# Returns and refund review > Returns and refund review: Policy met; abuse indicators Checked: vetoed by code. Produces approved or held refunds. Source: https://www.sumarity.ai/library/returns-and-refund-review/ Retail, hospitality and consumer # Returns and refund review Reads return requests, receipts, the refund policy. Produces approved or held refunds. Code works out every figure, the Judgement Engine answers the narrow questions, and whatever stays uncertain goes to a person. Vetoed by codeClaims, disputes and exceptions Start from this workflowMore for retail and hospitality What it decides ## Narrow questions, each with a check behind it. Policy met; abuse indicators Vetoed by code: receipt, dates. Code can't prove the answer right, but it can prove it wrong: an amount out of range, an impossible date, a name not on the document. Many settle; the rest go to a person. | Reads | Return requests, receipts, the refund policy | Produces | Approved or held refunds | Who signs | A person on your team, with every figure traced to its document and every call on record | Process | Claims, disputes and exceptions | Industry | Retail, hospitality and consumer See it run ## Every line, decided where it's safest. An example run. Each stage lights up as a line is decided there; pick a stage to see its lines. ReadsReturn requests, receipts, the refund policyCodeReads, matches, ties out and works out every figure.JudgementPolicy met; abuse indicators. Settles only when it clears the cutoff and no check in code rules it out.A personGets whatever stays uncertain, with the evidence. Their ruling tunes the next run.ProducesApproved or held refunds Returns and refund review · example run Example run | Ref | Item | Value | Decided by | RT-31 | Jacket, 12 days, receiptWithin 30 days, unused | 149.00 | Code · approved | RT-32 | Shoes, wornWorn: store credit per policy · 0.90 | 89.00 | Engine · checked | RT-33 | No receipt, high valueCard lookup found the sale · 0.93 | 420.00 | Engine · checked | RT-34 | 5th return this monthPattern: review | 60.00 | To a person | RT-35 | Electronics, 45 daysOutside the window | 299.00 | Code · declined | Settled 4 To a person 1 Settled wrong 0 Illustrative lines. Amber rows are the Judgement Engine's calls; each settles only above its cutoff and when its check agrees. What it could save ## Your volumes in. Your hours out. In our tests, between 24% and 42% of items still went to a person after tuning. Set your own share; a pilot measures it on your data. Items per run Runs per month Minutes per item by hand Cost per hour Share of items that still need a person: 30% By hand today80 ha month With Sumarity24 ha month, for the items people decide Saved56 hCHF 53,760 a year Related ## Start from what's closest. ### Claims, disputes and exceptions in other work Finance ### Customer deductions and short-pay disputes Valid deduction (promotion, damage) or not Proven by codeBanking ### Payment exceptions and investigations Why a payment failed or is queried; the action Proven by codeBanking ### Card disputes and chargebacks Network reason code; evidence enough to fight Cross-checked ### More for retail and hospitality ### Daily takings reconciliation (tills, cards, cash, bank) Which deposits cover which days and sites Proven by code ### Delivery platform payout reconciliation Orders, commissions, refunds in each payout Proven by code ### Marketplace payout reconciliation (Amazon, Shopify) Sales, fees, returns, reserves in each payout Proven by code Questions ## What buyers ask What happens when Sumarity isn't sure? Nothing settles below the cutoff, or when the check disagrees. The item goes to a person in the inbox with what the engine saw, its best answer and the runner-up. The person's ruling is kept and tunes the next run. How are its judgements checked? Code can't prove the answer right, but it can prove it wrong: an amount out of range, an impossible date, a name not on the document. Many settle; the rest go to a person. Can we change it to fit how we work? Yes. Start from this workflow and describe your differences in plain words. The design assistant revises it, the validator checks it, your expert reviews it on the canvas, and your admin publishes it. Does it read our files as they are? Yes: spreadsheets and CSV in any layout, PDFs, bank formats and e-mail attachments. Sumarity suggests how each column maps, proves the mapping on a sample and remembers it once a person confirms it. Where does our data go? Sumarity runs in Zurich. Your data, backups and logs are stored in Switzerland. How does it get better? Your team's rulings become an answer key. Better questions and cutoffs are proposed, tested on data they haven't seen, and published only when your admin approves. It improves itself, with permission. ## Start from this workflow. Run it on your data. A pilot runs alongside your own process for a few weeks, at our cost, and ends in a line-by-line comparison. Talk to usBack to the library