ReturnsIQ — User Guide

The complete handbook for AI-powered returns on your Shopify store.

This guide walks you through every feature in ReturnsIQ — what it does, when to use it, and how to get the most out of it. No technical background required. If you can run a Shopify store, you can run ReturnsIQ.

What ReturnsIQ Does
Getting Started
The Dashboard
The Branded Return Portal (Your Customer's Experience)
Working a Return: Approve, Reject, Refund, Exchange
Refunds & Store Credit
Exchanges
Return Reasons, Conditions & the Return Lifecycle
Return Policies & the AI Policy Builder
AI Fraud Scoring
Photo Authenticity
Predictive Return-Risk (At-Risk Orders)
Buyer-Remorse Interventions
The AI Decision Engine
Save the Sale
Carrier Labels & Return Shipping
Disposition & Warehouse Capture
Carrier Claims
Cross-Store Fraud Network
The AI Copilot
Customers
Analytics
AI Intelligence: Reason Clusters & Forecasts
Sustainability
The Audit Trail
Settings
Plans & Billing
Connections (Your Other Kernvel Apps)
FAQ
Glossary
Getting Help

What ReturnsIQ Does

ReturnsIQ runs your entire returns operation — from the moment a customer wants to send something back to the moment you've refunded them, swapped the item, or put it back on the shelf. It gives shoppers a polished, self-service return portal on your storefront, and it gives you a control room in your Shopify admin where every return arrives pre-scored, pre-sorted, and ready to action in one click.

Most returns apps stop at "here's a portal and a rules list." ReturnsIQ goes further: it actually protects your margin. Return fraud, wardrobing (wear-it-once-and-return), and serial returners drain profit you never see on a dashboard. ReturnsIQ catches those patterns before the money's gone.

The "AI" in ReturnsIQ is not a marketing label — every AI claim below maps to a real capability working on your returns:

Scores every return for fraud risk the moment it lands, with a plain-English breakdown of why — never a black box.
Checks return photos for authenticity — whether an image was reused from a past claim, edited, looks AI-generated, or doesn't match the damage the customer described.
Predicts which orders will come back before they ship — like a customer who bought three sizes of the same item — so you can step in early.
Recommends the smartest outcome for each approved return — refund, store credit, or exchange — based on how valuable the customer is, what's in stock, and your own rules.
Tries to save the sale when a customer starts a return, offering an exchange or a sweetened store-credit deal before the money leaves your store.
Generates real prepaid return labels, issues real Shopify refunds and gift cards, and creates real exchange draft orders — not placeholders.

You stay in control of everything. Every AI output is a recommendation with a threshold you set. Turn any feature on or off in Settings.

Who is ReturnsIQ for?

Apparel, footwear, and accessories stores where fit drives a lot of returns and wardrobing is a real cost.
Any store that processes enough returns that doing them by hand has become a job — and where fraud and "it arrived damaged" claims are eating margin.
Merchants who've outgrown a basic return portal and want fraud scoring, photo checks, and smart outcomes — without an enterprise invoice.

Getting Started

1. Install the App

Install ReturnsIQ from the Shopify App Store. The app is free to start — no credit card on the free plan, and you only pay if you upgrade. After installing, you're taken into the ReturnsIQ onboarding wizard inside your Shopify admin.

2. The Setup Wizard (~4 minutes)

ReturnsIQ walks you through six quick steps. Everything you set here can be changed later in Settings — nothing is permanent.

Welcome — A quick tour of what the AI does for you: damage forensics, the Save-the-Sale agent, the disposition decision agent, pre-purchase fit prediction, and the cross-merchant fraud network.
Store analysis — ReturnsIQ looks at your existing return history (if you have any) and sketches sensible starting settings, plus a few recommended policies.
AI configuration — Set your fraud thresholds (when to flag a return as high-risk, when to auto-reject), turn on auto-approve for clean returns, and set your Save-the-Sale and returnless-refund limits.
Customer experience — Pick your portal accent color, write your own welcome copy, and choose which storefront widgets to show (the VIP return-window badge and the "CO2 saved" footer).
Integrations — Choose your preferred carrier for return labels (Shippo, EasyPost, or manual for now) and, if you're eligible, opt into the cross-merchant fraud network.
Review & launch — Confirm what's switched on, then go live.

