Predictika · Merchant services proposals

An AI sales agent's merchant proposal, checked before it goes out

An AI sales agent prepares a card-acceptance proposal for a merchant: business type, plan, devices, pricing model, term and sales channel. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the plan, device, pricing and approval rules and returns a verdict, every reason and the figures it computed. There is no price list in this model.

Scope: merchant services proposals for point-of-sale devices, plans, pricing structure and approvals. See everything this validator covers
  • Merchant segments — counter service, full service dining, retail, professional services, personal services, and home and field services.
  • Plans and devices — Starter, Standard and Advanced plans; Station, Mini, Flex, Go and Kiosk devices; which segment may use which device; device count limits per plan.
  • Pricing structure — interchange-plus markup and per-transaction fee against a floor that depends on volume and channel; tiered rates for qualified, mid-qualified and non-qualified cards, their order and spread.
  • Risk and term — high-risk merchants, card-not-present share, term length, early termination fees.
  • Sales channel — direct sales, ISO partner and bank partner, with partner revenue share limits and approval thresholds.
  • Stated figures — the device count and the estimated monthly markup revenue written into the proposal are compared with what the validator computes.

Not covered: any list price, rate card or fee schedule. This validator models the structure of a proposal, not its prices, so it produces no price build-up. The only money figure is an estimated markup revenue computed from the proposal's own inputs. Assumptions: every rate floor, spread limit, approval band, device rule and channel share limit is a placeholder chosen for this demonstration and is not any provider's policy.

The AI sales agentBuilds the merchant proposal from the conversation and says it is ready.
Predictika validatorChecks every rule at once and returns the verdict, each finding, and the computed figures.

7 scenarios in 3 groups · each group runs from the agent’s first attempt to the outcome · click one to switch

What the agent proposes

The AI sales agent says

Illustrative agent statement. The validator receives the structured request below.

Proposing agent: AI sales agent (scripted)

Before sending, the agent must call Recorded run

Predictika validator


      
Waiting for the call.

Where rules-based CPQ gets it wrong

Click to expand

Our assessment of what a typical rules-based CPQ (product rules, price rules and validation rules) does with the same cases. It is a judgment about effort and reliability, not a claim that nothing can be built.

What the validator caughtTypical rules-based CPQPredictika
Plan and device eligibility that depends on the merchant segment and the device count togetherOnly with care
Each combination becomes a product rule. It works, but the table of allowed combinations grows with every new device or segment and has to be kept in step by hand.
The segment, plan and device quantities are all inputs to one model, and every combination is checked on every call.
A markup floor that depends on monthly volume and on the sales channelOnly with workarounds
The floor is a lookup on two fields, with a channel add-on. It usually ends up as a price-rule table or several rules that must be edited together.
The floor is computed from the volume band and the channel record, and the markup is compared with it, with a separate approval band near the floor.
The proposal states a monthly revenue figure that must match the inputsOnly with care
A validation rule can compare a stated figure with a calculated field, but only if the stated figure is a field on the proposal, which it rarely is.
The stated figure travels with the proposal and is compared with the computed one on every call.
Tiered rates that must rise in order and stay within a spreadOnly with care
Two simple rules, but they only fire when all three rates are filled in, and they interact with the volume and risk limits for tiered pricing.
Order, spread, volume ceiling and high-risk exclusion are all checked together.
A partner revenue share with a hard limit and a lower approval threshold, both per channelEasy
A lookup table and two comparisons.
The same, with the thresholds held in the channel record.
Every problem returned in one answerOnly with workarounds
Rules tend to fire as a step is completed or the quote is saved, so a rep often fixes one error and meets the next.
All findings come back together, whatever order the choices arrived in.

What this validator catches: its rules in brief

Click to expand

Every rule is checked on every call, in any order the choices arrive; one problem never hides another. Blocking findings stop the proposal; approval findings need a person to sign off. All limits are placeholders.

1Block: the proposal has no counter device (Station or Mini).
2Block: a mobile reader is not a primary device for counter, retail or full service merchants; the Kiosk is offered for counter service and retail only, and needs the Standard or Advanced plan.
3Block: the Starter plan covers up to 2 devices; full service dining with 5 or more devices needs the Advanced plan.
4Block: more than 25 devices is an enterprise deal, not a standard proposal.
5Block: interchange-plus markup below the floor for the monthly volume and channel; per-transaction fee below the minimum.
6Block: tiered rates must rise from qualified to mid-qualified to non-qualified, and the spread may not exceed the limit.
7Block: tiered pricing above the volume ceiling, or for a high-risk merchant.
8Block: card-not-present share above the limit; term above the maximum; early termination fee on a month-to-month agreement or above the cap.
9Block: partner revenue share above the channel maximum.
10Block: the stated device count or stated markup revenue does not match what the validator computes.
11Approval: more than 10 devices needs sales management sign-off.
12Approval: markup within the approval band of the floor needs pricing desk approval; a low qualified rate needs pricing desk approval.
13Approval: a high-risk merchant, or a high card-not-present share, needs underwriting review.
14Approval: a long term on low volume, or a partner share above a rep's authority.

