Every other tool stops at a recommendation. This is the step where ASOLOOP doesn’t.

Ranked candidates at create-time — 3 to 4 in Classic mode, one in Agent mode — every option traceable to its source signals.

Ranked candidates with explicit reasoning. Three to four in Classic, one in Agent — the surface shape itself communicates evidence confidence.

Compare any past experiment to the current proposal. Hypothesis cards show prior results inline, weighted by recency and tier.

Every post-install signal carries a precision tier label. The reasoning layer surfaces the tier in why-text on every hypothesis card.

Interrogate any Agent-mode recommendation — see the signals, the weights, and the alternatives ranked underneath the top candidate.

Every AI-generated variant runs through a claim-safety contract. Un-sourced revenue claims and unsupported absolutes are rejected before submission.

Storefront auto-detected; baseline, review, and competitor analysis captured in the listing language; claim-safe generation per language (es-MX gets es-419, not es-ES); experiments scope to the detected locale. ~30 mapped storefronts — outside them, generation degrades safely with a review flag.

Teach ASOLOOP your house rules. Banned phrases, mandatory copy patterns, vertical-specific guardrails — applied across every generated variant.

Apple PPO + CPP and Google SLE submitted, monitored, and finalized end-to-end via the Playwright automation engine. CSLs authored and measured; send-for-review staged for you.

Recurring ExperimentSchedule entities with daily / weekly / biweekly / monthly / quarterly cadences. Each iteration spawns a child experiment; signal contributions accumulate.

Run CPP_paid, CSL_paid, or owned-routed experiments at the per-variant deterministic tier — install assignment is deterministic via ASA / UTM / OneLink identifiers.

Native PPO, Google SLE, and organic CSL experiments fire at the cohort-window tier, with explicit precision-tier disclosure so low-volume programs still produce honest signal.

See the precision class your experiment will fire at before you launch. Per-variant deterministic, modeled, or cohort-window approximation — never a surprise.

On by default, and it ships only the statistically judged winner: ASOLOOP writes the winning variant to the live listing — reversible within the 14-day undo window, with ask-first one toggle away.

Roll back an applied winner for 14 days. Suspend any experiment, revoke any signal within its 7-day window, force-switch back to Classic — every action carries an undo path; irreversible mistakes are not part of the design.

Revenue flows from your connected AppsFlyer account into the workspace revenue band on the Teams-tier stakeholder dashboard today. ARPU, install-to-paid rate, and lifespan feeding every confidence surface — translating each CVR posterior into a revenue range — is in progress.

A workspace-level revenue band computed from your connected AppsFlyer revenue, on the Teams-tier stakeholder dashboard today — a range bucket, no point estimates. Per-hypothesis translation of the Bayesian CVR posterior into a revenue range, every assumption disclosed inline, is in progress.

Every hypothesis, metadata change, variant, and outcome recorded per app. Indexed, timestamped, searchable across PPO/CPP cycles.

Every LLM-rendered surface keeps its deterministic evidence in view — trace any claim to its source experiment. No black-box scores. No silent confidence claims.

Manage your portfolio from one surface. Per-app evidence stays isolated; portfolio-level views show what is rolling and what is converging.

A dedicated role for regulated-claim review. Approvals tracked per artifact. Brand or legal stays in the loop without slowing the experimentation cycle.

Event-level audit trail covering every signal, decision, and platform mutation. Exportable for compliance review.

Apple PPO + CPP submitted, monitored, and finalized via the Playwright automation engine. ASA creative IDs supported as routed-traffic identifiers.

Native SLE experiments (Default Listing A/B, the Android default) handled end-to-end. CSLs authored and measured — send-for-review staged for you. UTM-tagged Install Referrer supported for routed CSL traffic.

Your AppsFlyer revenue feeds the workspace-level revenue band today; ARPU, install-to-paid rate, and lifespan feeding per-experiment confidence is in progress. Works across AppsFlyer Zero, Growth, and Enterprise plans.

Your existing AppTweak keyword feed becomes part of ASOLOOP's experiment evidence. Same subscription, more durable learning.

Apple Search Ads creative IDs are first-class routed-traffic identifiers — paid CPP experiments fire at the per-variant deterministic precision tier.

Adjust, Branch, and Firebase Analytics connectors are live alongside AppsFlyer — pick the MMP your stack already runs on. Revenue-input derivation is AppsFlyer-based today.

Programmatic access to experiments, signals, and confidence surfaces — for QBR pipelines, BI tools, and internal portals.

Bring your own Anthropic, OpenAI, or fal.ai keys. Required on Starter and Pro; optional on Teams alongside managed AI.

Single sign-on via SAML 2.0. Centralized provisioning, deprovisioning, and group-mapped role assignment.

Pin your workspace to US, EU, or APAC. Data plane and processing stay in-region; SLAs and audit obligations follow.

Audit log with export, SSO, and data residency available as a Pro-tier add-on for regulated mid-market portfolios.