3. Put the Return Portal on Your Storefront (the one manual step)

This is the single step ReturnsIQ can't do for you, and it's the one thing your customers need:

In Shopify, go to Online Store → Pages → Add page and name it Returns. (The wizard has a Create Returns page button that does this for you and creates /pages/returns.)
Open the theme editor, choose Add section → Apps → "ReturnsIQ Portal", and drop it onto your Returns page.
In Settings → Branding, set the Returns page path (e.g. /pages/returns) so the one-click links in your customer emails open the right page.

ReturnsIQ detects the portal automatically the moment a shopper loads the page — you'll see a green "Portal detected" badge. Until it's placed, customers have nowhere to start a return.

4. (Optional) Connect a Carrier & Set Your Return Address

If you want ReturnsIQ to generate prepaid return labels, add your own Shippo or EasyPost API key in Settings → Carriers, and fill in your return address in Settings → Return address (this is the destination printed on labels). You can skip this and handle labels manually.

5. File a Test Return

Once the portal's live, file a test return through it yourself to watch the whole flow — fraud score, photo check, AI recommendation, and the one-click refund — end to end.

The Dashboard

The dashboard is your home base. It opens with the numbers that matter and a live view of what needs your attention right now.

Key numbers — Open returns, returns in the last 30 days, fraud caught (30 days), and dollars saved (30 days).
Setup checklist — For new stores: place the portal block, create your first auto-approval policy, connect a carrier, and file a test return.
Your AI engine — A grid of every AI capability (fraud scoring, photo authenticity, the decision engine, Save-the-Sale, predictive return-risk, buyer-remorse, reason clustering, the predictive returns forecast, the policy AI translator, fit prediction, and the cross-merchant network) showing which are active on your plan, each linking to its own page.
Today's attention — The highest-signal items that need action now, ranked by severity.
AI insights — Short, automatically composed summaries of what changed in your returns and what to do about it, refreshed on a regular schedule. Hit Refresh insights to recompute.
Sustainability impact — Your 30-day CO2 saved, items kept in circulation, and a breakdown of where returned items went.

The Branded Return Portal (Your Customer's Experience)

The portal is what your shoppers use to start and track a return. It lives right on your storefront — no separate website to redirect to, works with every Shopify theme, and carries your branding (accent color, logo, and your own welcome copy). A clear self-service portal is the single biggest thing you can do to cut "where's my refund?" support tickets.

Here's the journey from your customer's side:

Look up the order — The customer enters their order number and email. ReturnsIQ pulls the order from Shopify and shows the items that are still eligible to return. If an item is already part of an in-progress return, the portal explains that and won't let them request it twice.
Pick items and a reason — They choose what to send back, the quantity, the item's condition, and a return reason (e.g. "doesn't fit," "not as described," "defective"). They can add a free-text note explaining what happened and tell you their preferred outcome (refund, store credit, or exchange).
Attach a photo (when needed) — For damage or "arrived broken" claims, the customer can upload photos. Each one is automatically checked for authenticity (see Photo Authenticity).
Get an RMA and track it — They receive a return authorization (RMA) code and can come back any time to track status — submitted, approved, label sent, in transit, received, refunded — right on the same page. Every email you send links back to it in one click.

Two optional touches make the portal a loyalty surface, not just a form:

Dynamic return window badge — Shoppers can see a personalized window like "Your return window: 45 days (Gold member)", calculated from how valuable and how trustworthy the customer is. It turns your return policy from fine print into a perk.
Sustainability footer — An optional "This brand has saved X kg of CO2" badge that tells your sustainability story where it matters.

Working a Return: Approve, Reject, Refund, Exchange

The return detail page is where you do almost everything. Open any return from the Returns list and you'll see the full picture in one place, then take action with one click.

What you see at the top:

Status and outcome — Where the return is in its lifecycle, its total value, and when it was requested.
AI fraud score — A 0–100 score with a low/medium/high risk band and a list of the specific signals that drove it, each with its impact. (Details in AI Fraud Scoring.)
Customer context — If you have other Kernvel apps connected, trust and risk signals appear here (membership, wishlist or auction engagement, cross-merchant return history).
AI recommended outcome — The decision engine's pick (refund, store credit, or exchange), its confidence, and its reasoning. You can hit Re-run AI decision to refresh it.
Items and customer comments — What's being returned, in what condition, and the customer's own explanation (which the AI also reads when scoring fraud and checking photos).
Photo evidence — Thumbnails with an authenticity score and flags for anything suspicious.

