Receipt Scanning

Earn points on purchases you never saw.

Members photograph a receipt. AI reads it, your rules decide it, and points land automatically — even when the sale happened in someone else’s store, on someone else’s marketplace, entirely off your radar.

Receipt moderation Approved Review Rejected
GREEN VALLEY GROCERS
03 Aug 2026 · 14:22
OAT DRINK 1L x27.80
PASTA SAUCE 400G4.15
COFFEE BEANS 500G12.90
HOUSEHOLD SUNDRIES59.35
TOTAL$84.20
OCR confidence 0.94
Moderation rules · first match wins
Receipt dated within 90 daysPass
Store on approved retailer listPass
Total below $500 thresholdPass
Confidence at or above 0.900.94
Auto-approved
420 points credited in seconds
METRO MART SUPERSTORE
29 Jul 2026 · 19:05
SEASONAL BULK ORDER418.00
GIFT SET x4156.00
PROMO DISCOUNT-61.10
TAX100.00
TOTAL$612.90
OCR confidence 0.88
Moderation rules · first match wins
Receipt dated within 90 daysPass
Store on approved retailer listPass
!Total below $500 threshold$612.90
!Discount under 15% of totalFlagged
Sent to manual moderation
A moderator confirms before points are awarded
CORNER STORE #114
12 Feb 2026 · 08:41
SNACK PACK3.40
BOTTLED WATER x65.70
MISC ITEMS22.00
TOTAL$31.10
OCR confidence 0.71
Moderation rules · first match wins
Receipt dated within 90 days178 days
Store on approved retailer list
Total below $500 threshold
Confidence at or above 0.90
Auto-rejected
Member emailed with the reason, no ticket raised

Proof of purchase, without the paperwork

Manual review doesn’t scale. Blanket auto-approval invites fraud. Receipt Scanning sits between the two — and you decide exactly where.

Tune the automation

Start conservative. Dial up as OCR earns your trust.

Two settings govern the whole operation — and either one moves without a release.

Automation percentage — how much volume OCR handles
Confidence threshold — how sure a read must be to pay out
Human approval gate — optional final sign-off, always available
OCR settings Per 1,000 receipts
Automation percentage 25% 50% 75% 100%
Confidence threshold 0.7 0.8 0.9
Auto-approved by OCR370
Below threshold → moderator130
Never sent to OCR → moderator500
0.8 is the usual settling point once OCR accuracy has been validated on your receipt formats.
Independent retail Marketplaces Distributors
MEMBER
+420
points
Earning without a POS integration
Channels you don’t own

Reach the sales your data never sees

If you sell through independent retailers, marketplaces or distributors, most of your volume is invisible to your loyalty program. A photographed receipt turns that volume into member data — and members into people you can actually talk to.

Independent retail Marketplaces Distributors
Rules, not code

Fraud controls your team can change on a Tuesday

Stale dates, unapproved stores, suspicious discounts, unusually large totals — each is a configurable condition, evaluated top-down with auto-reject taking priority. Tighten your risk tolerance in an afternoon instead of a release cycle.

Receipt age Store allow-list Total threshold Discount limits
EVALUATED TOP-DOWN
Receipt over 90 days old Auto-reject
! Discount over 15% Review
Store on approved list Approve
Auto-reject first, review over auto-approve
Snap receipt
Upload
Approved +420
Received email Approved email Reason if rejected
Inline or popup, styled as yours
Member experience

A submission that feels like part of your brand

Build the upload widget as inline or popup, with your own button, upload page and thank-you page. Received, approved and rejected emails go out under your sender name, personalised down to the rejection reason — so a declined receipt still reads like service, not silence.

Web & app WhatsApp & SMS API
Reporting

See the queue, the throughput and the risk

Every submission is measured end to end — how much came in, how fast it cleared, what got rejected and why. Staff the review queue against evidence instead of instinct.

Volume & throughput Quality & risk
Receipt Scanning Created At Illustrative data
Total receipts created
1,537
All uploads, all sources
Total processed
396
Decision made
Processing rate
25.8%
Of everything received
Open backlog
1,141
Awaiting a decision
Open receipts by age
0–7 days24
8–14 days62
15+ days1,055
Days since upload
Current ticket status
Manual moderation700
Pending440
Points awarded205
Disapproved190
Open and closed tickets
Receipts by type
Automated891
Manual640
Unclassified6
Automated vs. human-handled
Receipts by source
API780
Dashboard430
WhatsApp280
SMS45
Where the upload came from
Closed same day
88%
Of all closed receipts
Flagged for fraud review
470
Routed to a person
Confirmed fraud
120
Blocked before payout
Duplicate rejections
53
Top rejection reason
Closed receipts by time to close
1 day355
2 days8
1 week12
2 weeks10
3 weeks+22
Most decisions land same-day
Rejection reason by volume
Duplicate receipt53
Test / manual reject33
Other7
Routed to moderation2
Fraud review1
Why receipts were declined
Fraud detection volume
Not fraud820
Requires review470
Unscored125
Fraud120
AI fraud screening outcome
Receipts created vs. decisions
Wk 3092
Wk 31148
Wk 3261
Wk 33171
Wk 34156
Weekly peaks drive the backlog

Where brands point it

CPG

Points for the grocery aisle

Sold through thousands of independent stores with no shared POS data. Receipts become the proof of purchase, with eligibility limited to an approved retailer list.

Earning extended to the majority of actual sales volume
Retail

Manual review, only where it pays

Routine low-value receipts clear automatically; anything above the dollar or discount threshold is routed to a moderator before points move.

A review queue sized to risk, not to volume
Beauty

Reward trial in someone else’s store

A new line launches at a department store partner. Receipt Scanning credits the trial purchase and identifies who bought it — data the brand had no other route to.

First-party data from third-party shelves
Receipt Scanning

Every purchase counts, wherever it happened

Extend earning to the channels you can’t measure, keep fraud on a leash, and let your team review only the receipts that genuinely need a person.

Offline purchases captured Fraud rules you control Only the hard cases reach a person
Learn more

You might also be interested in

View all resources