Predictika · Workplace supplies orders

An AI sales agent's workplace supplies order, checked before it goes out

An AI sales agent prepares an order for workplace supplies: A4 copy paper, HP LaserJet Pro toner for a named printer and hi-vis vests, with a delivery option, the order time, a discount, payment terms and the total it states in pence. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the order rules and returns a verdict, every reason and the totals it computed.

Scope: Lyreco UK workplace-supplies orders: copy paper, HP toner for named printers and hi-vis vests, delivery option, order time, discount tier, payment terms, VAT and the stated total. See everything this validator covers

This validator models workplace-supplies orders of this kind, not just the scenarios shown below.

  • Items and printer fit — paper by the box, hi-vis vests, and the HP toner cartridge that fits each of four LaserJet Pro printer models.
  • Delivery and timing — next-day, express, pallet and scheduled delivery, the order cut-off, the full pallet of paper that pallet delivery needs, and who may use scheduled delivery.
  • Discount, credit and total — the discount each customer tier may be given, the level that needs a sales manager, credit terms and the credit limit, VAT, and the stated total compared with the computed one.

Any order of these items that touches these rules is in scope — not only the scenarios pictured here. Lyreco's category names are public; the printer and cartridge fit follows HP's product naming, and every price, the delivery charges, cut-off time, pallet size, discount tiers and credit rules are modeled and not published by Lyreco, so the amounts on this page are illustrative. Not covered: furniture, catering, hygiene, other brands and returns.

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

6 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, as in Salesforce CPQ) 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
Toner that fits only some printersOnly with care
Each cartridge fits two of the four printer models, so every pairing needs its own rule.
The printer and cartridge are fields of one order and are checked together.
A delivery option that depends on the order time, the customer tier and the paper countOnly with workarounds
Next-day cut-off, a full pallet of paper and a contract for scheduled delivery each depend on other choices on the order.
Delivery, order time, tier and quantities are in one request and every condition is evaluated on every call.
Discount limits that depend on the tier and the size of the orderOnly with workarounds
The limit is none for a standard customer, 5 percent for a contract and 10 percent on a large order, with approval above it and a hard stop at 15 percent.
The goods value sets the limit on every call; the discount is compared with it and with the hard stop.
A total built from many partsOnly with workarounds
The total depends on the items, a discount rounded down to the penny, delivery that is free over a net goods value, and VAT, so it must be computed before it can be compared with the total stated.
The total is computed on every call and compared with the total stated in the order.
Every problem returned in one answerOnly with workarounds
Rules tend to fire as a step is saved or completed, so a person often fixes one error and meets the next.
All findings come back together, whatever order the choices arrived in.

What this validator catches: findings from its test cases

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 order; approval findings need a person to sign off; everything else is computed.

1Pallet delivery needs at least 20 boxes of paper; the order has 19.
2An order placed at hour 17 misses the 17:00 cut-off for next-day or express delivery.
3Scheduled standing-order delivery is only offered to contract customers.
4The order has no items.
5Toner is ordered but no toner cartridge model is chosen.
6The chosen toner cartridge does not fit the chosen printer model.
7A standard customer without a contract pays pro-forma; credit terms are not available.
8A discount of 1 percent is above the 0 percent tier limit; a sales manager should approve it.
9A discount of 6 percent is above the 5 percent tier limit; a sales manager should approve it.
10A discount of 11 percent is above the 10 percent tier limit; a sales manager should approve it.
11A discount of 15 percent is above the 10 percent tier limit; a sales manager should approve it.
12A discount of 16 percent is outside the 0 to 15 percent range that can be approved.
13Sixty-day payment terms need credit control approval.
14An order of 500,256 pence on thirty-day terms is above the credit limit that can be released without approval.
15The order states a total of 127,238 pence; the modeled prices give 127,338 pence.
16The order states a total of 127,339 pence; the modeled prices give 127,338 pence.
17The order states a total of 100 pence; the modeled prices give 127,338 pence.
18Pallet delivery needs at least 20 boxes of paper; the order has 5.
19A discount of 20 percent is outside the 0 to 15 percent range that can be approved.
20An order placed at hour 18 misses the 17:00 cut-off for next-day or express delivery.
21Pallet delivery needs at least 20 boxes of paper; the order has 12.
22A discount of 12 percent is above the 5 percent tier limit; a sales manager should approve it.
23A discount of 8 percent is above the 0 percent tier limit; a sales manager should approve it.

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. Toner, cut-off and discount: An order with the wrong toner for the printer, placed after the cut-off, with a large discount and a total from an earlier draft

The agent's request. Please quote 30 boxes of A4 copy paper, 6 HP 59A toner cartridges for the M402 printers, 20 hi-vis vests on a contract account, with next-day delivery, placed at 18:00 on thirty-day terms. The discount is 20 percent. The order states a total of 127,338 pence. It is ready to send.