The actions you can take (ReturnsIQ greys out anything that doesn't apply to the current status, so you can't double-refund or reject an already-approved return):

Approve — Accept the return. If you have a carrier connected and auto-labels on, a prepaid label can be generated automatically.
Request photo — Ask the customer for photo evidence before deciding; this emails them and moves the return to "photo required."
Reject — Decline with a reason (which is recorded in the audit trail and, if you've configured it, shown to the customer).
Issue refund / store credit — Pay the customer back as cash or as a Shopify gift card. (See below.)
Create exchange draft order — Swap for a replacement item. (See Exchanges.)
Generate return label — Buy a prepaid label in one click. (See Carrier Labels.)

Lower on the page you can manage return-shipment tracking, leave internal staff notes (never shown to the customer), and — for edge cases — manually override the status. There's also a link to the full audit trail for that return.

The Returns List

The Returns screen is the master list of everything filed. Use the tabs (All, Needs review, Approved, Refunded, Rejected), the search box (by RMA, email, or name), the risk-band filter, and the sort options (date, fraud score, or value) to find exactly what you need. Select multiple returns with checkboxes to bulk-approve, bulk-request-photos, or bulk-reject up to hundreds at once — ReturnsIQ tells you how many succeeded.

Filing a Return for a Customer

When a customer reaches you by phone or email instead of the portal, open Returns → Create return. Look up their order by order number and email, pick the items with per-item condition and reason, add a note, and create the return. It lands in your queue and on the return detail page just like a portal-filed one.

Refunds & Store Credit

ReturnsIQ issues real payouts inside Shopify — not IOUs in a separate ledger.

Refunds create a genuine Shopify refund against the order, so your accounting stays correct.
Store credit issues a real Shopify gift card the customer can spend at checkout.

On the return detail page you set the amount (pre-filled with the order value), an optional restocking fee (subtracted from the payout and recorded on the return), and — for store credit — a bonus % to sweeten the credit so customers choose it over cash. Click Issue refund or Issue store credit and ReturnsIQ does the rest, showing you the exact net figure.

ReturnsIQ has guardrails baked in: it never refunds more than the order's value, never issues both a cash refund and a gift card on the same return, and is protected against accidental double-refunds.

Exchanges

To swap an item rather than refund it, use the exchange action on the return detail page. Enter the Shopify variant ID of the replacement item and a quantity, and ReturnsIQ creates a Shopify draft order for the replacement. You finish the exchange from the draft order in Shopify like any other order.

Exchanges are the smart move when an item is a hot seller or low in stock — you keep the revenue instead of giving it back. The decision engine takes inventory into account when recommending an exchange (see The AI Decision Engine).

Return Reasons, Conditions & the Return Lifecycle

Return reasons capture why something came back. ReturnsIQ ships with a built-in set every store gets — doesn't fit, not as described, defective or damaged, wrong item received, quality not as expected, changed my mind, ordered multiple sizes to try, arrived too late, found a better price elsewhere, and unwanted gift — and you can add your own. These power your reason analytics and clustering, so you can see the root causes (and fixes) behind your returns.

Item conditions capture what state the item is in — new, opened, damaged, or broken. Condition feeds fraud scoring (a "new" item with a "damaged" claim is a red flag) and disposition (items marked damaged or broken are automatically held back from restock).

The return lifecycle moves through clear states, all visible on the return detail page and the customer's tracker:

Requested → Photo required → Under review → Approved → Label sent → In transit → Received → Restocked → Refunded (or Rejected / Cancelled at any decision point)

You normally never touch the lifecycle manually — approvals, labels, tracking webhooks, and refunds advance it for you — but a manual status override is there for corrections.

Return Policies & the AI Policy Builder

Policies are your automation rules. Set them once and ReturnsIQ runs them on every incoming return so clean returns sail through and risky ones get held.

