You pay for
decisions.
One rate, charged per verdict: every submission that came out of the pipeline with a decision, priced by the kind of evidence. Seats, workflows, the API and the platform itself are part of the product, not line items.
SteveOne line
on the invoice.
From €0.12 a decision
Verdicts × your rate
Every submission that leaves the pipeline with a decision — approved, flagged, blocked, or rejected by your rules — at the rate agreed for that kind of evidence. Higher volume lowers the rate. Nothing else is counted.
What never moves that number
- RefusalsEvidence the quality gate turns away never reaches a rate. Neither do processing failures.
- SeatsReviewers, admins and API users, as many as the work needs. There is no per-seat charge.
- The platformAdmin panel, review queue, workflows and rules, the API, webhooks, integrations, setup and support.
One pricing model. Many kinds of work.
From purchase verification and product registration to POSM placement, claims evidence, and call compliance: many workflows share the same evidence formats. Rates reflect the format and processing involved. Your agreement lists the ones you use.
Receipts and documents
Priced per verdict
Receipts, invoices, warranty cards, claim documents, forms
- What counts
- One image or PDF that receives a verdict: approved, flagged, blocked, or rejected by your rules.
- Not billed
- Photos refused by the quality gate. Processing failures.
Photos and visual evidence
Priced per verdict
Shelf execution, POSM placement, store standards, damage evidence, maintenance photos
- What counts
- One photo checked against your questions and your product gallery, with a verdict. A visit with five photos is five verdicts.
- Not billed
- Photos refused by the quality gate. Photos that could not be matched to a store or task when that is required.
Call recordings
Priced per verdict, by duration band
Mystery calls, sales and service calls, drive-through audio
- What counts
- One recording transcribed and scored against your script. Banded by length: up to 5 minutes, up to 15, up to 30. Longer on request.
- Not billed
- Recordings too poor to transcribe. Silence and dead air.
Messages and transcripts
Priced per verdict
Chat transcripts, emails, support replies, bot conversations
- What counts
- One conversation or message thread checked against your standard. Extraction only is a lower rate.
- Not billed
- Empty or system-only threads.
Rates depend on the evidence class and on the volume you commit to. Ask for your numbers and we will send them after the first call.
Included, at no extra charge.
Everything you need to run, see, and act on verdicts. None of it is metered, and none of it is priced per seat.
| Part of the product | |
|---|---|
| Seats | No per-seat pricing. As many reviewers, admins, and API users as you need, at no extra cost. |
| Admin panel and review queue | Evidence, fields, and the reason side by side. One click to confirm or clear. |
| Workflows and rules | As many workflows as you have use cases. Declarative rules you can read, versioned on every change. |
| Fraud detection | Perceptual image matching and identity fingerprints against the full history of each workflow. |
| Capture quality gate | Runs before processing on every submission. Rejections are not billed. |
| API, SDK, and webhooks | Submit evidence and receive fields and verdicts. Idempotent submissions. No call volume limits tied to price. |
| Open Loyalty integration | Verified receipts post as transactions. Native, no connector fee. |
| Reporting and statement | Submitted, gated, approved, flagged, blocked, and billed, per workflow, per month. |
| Workflow setup | We configure the first workflows with you on your real evidence before go-live. |
| Support | Email and shared channel. Response times in the agreement. |
| Data retention controls | Retention per workflow. Originals canonicalised, not retained beyond the window you set. |
On request
- Volume commitments
Committed monthly verdict volume lowers the per-verdict rate. - Dedicated environment
Single-tenant deployment and regional hosting for regulated workloads. - Custom SLA
Verdict latency and availability targets above the standard agreement.
How much of the month
goes into checking?
Tick what you check and set your volume. Each kind of evidence is priced on its own and the lines add up. We fill in the rest: the discount your volume earns, €22 an hour of review time, and the share that still reaches a person.
Past 100,000 submissions a month the rate is quoted rather than listed, and the platform side of the agreement changes with it.
Rates here are list prices before the discount your volume earns, and your own rate is confirmed in your quote. Review time is costed at €22 an hour and a full-time reviewer at 160 hours a month. Steve checks everything; about 15% still reaches a person, at roughly half the time because the evidence and the reason arrive with them, and those hours are held back from the reviewer time we show you getting back.
Get your rateHow a submission becomes a line on the statement.
A submission arrives
Through the app, the API, or Open Loyalty.
The quality gate decides
Unreadable evidence is refused with a reason. Nothing is billed.
The pipeline produces a verdict
Extraction, rules, fraud checks. Approved, flagged, blocked, or rejected by your rules. This is the billed unit.
Your statement shows both
Submitted and billed, side by side, per workflow, with the outcomes.
Pricing questions.
If yours is not here, ask it in the demo. We answer pricing questions in the first call, not the third.
Why not one price for everything?
Because the work is not the same. Reading a receipt and checking it against rules costs a fraction of transcribing a twenty-minute call and scoring it against a script. One blended price would overcharge receipt programmes to subsidise audio. Rates are set per evidence class so each use case pays for its own work.
What exactly is a verdict?
A submission that left the pipeline with a decision: approved, flagged for review, blocked, or rejected by your rules. Extraction-only workflows count a submission as resolved when its fields are delivered. A flagged submission that a reviewer later clears is still one verdict; the review is included.
Do duplicates count?
Yes. Detecting that a receipt was submitted before is a verdict, and usually the most valuable one. Refusing an unreadable photo at the quality gate is not a verdict and is not billed.
What else is on the invoice?
Nothing. Verdicts are the only billed unit. The admin panel, the review queue, workflows and rules, the API, webhooks, the integrations, setup and support are part of the product, not line items, and there is no per-seat charge on top. If your deployment needs something beyond the standard agreement, that is a conversation, not a surcharge you discover later.
Is there a minimum on verdicts?
Pilots run on a short fixed-term agreement with no volume commitment. Production agreements can include a committed monthly verdict volume in exchange for a lower rate; pay as you go is available at the list rate. We will show you both options with numbers in the first call.
How do I know what I will pay this month?
Your statement shows submitted, gated, and billed volumes per workflow, updated daily, multiplied by the rate you agreed. Set an alert on any workflow's volume through the API.
What if the pipeline gets it wrong?
Flag it in the review queue. Every overturned verdict is logged and feeds threshold tuning. Systematic errors on a workflow are credited. That is in the agreement, not a favour.
Working with teams at
Your rate,
in one call.
Bring your volumes and one workflow. You leave the call with your rate per verdict and a pilot plan.
Book a demo 30 minutes. Your use case. Real possibilities.