Predictika · Medical equipment

An AI sales agent's barrier-free exam chair order, checked before it goes out

An AI sales agent prepares an order for a clinic: a number of Midmark barrier-free power exam chairs of one model, the seat height, the heaviest patient, the optional features the clinic needs and the weight capacity it tells the buyer the chair has. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the public Midmark barrier-free model matrix and returns a verdict, every reason and the figures it computed.

Scope: Midmark 626, 229 and 228 barrier-free chairs: model, uncompressed seat height, nine optional features and weight capacity. See everything this validator covers

This validator models one barrier-free chair order at a time, based on the public Midmark barrier-free model comparison matrix, not just the scenarios shown below.

  • Models — 626-001 to 626-006, 229-002, 229-003 and 228-001 to 228-003, each with the seat height, features and weight capacity the matrix lists.
  • Seat height — 15.5 inch on the 626 models and 17 inch on the 229 and 228 models; a different seat height than the selected model lists is blocked.
  • Nine optional features — powered back, clean assist roller system, active sensing technology, lockout hand and foot controls, heated upholstery, integrated digital scale, standard rail accessory receivers, tamper-resistant electrical receptacles, and drawer heater with pelvic tilt. A feature the order needs that the selected model does not list is blocked.
  • Weight capacity — the heaviest patient must not exceed the model capacity (650 pounds on 626, 500 on 229 and 228), and the capacity the order states to the buyer must equal the matrix figure.
  • Matrix twins — the matrix shows no difference between 626-003 and 626-005, or between 626-004 and 626-006, so those orders go to a person to confirm the model number with Midmark.

Any order for these models that touches these rules is in scope — not only the scenarios pictured here. Not covered: prices (none are public), accessories and upholstery colors, delivery and installation, other Midmark chairs, and clinical suitability.

The AI sales agentBuilds the chair 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.

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, 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
Features that exist only on some modelsOnly with workarounds
Each of the twelve models lists a different subset of nine features, so a rules-based CPQ needs a model-by-feature table or a rule per feature, edited whenever the matrix changes.
Model and needed features arrive in one request and are compared with the published matrix row on every call.
Seat height tied to the model familyOnly with care
The seat height depends on the family chosen elsewhere in the order, so the check is a lookup against another choice.
The model row supplies the seat height and the order is compared with it, in any order of choices.
Patient weight against model capacityOnly with care
The capacity differs by family, and the heaviest patient and the capacity quoted to the buyer are both compared with it.
Both the patient weight and the stated capacity are compared with the matrix figure on every call, including at the exact limit.
Models the matrix does not tell apartOnly with workarounds
Two pairs of 626 models show identical rows, which a rules engine either ignores or hard-codes.
The pair is flagged for a person to confirm the model number, while the other checks still run.
Every problem returned in one answerOnly with workarounds
Rules tend to fire as the order is saved or a step is 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 quote; approval findings need a person to sign off; everything else is computed.

1The order needs a chair for patients up to 651 pounds; the selected model is listed for 650 pounds.
2The order needs a chair for patients up to 501 pounds; the selected model is listed for 500 pounds.
3The order has 0 chairs; at least 1 is needed.
4The order needs a different uncompressed seat height (15.5 inch or 17 inch) than the selected model lists.
5The order needs Heated Upholstery; the selected model does not list it.
6The order needs an Integrated Digital Scale; the selected model does not list it.
7The model matrix shows no difference between 626-003 and 626-005, or between 626-004 and 626-006; a person should confirm the model number with Midmark.
8The order needs Powered Back; the selected model does not list it.
9The order needs the Clean Assist Roller System; the selected model does not list it.
10The order needs Active Sensing Technology; the selected model does not list it.
11The order needs Lockout Hand and Foot Controls; the selected model does not list it.
12The order needs Tamper-Resistant Electrical Receptacles; the selected model does not list it.
13The order needs a Drawer Heater and Pelvic Tilt; the selected model does not list it.
14The order states a weight capacity of 700 pounds; the model matrix lists 650 pounds for the selected model.
15The order states a weight capacity of 650 pounds; the model matrix lists 500 pounds for the selected model.

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. Seat height and weight capacity: Six chairs for a bariatric clinic: 15.5 inch seat, heated upholstery, patients up to 600 pounds

The agent's request. Please prepare an order for 6 Midmark 228-003 barrier-free chairs with a 15.5 inch uncompressed seat height, for patients up to 600 pounds, needing heated upholstery. It is ready to send.

Verdict: Blocked.

  • Patient weight above the model capacity. The order needs a chair for patients up to 600 pounds; the selected model is listed for 500 pounds.
  • Seat height not as required. The order needs a different uncompressed seat height (15.5 inch or 17 inch) than the selected model lists.
  • Heated upholstery. The order needs Heated Upholstery; the selected model does not list it.