On the Policies screen you see every rule with its priority, status, the action it takes, and how effective it's been (times matched, times approved, fraud caught). Policies are checked in priority order — the first one that matches wins.

A policy has two parts:

Conditions — Combine things like order value, fraud score, customer VIP status, days since the customer's first order, and more, using AND/OR logic. For example: order value ≤ $50 AND fraud score ≤ 30.
Action — What to do when the conditions match: auto-approve, request photos, or flag for fraud review.

Write Policies in Plain English

You don't have to build conditions by hand. In the policy builder, type what you want in plain English — "Auto-approve orders under $50 with low fraud scores" — and the AI policy translator turns it into the structured conditions for you, grounded in your store's real reason names and customer segments. Review the result, tweak if needed, and save. (Available on the Growth plan and up.)

AI Fraud Scoring

Every return is scored for fraud risk the moment it's filed — a number from 0 to 100, sorted into a low, medium, or high risk band. The score isn't a black box: the return detail page lists the exact signals behind it and how much each one contributed, so you can trust (and defend) the decision.

ReturnsIQ looks at six families of signals:

Customer history — Their past return rate, prior fraud flags, whether they're on your blocklist, whether they're brand-new, and return velocity (e.g. three or more returns in 30 days).
Order context — Order value, how long after delivery the return came (very late returns are riskier), full vs. partial returns, and bracket-buying (buying several variants of one item).
Item signals — Bracket-buying, return-prone categories, and condition/reason mismatches (claiming "damaged" on an item marked "new").
Behavioral signals — A mismatch between the purchase and return locations, copy-pasted reason text, or a suspiciously fast portal submission.
Photo evidence — How the photo authenticity check came out (reused images, AI-generated images, and edits all push risk up).
Cross-store intelligence — Signals from the fraud network and your connected Kernvel apps (and trust signals, like an engaged member, that push risk down).

You control the thresholds in Settings → Fraud thresholds: where low becomes medium, where medium becomes high, and the score above which a return is auto-rejected. Set them strict to block more, looser to approve more.

It gets smarter with use. ReturnsIQ starts with a transparent rules-based score that works from day one. Once your store has accumulated enough reviewed returns (around 200 labelled examples), it can blend in a model trained on your store's outcomes for sharper, store-specific scoring.

Teaching the Scorer (the accuracy loop)

On a return, you can mark whether a fraud call was right or wrong. That feedback flows into the Fraud detection dashboard, which shows your precision (of the returns you flagged, how many were really fraud) and recall (of all the real fraud, how much you caught), plus the dollars you protected versus the dollars that leaked. Marking decisions over time tightens the scoring for your store.

Photo Authenticity

When a customer uploads a photo to back a damage claim, ReturnsIQ checks whether the image is trustworthy — so "it arrived broken" with a stock photo never earns a free refund. Each photo gets an authenticity score (0–100) and clear flags for anything off, shown right on the return.

The checks include:

Recency and edits — Reading the photo's embedded data to spot images taken long before the return, or ones run through editing software like Photoshop.
Reuse / duplicate detection — A visual fingerprint of every photo is compared against past uploads, so an image recycled from an earlier claim (even a slightly altered copy) gets caught.
AI-generated detection — Looking for the tell-tale markers and provenance tags of images created by AI image generators, with a vision model as a second opinion.
Damage-claim match — A vision model checks whether the photo actually shows the damage the customer described.

When something's wrong, the photo is badged on the return detail page — for example AI-generated, reused, edited, or doesn't match claim — and that result feeds the fraud score. You can turn the vision-model check on or off in Settings → Vision LLM (the metadata, reuse, and AI-generation checks always run regardless).

Coming soon: an additional layer that searches the public web (stock-photo sites, marketplaces, social) for a customer's return photo, to catch images pulled from the internet. This appears as Coming soon in Settings until it's switched on.

Predictive Return-Risk (At-Risk Orders)

Some returns you can see coming. ReturnsIQ scores orders for return risk before they ship, so you can step in while it's still cheap to do so — a prevented return costs far less than a processed one.

The At-risk orders screen lists flagged orders with a risk band, a score, an expected return rate, and a recommended action. Click "Why?" on any order to see the factors behind the score in plain English — for example, a customer buying three sizes of the same item, a historically return-prone shopper, or a high-return product category.