Scenarios and results

Click to expand

The recorded result of the validator for each scenario on this page: the request the AI agent sent, the verdict, and every finding, in the order they appear above.

1. Plan, term and markup together: Full service restaurant on the Standard plan with a long term, markup under the floor and revenue overstated

The agent's request. Please quote a full service dining merchant on the Standard plan through the direct sales channel, with 2 Station, 4 Flex devices (6 in total). Monthly card volume is $300,000 with an average ticket of $45, on a 48 month term. Pricing is interchange-plus with a markup of 18 basis points and 10 cents per transaction. The proposal states 6 devices and estimated monthly markup revenue of $1,500.00. It is ready to send.

Verdict: Blocked.

  • Plan too small for the device count. Full service dining with 6 devices needs the Advanced plan; the Standard plan was selected.
  • Markup below the floor. Interchange-plus markup of 18 bps is below the 20 bps floor for $300,000 monthly volume through the Direct sales channel.
  • Term longer than the maximum. A 48 month term is longer than the 36 month maximum.
  • Stated revenue does not match. The proposal states $1,500.00 monthly markup revenue; the validator computes $1,206.60.

2. Plan, term and markup together: Advanced plan and a 36 month term, markup raised but still close to the floor

The agent's request. Please quote a full service dining merchant on the Advanced plan through the direct sales channel, with 2 Station, 4 Flex devices (6 in total). Monthly card volume is $300,000 with an average ticket of $45, on a 36 month term. Pricing is interchange-plus with a markup of 22 basis points and 10 cents per transaction. The proposal states 6 devices and estimated monthly markup revenue of $1,326.60. It is ready to send.

Verdict: Needs approval.

  • Markup needs pricing desk approval. Markup of 22 bps is within 10 bps of the 20 bps floor and needs pricing desk approval.
  • Proposal summary. 6 devices on the Advanced plan for Full service dining; about 6,666 transactions a month; interchange-plus markup floor 20 bps; estimated monthly markup revenue $1,326.60; no price list is modeled.

3. Plan, term and markup together: The agent resubmits after fixing the errors

The agent's request. Please quote a full service dining merchant on the Advanced plan through the direct sales channel, with 2 Station, 4 Flex devices (6 in total). Monthly card volume is $300,000 with an average ticket of $45, on a 36 month term. Pricing is interchange-plus with a markup of 30 basis points and 10 cents per transaction. The proposal states 6 devices and estimated monthly markup revenue of $1,566.60. It is ready to send.

Verdict: Allowed.

  • Proposal summary. 6 devices on the Advanced plan for Full service dining; about 6,666 transactions a month; interchange-plus markup floor 20 bps; estimated monthly markup revenue $1,566.60; no price list is modeled.

4. Markup floor and stated revenue: Markup under the floor and revenue overstated

The agent's request. Please quote a retail merchant on the Standard plan through the bank partner channel, with 3 Station, 1 Mini, 2 Flex devices (6 in total). Monthly card volume is $450,000 with an average ticket of $60, on a 36 month term. Pricing is interchange-plus with a markup of 22 basis points and 10 cents per transaction. The proposal states 6 devices and estimated monthly markup revenue of $2,325.00. It is ready to send.

Verdict: Blocked.

  • Markup below the floor. Interchange-plus markup of 22 bps is below the 25 bps floor for $450,000 monthly volume through the Bank partner channel.
  • Stated revenue does not match. The proposal states $2,325.00 monthly markup revenue; the validator computes $1,740.00.

5. Markup floor and stated revenue: Markup raised, still close to the floor

The agent's request. Please quote a retail merchant on the Standard plan through the bank partner channel, with 3 Station, 1 Mini, 2 Flex devices (6 in total). Monthly card volume is $450,000 with an average ticket of $60, on a 36 month term. Pricing is interchange-plus with a markup of 28 basis points and 10 cents per transaction. The proposal states 6 devices and estimated monthly markup revenue of $2,010.00. It is ready to send.