Verdict: Blocked.

  • Toner does not fit the printer. The chosen toner cartridge does not fit the chosen printer model.
  • Order placed after the next-day cut-off. An order placed at hour 18 misses the 17:00 cut-off for next-day or express delivery.
  • Discount outside the allowed range. A discount of 20 percent is outside the 0 to 15 percent range that can be approved.
  • Stated total does not match. The order states a total of 127,338 pence; the modeled prices give 107,232 pence.

2. Toner, cut-off and discount: The agent resubmits after fixing the errors

The agent's request. Please quote 30 boxes of A4 copy paper, 6 HP 26A toner cartridges for the M402 printers, 20 hi-vis vests on a contract account, with next-day delivery, placed at 15:00 on thirty-day terms. The discount is 5 percent. The order states a total of 125,970 pence. It is ready to send.

Verdict: Allowed.

  • Order summary. 56 item(s), goods 110,500 pence, discount 5,525 pence, delivery 0 pence, total 125,970 pence including VAT.

3. New customer terms and delivery: A new customer without a contract asks for credit terms, scheduled delivery and a discount

The agent's request. Please quote 30 boxes of A4 copy paper, 6 HP 59A toner cartridges for the M404 printers, 20 hi-vis vests on a standard account, with scheduled standing-order delivery, placed at 10:00 on thirty-day terms. The discount is 3 percent. The order states a total of 130,018 pence. It is ready to send.

Verdict: Blocked.

  • Scheduled delivery needs a contract. Scheduled standing-order delivery is only offered to contract customers.
  • Credit terms need a contract. A standard customer without a contract pays pro-forma; credit terms are not available.
  • Discount above the tier limit. A discount of 3 percent is above the 0 percent tier limit; a sales manager should approve it.

4. New customer terms and delivery: The agent resubmits after fixing the errors

The agent's request. Please quote 30 boxes of A4 copy paper, 6 HP 59A toner cartridges for the M404 printers, 20 hi-vis vests on a standard account, with next-day delivery, placed at 10:00 on pro-forma terms. The order states a total of 134,040 pence. It is ready to send.

Verdict: Allowed.

  • Order summary. 56 item(s), goods 111,700 pence, discount 0 pence, delivery 0 pence, total 134,040 pence including VAT.

5. Pallet discount and credit terms: A pallet of paper with sixty-day terms and a discount above the tier limit

The agent's request. Please quote 24 boxes of A4 copy paper on a contract account, with pallet delivery, placed at 10:00 on sixty-day terms. The discount is 12 percent. The order states a total of 54,744 pence. It is ready to send.

Verdict: Needs approval.

  • Discount above the tier limit. A discount of 12 percent is above the 5 percent tier limit; a sales manager should approve it.
  • Sixty-day credit terms. Sixty-day payment terms need credit control approval.
  • Order summary. 24 item(s), goods 51,840 pence, discount 6,220 pence, delivery 0 pence, total 54,744 pence including VAT.

6. Pallet discount and credit terms: The agent resubmits on thirty-day terms at the tier limit

The agent's request. Please quote 24 boxes of A4 copy paper on a contract account, with pallet delivery, placed at 10:00 on thirty-day terms. The discount is 5 percent. The order states a total of 59,097 pence. It is ready to send.

Verdict: Allowed.

  • Order summary. 24 item(s), goods 51,840 pence, discount 2,592 pence, delivery 0 pence, total 59,097 pence including VAT.

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 an order for workplace supplies: A4 copy paper, HP LaserJet Pro toner for a named printer and hi-vis vests. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the order rules and returns a verdict, every reason and the totals it computed.
  2. I will start with the first request by the agent.
  3. The agent says. Please quote 30 boxes of A4 copy paper, 6 HP 59A toner cartridges for the M402 printers, 20 hi-vis vests on a contract account. The discount is 20 percent. The rest of the request is on screen. The order states a total of 127,338 pence.
  4. Before it goes out, the agent calls the validator.
  5. Blocked, with four findings, all in one answer. The chosen toner cartridge does not fit the chosen printer model. An order placed at hour 18 misses the 17:00 cut-off for next-day or express delivery. A discount of 20 percent is outside the 0 to 15 percent range that can be approved. 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 order are checked against each other and against the stated total.
  7. What this shows: for toner that fits only some printers, a rules-based CPQ is error-prone for this.
  8. Now the agent resubmits.
  9. The corrected request. It changes the toner model to HP 26A, the order hour to 15, the discount to 5 percent and the stated total to 125,970.
  10. It calls the validator again.
  11. This time the validator clears it.
  12. All four earlier findings are cleared.
  13. Four problems caught in one call, fixed in one pass, and a clean order. That is the value of Predictika's validator: catching mistakes before they reach the customer. Now you can try the other scenarios yourself.