ReturnsIQ recommends an intervention based on how risky the order is:

Medium risk → send a size-guide email to head off a fit return.
High risk → hold the order for 24 hours and alert you for a quick look.
Very high risk → offer an exchange before the order even ships.

Click Mark sent to log that you've intervened. (Available on the Growth plan and up.)

Buyer-Remorse Interventions

Many returns are decided within a day or two of purchase — sometimes before the parcel even moves. The Buyer Remorse queue flags orders that show remorse signals (late-night impulse buys, a first-time customer in a usually-returned category, bracket-buying) and lets you send a gentle nudge — a reassurance message, a size guide, or a pre-emptive exchange offer — before a return is ever filed.

You'll see each flagged order with its top signals and a recommended message. Click Send to dispatch it. The queue then tracks the outcome: did the shopper keep the order, or return it anyway. The top of the page shows your 30-day totals — interventions sent, items kept, items returned. (Available on the Growth plan and up.)

The AI Decision Engine

When you approve a return, picking the wrong outcome quietly costs you money — refund a hot seller and you lose the sale; force store credit on a VIP and you annoy your best customer. The decision engine recommends the outcome that protects the most lifetime margin: refund, store credit, or exchange.

It weighs things like:

The customer's stated preference and their lifetime value / VIP status (lean toward a clean refund for your best customers).
Fraud history (prior flags push away from a cash refund).
Inventory (if you have StockIQ connected: lean toward an exchange on hot or low-stock SKUs, toward a refund on dead stock).
Item margin (push store credit on thin-margin items) and any brand rules you've set.
Item condition.

The recommendation appears on the return detail page with a confidence level and a short, readable reason. When store credit wins but the customer asked for cash, the engine can suggest a small bonus to make the credit more appealing. It's always advisory — you make the final call, and you can re-run it any time. (Available on the Growth plan and up.)

Save the Sale

A flat return form turns every request straight into a refund. The Save-the-Sale agent inserts a brief, helpful conversation first — offered to the customer in the portal as they file a return — to turn a refund into kept revenue when it makes sense.

It can make two kinds of offers, within limits you set:

A sweetened store-credit offer — keep the value in your store, plus a bonus you control (e.g. an extra 15%).
A "keep it" (returnless) refund — for low-value items where shipping back costs more than it's worth, the customer keeps the item and still gets refunded.

When a customer accepts, ReturnsIQ executes it immediately — a real gift card or refund — so they leave with a resolution, not a "we'll be in touch." High-risk returns are never auto-resolved this way; they're sent to you for review.

The Save the Sale screen in your admin shows how it's performing: saves in the last 30 days, revenue retained, a breakdown by offer type, and recent saves. You set the guardrails — the maximum store-credit bonus and which offers are allowed — in Settings → Save-the-Sale agent. (Available on the Growth plan and up.)

Carrier Labels & Return Shipping

ReturnsIQ generates real prepaid return labels in one click. Connect your own Shippo or EasyPost account (your API key is encrypted at rest and used only for your store), set your return address, and click Generate return label on any return. ReturnsIQ buys the label, fills in your return address as the destination, and hands back a printable label with a tracking number.

Bring your own carrier account — Choose Shippo, EasyPost, or manual (no auto-generation) in Settings → Carriers. Your rates are your rates.
Tracking that updates itself — When the parcel moves, carrier tracking updates the return's status automatically (label sent → in transit → received).
Skip labels when you don't need them — Refund-only and "keep it" returns don't require a label at all.

On the return detail page you can also record tracking manually, choose whether to email the customer the tracking link, and click Mark received when the parcel arrives (which always emails the customer). Return-label generation is available on paid plans.

Disposition & Warehouse Capture

Once an item is back, the question is what to do with it — and that decision is real money. ReturnsIQ's disposition tools score a returned item's condition and recommend where it should go: resell, refurbish, liquidate, donate, or discard.

Warehouse capture is a mobile-friendly intake flow. A warehouse worker opens the return on a phone or tablet, photographs the item from a few angles (the camera opens right in the browser), enters its MSRP and category, and hits Evaluate. A vision model returns a condition score, a recommended disposition, an estimated resale value, and the best resale channel — plus flags for any damage it spots.
The Disposition screen collects every verdict, with totals for recoverable value and CO2 saved, a breakdown by outcome, and a searchable list of decisions. Borderline-but-valuable items are badged "Worth reviewing" so you can give them a human glance.