Verdict: Needs approval.

  • Markup needs pricing desk approval. Markup of 28 bps is within 10 bps of the 25 bps floor and needs pricing desk approval.
  • Proposal summary. 6 devices on the Standard plan for Retail; about 7,500 transactions a month; interchange-plus markup floor 25 bps; estimated monthly markup revenue $2,010.00; no price list is modeled.

6. Markup floor and stated revenue: The agent resubmits after fixing the errors

The agent's request. Please quote a retail merchant on the Standard plan through the bank partner channel, with 3 Station, 1 Mini, 2 Flex devices (6 in total). Monthly card volume is $450,000 with an average ticket of $60, on a 36 month term. Pricing is interchange-plus with a markup of 40 basis points and 10 cents per transaction. The proposal states 6 devices and estimated monthly markup revenue of $2,550.00. It is ready to send.

Verdict: Allowed.

  • Proposal summary. 6 devices on the Standard plan for Retail; about 7,500 transactions a month; interchange-plus markup floor 25 bps; estimated monthly markup revenue $2,550.00; no price list is modeled.

7. Tiered rate spread: Non-qualified rate too far above qualified

The agent's request. Please quote a counter service merchant on the Standard plan through the ISO partner channel, with 5 Station, 3 Mini devices (8 in total). Monthly card volume is $250,000 with an average ticket of $30, on a 24 month term. Pricing is tiered at 180, 250 and 340 basis points for qualified, mid-qualified and non-qualified cards, with a 45 percent partner revenue share. The proposal states 8 devices. It is ready to send.

Verdict: Blocked.

  • Tiered rates spread too wide. The non-qualified rate of 340 bps is more than 150 bps above the qualified rate of 180 bps.

8. Tiered rate spread: The agent resubmits after fixing the errors

The agent's request. Please quote a counter service merchant on the Standard plan through the ISO partner channel, with 5 Station, 3 Mini devices (8 in total). Monthly card volume is $250,000 with an average ticket of $30, on a 24 month term. Pricing is tiered at 180, 250 and 320 basis points for qualified, mid-qualified and non-qualified cards, with a 45 percent partner revenue share. The proposal states 8 devices. It is ready to send.

Verdict: Allowed.

  • Proposal summary. 8 devices on the Standard plan for Counter service; about 8,333 transactions a month; tiered rates of 180, 250 and 320 bps for qualified, mid-qualified and non-qualified cards; no price list is modeled.

Narration transcript

Click to expand

The text of the narrated walkthrough, step by step.

  1. Welcome to a demo of Predictika's Validator. An AI sales agent prepares a card-acceptance proposal for a merchant: business type, plan, devices, and more. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the plan, device, pricing and approval rules and returns a verdict, every reason and the figures it computed. There is no price list in this model.
  2. I will start with the first request by the agent.
  3. The agent says. Please quote a full service dining merchant on the Standard plan through the direct sales channel, with 2 Station, 4 Flex devices (6 in total). Monthly card volume is $300,000 with an average ticket of $45, on a 48 month term. The rest of the request is on screen. The proposal states 6 devices and estimated monthly markup revenue of $1,500.00.
  4. Before it goes out, the agent calls the validator.
  5. Blocked, with four findings, all in one answer. Full service dining with 6 devices needs the Advanced plan; the Standard plan was selected. Interchange-plus markup of 18 bps is below the 20 bps floor for $300,000 monthly volume through the direct sales channel. A 48 month term is longer than the 36 month maximum. The other one is on screen.
  6. The validator returned four blocking findings in one answer. A check that reads one field at a time would pass most of this; it shows only when the choices in the proposal are checked against each other.
  7. What this shows: for plan and device eligibility that depends on the merchant segment and the device count together, a rules-based CPQ is error-prone for this.
  8. Now the agent resubmits.
  9. The corrected request. It changes the plan to advanced, the term to 36 months, the markup to 22 bps and the stated monthly markup revenue to $1,326.60.
  10. It calls the validator again.
  11. Not blocked, but held for a person.
  12. The four blocking findings are cleared; one approval remains for a person: markup needs pricing desk approval.
  13. The agent resubmits once more.
  14. The revised request. It changes the markup to 30 bps and the stated monthly markup revenue to $1,566.60.
  15. It calls the validator once more.
  16. This time the validator clears it.
  17. The earlier finding is cleared, and the validator's total, $1,566.60, matches the proposal.
  18. Four problems caught in one call, fixed in two passes, and a clean proposal. That is the value of Predictika's validator: catching mistakes before they reach the merchant.