Request access

Use case · worked example

Paid for the shelf. Proof it's there.

Brands pay stores to put their products on end caps, at eye level, on promotional displays. Most of that money goes out on trust. Teleport checks every participating store on day 1 and day 7, sends a fix when a display is missing, and gives the brand a proof for each one it pays for.

Talk to usAll use cases
A day-1 display audit at the aisle 7 end capA supermarket from above on day 1 of a promotion; day 7 is marked for the second check. An auditor walks from the door along the side aisle to the promotional end cap on aisle 7 and photographs it three times. An inset of the end cap shows its facings counted shelf by shelf and the shelf tag read as four ninety-nine. A weighted score fills past the 8,500 line to 9,375 out of 10,000. Verified, and a receipt prints: compliant.DOORA4A5A6A7END CAP A7PROMO WK 461234567DAY 1 · DAY 7$4.99SCORE9,3758,500VERIFIEDSCORE 9,375COMPLIANT
Example · fictional store · products unbranded
§ 01

Two thousand end caps. Two checks each.

Today a brand finds out how its promotion looked from a sales rep's report, or a sample of store visits weeks later. By then the promotion is over and the payment has gone out.

Who asks
A consumer brand's trade-marketing agent, running a whole promotion within a budget the brand sets.
Who does it
An approved shopper-auditor, the store's own staff, the brand's merchandiser, or a shelf-scanning robot where the store runs one.
Where
Only stores whose retailer has agreed to it, in the aisles it names.
Evidence
Three guided photos: the end cap, the shelf, the price tag. Captured live, checked against the planned layout.
Pay
$5 per accepted audit, whatever the display looks like. Example figure.
You get
A score for every store, twice per promotion, and a proof behind each one.
Example · photo to come

The brand paid for the end cap. Somebody should look at it.

The short version

Each audit is three photos and a few minutes. Most of the scoring is counting and reading, so a store's result is back the same day, while there's still time to fix the display.

Northfield Snacks, Retailer X and store 2291 are fictional. Prices are illustrative.

§ 02

Not pass or fail. A score out of 10,000.

Displays are rarely perfect and rarely absent. Five weighted conditions add up to a score: 8,500 or more is compliant, 5,000 to 8,499 is partial and gets a fix task, below 5,000 gets a fix task and a report.

Pick what the auditor finds
PROMO 46NO HEADER$4.994321DASHED = THE PLAN● = HERO PACKCONDITIONPOINTS / WEIGHTIN ITS SLOT03,0003,000/3,000FACINGS01,8751,875/2,500HERO AT EYE LEVEL001,500/1,500PROMO PRICE01,5001,500/1,500MATCHES THE PLAN001,500/1,500SCORE0 / 10,0006,375 / 10,0009,375 / 10,0005,0008,500NON-COMPLIANT · FIX + REPORTPARTIAL · FIX SENTCOMPLIANT

Day 1 · as found · 0Day 1, and the end cap still holds last week's seasonal stock. Nothing on it is ours, so nothing scores. A fix task goes out and the store is reported.

Day 1 · as found · 6,375The display is up and the price is right, but the hero pack is on the bottom shelf and one SKU is short of facings. A fix task goes to the merchandiser.

Re-audit · same day · 9,375Hero pack moved to eye level, shelf rebuilt to the plan. One SKU is still a facing short, and that's fine: compliant doesn't have to mean perfect.

§ 03

From campaign to paid, in six steps.

The audit finds the problem. The fix task solves it. The re-audit proves it, and the before and after travel together.

  1. 01

    Campaign

    Promo week 46: 2,000 stores, checked on day 1 and day 7. One task per store, per check.

  2. 02

    Audit

    An approved auditor photographs the end cap, the shelf and the tag, guided by the app.

  3. 03

    Score

    Compared with the planned layout: in its slot, facings, eye level, price, the look of it.

  4. 04

    Fix

    Score too low? A fix task goes to store staff or the merchandiser: build it, restock 12 facings.

  5. 05

    Re-audit

    A linked check confirms the fix. The before and after are proven together, as a pair.

  6. 06

    Report + pay

    Compliance per store and per region. The brand releases or holds its own payment to the store.

§ 04

Four of five conditions are counting and reading.

The fifth asks whether the display looks like the plan. That one goes to a model, and a versioned policy, not the model, turns the answers into a score.

ConditionWeightCheckByResult
In its slot3,000The display is on the end cap it was paid for (aisle 7, north), not parked in the back aisle.RuleChecked
Facings2,500At least 12 facings of each of the four promoted packs, counted against the brand's product images.RuleChecked
Hero at eye level1,500The lead pack sits on shelf 3 or 4.RuleChecked
Promo price1,500The shelf tag reads $4.99, matching the promotion. Read by OCR.RuleChecked
Looks like the plan1,500Does the display match the reference planogram? Catches the messy and the half-built.ModelChecked

Every outcome gets a proof, including a missing display. A proof of non-compliance is worth as much to the brand as a proof of compliance. Not sure about a pack? It goes to review, not to “missing”.

§ 05 · Consent and competitor data

The retailer says yes first. Rivals stay out of frame.

A shelf photo is also a photo of every other brand on the shelf. Without the retailer's agreement and strict rules on what the brand gets back, this is espionage, not an audit.

OWN BRAND · KEPTRIVAL PACKS · RIVAL PRICES · FACESCOVERED BEFORE THE BRAND SEES THE PHOTO
Site consent · Retailer X store 2291Active · v1
Zones
Aisle 7 end caps and the promo shelf. Nowhere else.
Capture
Three guided photos per audit
Who
Approved auditors, store staff, the brand's merchandiser
Hours
Outside the 17:00–19:00 rush
Rivals
Covered before delivery. No share-of-shelf data.
Shoppers
Faces covered by default

Part of the trade agreement itself. Its hash goes into every proof from this store.