These operational tools are part of the Enterprise plan — see Plans & Billing.

Carrier Claims

When a return comes in as damaged in transit, the shipping cost shouldn't be your loss. ReturnsIQ helps you recover it: it composes the claim narrative and queues a carrier claim. For supported carriers (via EasyPost), you can file the claim directly from the Carrier claims screen; for carriers that aren't supported, the claim is marked for manual filing through the carrier's own portal, so nothing falls through the cracks.

The screen tracks claims queued, filed, needing manual filing, and dollars recovered. Carrier claims are part of the Enterprise plan — see Plans & Billing.

Cross-Store Fraud Network

A serial returner who's burned other stores is invisible to you on their first order — until ReturnsIQ's cross-store fraud network gives you an early warning. It's a shared, privacy-safe signal: when stores flag fraud, the network warns other members about the same bad actors, so you can catch a repeat fraudster you've never seen before.

Privacy is the whole design:

Only one-way hashed identifiers are shared (a scrambled fingerprint of things like email or address) — never raw customer data, and never anything you could read back.
You never see another store's customers or signals. You only get an aggregated risk indication on your own returns.
It's opt-in and off by default, and you can opt out anytime — opting out immediately purges every signal your store contributed to the network.

You manage it on the Fraud network page and in Settings → Cross-merchant fraud network. This is an enterprise-tier capability — see Plans & Billing.

Note: the cross-store fraud network is different from sister-app intelligence (using your own connected Kernvel apps as fraud context). Sister-app intelligence is covered in Connections and in the separate Integrations guide.

The AI Copilot

The Copilot lets you ask questions about your returns data in plain English and get an instant answer — as a number, a table, or a chart — without learning any reports. Ask things like "Show me the top fraud-flagged returns this week," "What are the most common return reasons?," "Which SKU has the highest return rate?," or "What's my return-rate trend over the last 8 weeks?"

It's built to be safe: the Copilot chooses from a fixed set of pre-built, read-only tools — it never runs arbitrary queries against your data. Start from the suggested question chips (grouped into Fraud & Risk, Reasons & Products, Operations, and Trends) or type your own. The Copilot is available on paid plans.

Customers

The Customers screen is a per-shopper view of return behavior. The list shows each customer's lifetime orders, lifetime returns, return rate, lifetime value, and any flags (VIP, fraud, blocked), with search and filters for VIP-only or flagged-only.

Open a customer to see their full profile:

Their numbers — orders, returns, return rate, and lifetime value, plus VIP / fraud / blocked status and first and last return dates.
Cross-merchant network signal — if the network flags them, a risk indication (with how many stores contributed) appears here.
Fit profile — the customer's preferred size per category and brand, built from their order and return history. This is what powers the pre-purchase fit recommendation on your product pages (available on the Scale plan and up).
Recent returns — their return history, each linking to its full detail.

Analytics

The Analytics screen is your returns overview, sliceable by 7, 30, or 90 days:

Headline numbers — total returns, dollars refunded, dollars of fraud caught, and (on eligible plans) how many orders were flagged as high return-risk.
Returns over time — a daily trend so you can spot spikes.
Outcomes — the mix of refunds vs. store credit vs. exchanges.
Risk distribution — how your returns split across low, medium, and high risk.
Fraud report — dollars and counts for fraud caught and estimated fraud leaked.
Predictive risk — on eligible plans, how well the pre-fulfillment risk model is performing.

AI Intelligence: Reason Clusters & Forecasts

The AI Intelligence screen turns the raw stream of returns into patterns you can act on. It brings together four AI surfaces in one place:

Reason clusters — Instead of reading hundreds of free-text return comments one by one, ReturnsIQ groups similar reasons into clusters using text embeddings and labels each one, then suggests a concrete product fix to ship (e.g. "runs small — update the size chart"). Fixing the cluster cuts the returns at the source. Hit Regenerate to recompute against your latest returns.
Predictive returns forecast — A 30/60/90-day projection of return volume and value, broken down by SKU and reason, so you can clear hot SKUs before peak and staff your returns desk for what's coming. Switch the horizon and regenerate on demand.
Fraud accuracy — The same precision/recall and dollars-protected-vs-leaked view described in AI Fraud Scoring, summarized here alongside the other intelligence.
The policy AI writer — A jumping-off point into the plain-English policy translator.

