customer's team can "review, override, or roll back", and decision traces are described as remaining in the customer's own data cloud as a governed asset they own rather than a vendor copy. Customer-reviewable and customer-held, which is the substance of a 2. Not a 3: the capture makes no claim of tamper-evidence, hash-chaining or immutability - the words "tamper" and "immutab" do not appear anywhere in the document.

Our automated recount on Aug 14, 2026 counted enough evidenced fields for 1/3. It never reads what those fields say, so it does not overturn the score above, judged on Aug 18, 2026 — but where the two disagree it is a reason to read this dimension again. Why there are two scores

Are compliance obligations attached per engagement?

SOC 2 Type II is claimed, and a Trust Center is referenced for "current certifications and security documentation". Only a 1: the claim is self-declared on the vendor's own page with no report, auditor, certificate or scope statement published on the capture, and there is no per-engagement compliance schedule. No ISO 27001, HIPAA, EU AI Act or NIST AI RMF posture appears.

Does payment depend on a verified result?

Payment is a free tier with usage-based pricing as usage grows, and an enterprise fee that scales with audience under management - charged on volume, not on result. The gap is worth naming precisely because this vendor markets causal outcome measurement heavily ("measures the causal, incremental impact of every decision"): it measures outcomes but does not settle on them. No escrow, milestone release or clawback appears on the capture.

Can the deployment be insured, and is the customer indemnified?

No public evidence found. The captured document contains no occurrence of "insur", "indemn" or "liabilit" - no cover, no named carrier, no customer indemnity, no liability cap, despite the product taking autonomous budget-shifting actions in paid media channels. Absence is the finding.