§ 06

What needs care.

Photos in someone else's store, of someone else's products, paid for by a brand with an interest in the answer. Each of these keeps the audit worth trusting.

  • The retailer says yes first

    Plenty of chains ban photos in store. A shopper photographing shelves for a brand without permission gets asked to leave, and the brand pays for it with the retailer. No consent, no audit.

  • Rivals stay out of frame

    Every shelf photo shows other brands' packs, prices and space. The brand gets its own results only, with rival products covered, unless the retailer agrees to share more.

  • Pay never depends on the result

    Pay more for “compliant” and auditors find compliance. Every audit that passes quality checks earns the same flat rate, and spot checks by robot scans or repeat visits keep it honest.

  • Packs look alike

    Flavour variants, near-identical boxes, one pack hiding another. The brand's product images, two or three angles, and a review queue for low-confidence packs, never “missing” by default.

  • Shoppers in the aisle

    Faces and trolleys end up in frame. They're covered by default, as in every capture Teleport takes in a shared space.

  • Partial is normal

    Retail is rarely all or nothing. Weighted conditions give a score with bands, so 72% compliant is a result the brand can act on, not a failed task.

§ 07

Found partial. Fixed. Proven, as a pair.

Store 2291 scored 6,375 on day 1. The merchandiser moved the hero pack to eye level, the re-audit scored 9,375, and one receipt carries both.

The auditor was paid the same $5 for each visit, whatever the shelf looked like. Dollars or USDC pay the work. Nobody needs $PORT to run a campaign or audit a shelf.

PROOF OF PHYSICAL WORKworkflow · example
VERIFIED
Task
Promo display · store 2291
Requested by
Trade agent, for Northfield Snacks
Performed by
Human · approved auditor
Evidence
Teleport capture: 3 photos, GPS
Checked against
Planogram promo46 v3
Site consent
On file · v1
Score
6,375 partial → 9,375
Decision
Compliant after fix
Paid
$5.00 per audit · either way
STATUS: FINAL

Photos: hashes only on-chain. The brand gets its own results, rivals covered.

Example receipt. Fictional brand and store; amounts illustrative.
§ 08

For builders.

A SCHEDULE trigger, an INSPECT task, and conditions marked required: false with weights, so the task is accepted with a score instead of failing outright. A low score triggers a linked STOCK fix and a re-inspection.

PhysicalTaskSpec · day-1 audit, one storeTAP v0.2 · preview, subject to change
{
  "schema": "teleport.tap.physical-task",
  "version": "0.2.0",
  "title": "Check promo display — store 2291, day 1",
  "category": "INSPECT",
  "requester": {
    "requesterType": "AI_AGENT",
    "principal": { "principalType": "ORGANIZATION" },
    "mandateId": "mdt_cpg_promo_audit_q4"
  },
  "trigger": { "type": "SCHEDULE", "sourceRef": "campaign:promo46" },
  "fulfillment": { "mode": "MARKETPLACE" },
  "relations": { "workflowId": "wf_promo46_store_2291" },
  "environment": {
    "siteRef": "site:retailer_x_2291",
    "siteConsentRef": "consent_retailer_x_aisle_capture_v1",
    "referenceAssets": ["planogram:promo46_endcap_v3", "catalog:cpg_skus_2026q4"]
  },
  "actorRequirement": {
    "permittedActorTypes": ["HUMAN", "ROBOT"],
    "requiredCapabilities": [{ "capability": "CAPTURE.RETAIL_SHELF" }]
  },
  "completionConditions": [
    { "type": "LOCATION", "id": "display_in_slot", "subjectRef": "display:promo46",
      "zoneRef": "site:retailer_x_2291:endcap_a7n", "relation": "IN",
      "weight": 3000, "required": false, "verificationMethod": "PERCEPTION" },
    { "type": "COUNT", "id": "facings_sku_a", "subjectRef": "sku:012345678905", "min": 12,
      "weight": 625, "required": false, "verificationMethod": "PERCEPTION" },
    { "type": "STATE", "id": "promo_price", "subjectRef": "shelf_tag:sku_a",
      "property": "price_usd_cents", "expected": 499,
      "weight": 1500, "required": false, "verificationMethod": "DETERMINISTIC" }
  ],
  "rightsPolicy": { "policyId": "retail-audit-own-brand-only-v1" },
  "economics": {
    "performerPayUsdCents": 500,
    "buyerPriceUsdCents": 900,
    "settlementAssets": ["FIAT", "USDC"]
  },
  "popw": { "required": true, "assurancePolicyId": "VERIFIED", "proofTiming": "WORKFLOW" }
}

Proposed additions this use case surfaced

Reference assets
The planned layout, the product catalogue, the store's reference scan: a hashed list of what the task is checked against, committed to the proof so “expected” can't quietly change afterwards.
Campaign rollups
One result for the whole promotion (compliance, coverage, the spread of scores) with its own root over every store's proof. The brand buys the campaign; the stores back it up.
Score-based outcomes
v0.2 has weights but still decides accept or reject. Add a score and its band to the decision and the proof, so “6,375, partial” is a valid outcome.
Incidental data
Rules for what the requester didn't ask for but the camera caught: rival products, bystanders. Cover, crop, or keep for audit only, and say which in the proof.
Outcome-independent pay
A policy flag that pays the performer for the quality of the evidence, never for what it shows. Protects any inspection market from its own incentives.

Example economics: $9 to the brand per audit, $5 to the auditor; the rest covers scoring, redaction and the proof. Developer docs

Consumer brands

Paying for shelf space you've never seen?

Tell us the promotion and the stores. We're working with design partners on display audit campaigns.