Reason clustering and the predictive forecast are part of the Growth plan and up; if your plan doesn't include them, the screen shows an "Upgrade to unlock" panel that explains what each one does.

Sustainability

Returns have a carbon cost, and handling them well is a story worth telling. The Sustainability screen turns your disposition data into ESG reporting: CO2 saved, items kept in circulation, and a breakdown of where returned items went, over the last 30 and 90 days. It also surfaces relatable equivalences (trees planted, miles not driven) — clearly labelled as illustrative estimates — and gives you a storefront trust badge you can toggle on to show shoppers your impact. The figures come from real disposition and shipping data, not invented numbers. Sustainability reporting is available on paid plans.

The Audit Trail

Every return keeps a complete, time-stamped event timeline — open it from the View audit trail link on any return. It records every status change, every fraud score, every policy that matched, every AI recommendation (and what you chose instead), and who did what, in plain English. It's exactly what you need to defend a decision to a customer or a chargeback arbiter.

Settings

Settings is where you make ReturnsIQ work the way your store works. Everything ships with sensible defaults — change only what you care about. The main areas:

General — Where merchant alerts are emailed.
Return address — The destination on your return labels.
Carriers & return labels — Connect your own Shippo or EasyPost account (keys encrypted, used only for your store) or choose manual labels.
Branding — Portal accent color, logo, custom welcome copy, dark theme, and your Returns page path.
AI feature toggles — Master on/off switches for fraud scoring, photo authenticity, predictive return-risk, the decision engine, and auto-approve.
Fraud thresholds — The low/medium/high boundaries and the auto-reject cutoff.
Membership influence — How much an active GateIQ membership favors a refund (off / advisory / strong). The fraud gate always runs after, so a flagged member can never auto-approve a cash refund.
Sister-app intelligence — Use your connected Kernvel apps as fraud context (available on the Scale plan and up; separate from the cross-merchant network).
Save-the-Sale agent — Maximum store-credit bonus and which offers are allowed.
Dynamic return window — Base days, a VIP multiplier, and minimum/maximum bounds.
Returnless refund — The maximum item value to offer "keep it," and whether to require a clean fraud history.
Buyer-remorse intervention — The score that triggers an intervention.
Vision LLM — Turn the vision-model photo check on or off.
Notifications — Pick which events email you (new return, approval, high-risk, refund).
Storefront widgets — Toggle the dynamic return-window badge and the sustainability footer.
Cross-merchant fraud network — Opt in/out (enterprise).
Restocking quality threshold — The minimum condition score to call an item resaleable (enterprise).
Data export & danger zone — GDPR deletion runs automatically through Shopify's compliance webhooks; for exports, contact support.

Tier-gated sections show an "Upgrade to unlock" badge when they're not on your plan — nothing to set up until you upgrade.

Plans & Billing

ReturnsIQ is free to start and priced to grow. The free plan needs no credit card, and every paid plan comes with a 14-day free trial. You can switch between monthly and yearly billing at any time, and annual billing saves you up to 17%.

The Billing screen shows your current plan, status, next billing date, and usage meters (returns this month and active policies against your limits). From there you can upgrade, downgrade, switch billing intervals, or cancel. There's also a built-in FAQ covering common billing questions.

At a high level, higher plans raise your monthly return allowance and unlock more of the AI:

Free — The branded portal, basic fraud rules, real refunds and store credit, and one policy — enough for a small store to run real returns.
Starter — Adds AI fraud risk scoring, photo authenticity checks, prepaid carrier labels, and more policies.
Growth — Adds the predictive return-risk and buyer-remorse tools, the AI decision engine, Save-the-Sale offers, reason clustering, the predictive forecast, and the AI policy translator.
Scale — Higher volume plus cross-store fraud intelligence and the pre-purchase fit prediction.
Enterprise — Adds the operational depth: the cross-merchant fraud network (opt-in), warehouse capture and disposition decisions, and carrier-claim filing.