2. Seat height and weight capacity: The agent resubmits after fixing the errors

The agent's request. Please prepare an order for 6 Midmark 626-002 barrier-free chairs with a 15.5 inch uncompressed seat height, for patients up to 600 pounds, needing heated upholstery. It is ready to send.

Verdict: Allowed.

  • Order summary. 6 chairs of the selected model; it lists 8 of the 9 optional features and a weight capacity of 650 pounds.

3. Features the model does not list: Four chairs with lockout controls, heated upholstery and a drawer heater

The agent's request. Please prepare an order for 4 Midmark 229-003 barrier-free chairs with a any seat height uncompressed seat height, for patients up to 450 pounds, needing lockout hand and foot controls, heated upholstery and drawer heater and pelvic tilt. The order tells the buyer the chair is rated for 500 pounds. It is ready to send.

Verdict: Blocked.

  • Lockout hand and foot controls. The order needs Lockout Hand and Foot Controls; the selected model does not list it.
  • Heated upholstery. The order needs Heated Upholstery; the selected model does not list it.

4. Features the model does not list: The agent resubmits after fixing the errors

The agent's request. Please prepare an order for 4 Midmark 626-004 barrier-free chairs with a 15.5 inch uncompressed seat height, for patients up to 600 pounds, needing lockout hand and foot controls, heated upholstery and drawer heater and pelvic tilt. The order tells the buyer the chair is rated for 650 pounds. It is ready to send.

Verdict: Needs approval.

  • Model rows identical in the matrix. The model matrix shows no difference between 626-003 and 626-005, or between 626-004 and 626-006; a person should confirm the model number with Midmark.
  • Order summary. 4 chairs of the selected model; it lists 9 of the 9 optional features and a weight capacity of 650 pounds.

5. Features the model does not list: The agent resubmits after fixing the errors

The agent's request. Please prepare an order for 4 Midmark 626-002 barrier-free chairs with a 15.5 inch uncompressed seat height, for patients up to 600 pounds, needing lockout hand and foot controls, heated upholstery and drawer heater and pelvic tilt. The order tells the buyer the chair is rated for 650 pounds. It is ready to send.

Verdict: Allowed.

  • Order summary. 4 chairs of the selected model; it lists 8 of the 9 optional features and a weight capacity of 650 pounds.

6. Stated capacity: Ten chairs with a powered back and a stated capacity of 650 pounds

The agent's request. Please prepare an order for 10 Midmark 228-002 barrier-free chairs with a 17 inch uncompressed seat height, for patients up to 450 pounds, needing powered back, active sensing technology and tamper-resistant electrical receptacles. The order tells the buyer the chair is rated for 650 pounds. It is ready to send.

Verdict: Blocked.

  • Powered back. The order needs Powered Back; the selected model does not list it.
  • Stated capacity does not match. The order states a weight capacity of 650 pounds; the model matrix lists 500 pounds for the selected model.

7. Stated capacity: The agent resubmits after fixing the errors

The agent's request. Please prepare an order for 10 Midmark 229-002 barrier-free chairs with a 17 inch uncompressed seat height, for patients up to 450 pounds, needing powered back, active sensing technology and tamper-resistant electrical receptacles. The order tells the buyer the chair is rated for 500 pounds. It is ready to send.

Verdict: Allowed.

  • Order summary. 10 chairs of the selected model; it lists 5 of the 9 optional features and a weight capacity of 500 pounds.

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 a clinic: a number of Midmark barrier-free power exam chairs of one model, the seat height, the heaviest patient, and more. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the public Midmark barrier-free model matrix and returns a verdict, every reason and the figures it computed.
  2. I will start with the first request by the agent.
  3. The agent says. Please prepare an order for 6 Midmark 228-003 barrier-free chairs with a 15.5 inch uncompressed seat height, for patients up to 600 pounds, needing heated upholstery. It is ready to send.
  4. Before it goes out, the agent calls the validator.
  5. Blocked, with three findings, all in one answer. The order needs a chair for patients up to 600 pounds; the selected model is listed for 500 pounds. The order needs a different uncompressed seat height (15.5 inch or 17 inch) than the selected model lists. The order needs Heated Upholstery; the selected model does not list it.
  6. The validator returned three 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 quote are checked against each other.
  7. What this shows: for patient weight against model capacity, a rules-based CPQ is error-prone for this.
  8. Now the agent resubmits.
  9. The corrected request. It changes the model to 626-002.
  10. It calls the validator again.
  11. This time the validator clears it.
  12. All three earlier findings are cleared.
  13. Three problems caught in one call, fixed in one pass, and a clean quote. That is the value of Predictika's validator: catching mistakes before they reach the customer. Now you can try the other scenarios yourself.