Check the pricing on the ReturnsIQ overview page and the in-app Billing screen for the current limits and prices, which are always the source of truth.

If you cancel: the AI agents (Save-the-Sale, disposition, buyer-remorse, photo authenticity, and the ML fraud overlay) are paused, and ReturnsIQ falls back to its transparent rule-based behavior. Your data and history stay intact.

Connections (Your Other Kernvel Apps)

ReturnsIQ stands on its own, but it gets sharper when you run it with the rest of the Kernvel suite. The Connections screen lets you turn each link on or off — they switch on automatically once both apps are installed, and they share only the minimum signal needed (counts, flags, hashed IDs), never raw customer lists.

In short, your returns data makes the suite smarter (StockIQ stops over-ordering high-return items, and your other apps get a "serial returner" trust signal), while your other apps make ReturnsIQ smarter (inventory, membership, and engagement signals improve every return decision).

For the full picture of each connection and the outcome it delivers, see the ReturnsIQ Integrations guide.

FAQ

Do I need to write any code? No. The only manual step is adding the ReturnsIQ Portal block to a page in your theme editor — a drag-and-drop, no code. Everything else is point-and-click.

Does ReturnsIQ issue real refunds? Yes. Refunds create genuine Shopify refunds, store credit issues real Shopify gift cards, and exchanges create real Shopify draft orders. Nothing is a placeholder.

Will the AI auto-refund or auto-reject without me? Only where you allow it. You set the auto-reject threshold and decide whether clean, policy-matched returns auto-approve. High-risk returns are always sent to you, and self-service Save-the-Sale offers are refused on high-risk returns. Every AI output is a recommendation you can override.

What counts toward my monthly return limit? Each return filed in a billing month counts toward your plan's allowance. The Billing screen shows your current usage against the limit.

Is the fraud scoring a black box? No. Every score lists the exact signals behind it and how much each contributed, and the full history is in the audit trail.

Do I have to use my own shipping carrier? Yes — you connect your own Shippo or EasyPost account so you get your own rates. You can also choose manual labels and skip auto-generation entirely.

Is the cross-merchant fraud network safe for my customers' privacy? It shares only one-way hashed identifiers — never raw customer data — it's opt-in and off by default, you never see other stores' data, and opting out immediately purges every signal your store contributed.

What happens to my data if I cancel? Your data and history remain. The AI agents pause and ReturnsIQ reverts to its rule-based behavior. GDPR deletion requests are always honored automatically through Shopify's compliance webhooks.

Will it slow down my checkout or storefront? No. The portal is a self-contained app block, and returns are scored after they're filed — your checkout and product pages are unaffected.

Glossary

RMA (Return Merchandise Authorization) — The unique code assigned to a return; how you and the customer reference it.
Fraud score — A 0–100 risk number on every return, sorted into a low/medium/high band, with a breakdown of contributing signals.
Risk band — The low / medium / high grouping a fraud score falls into; you set the boundaries.
Authenticity score — A 0–100 trust rating for an uploaded return photo.
Wardrobing — Buying an item, using it once, and returning it.
Bracket-buying — Ordering several sizes or variants of the same item intending to return most of them.
Disposition — The decision of what to do with a returned item: resell, refurbish, liquidate, donate, or discard.
Returnless / "keep it" refund — Refunding the customer without requiring them to ship the item back, used when return shipping costs more than the item is worth.
Save-the-Sale — The agent that offers an exchange or sweetened store credit during a return to retain revenue.
Decision engine — The feature that recommends refund vs. store credit vs. exchange for each approved return.
Predictive return-risk — Scoring an order for the likelihood it'll be returned, before it ships.
Buyer-remorse intervention — A proactive nudge (reassurance, size guide, or exchange offer) to a shopper likely to return, before they file.
Dynamic return window — A per-customer return window length based on their value and trustworthiness.
Cross-merchant fraud network — The opt-in, privacy-safe network that shares hashed fraud signals across stores.
Connections — The screen where you turn sister-app (other Kernvel apps) links on and off.

Getting Help

Questions, a stuck return, or want a hand with setup? Email support@kernvellabs.com and the Kernvel Labs team will help. For how ReturnsIQ works with your other Kernvel apps, see the Integrations guide.