Predictika · Industrial robots

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

An AI sales agent prepares an order for KR AGILUS small robots from the public KUKA product page: which of the 17 listed variants, which controller, what environment the robot works in, and the load and reach it must handle. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of that page and returns a verdict, every reason and the figures it computed.

Scope: KR AGILUS robots (the 17 variants listed on the product page): variant, controller, environment version, total load and reach. See everything this validator covers

This validator models orders for the KR AGILUS variants listed on the public KUKA product page, not just the scenarios shown below.

  • Variants — KR 6 R700-2, R900-2 and their EP and HO versions; KR 10 R900 and R1100 in CR, EX, WP, HM-SC and the -2 versions, each with its listed total load (6, 10 or 11 kg) and maximum reach (726, 901 or 1101 mm). The load and reach the agent states must equal the listed figures, and the application must not exceed them.
  • Environment versions — cleanroom needs a CR version, a potentially explosive area an EX version, food or pharmaceutical contact an HO version, electrostatic protection an EP version, and a wet area a robot listed with protection class IP 67.
  • Controllers — the KR 10 R900 HM-SC and KR 10 R1100 HM-SC are listed with the KR C5 micro only; the other listed variants with the KR C5 micro or the KR C5 micro-2.

Any order for these robots that touches these rules is in scope — not only the scenarios pictured here. Not covered: prices (KUKA publishes none for these robots, so this validator produces no price build-up or line-item quote), other KUKA robot families, tools, safety equipment, software options and integration. Assumptions: the page lists the same four mounting positions for every variant, so mounting is not a rule; the listed protection class "IP 65, IP 67" is read as IP 67 available; an explosive-area order with an EX robot goes to a person to confirm the zone classification and approvals.

The AI sales agentBuilds the robot 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
A robot version that must match the siteOnly with care
Each environment needs a different version, so a rep has to remember which of the suffixes applies and check it by hand.		The environment is part of the request and is compared with the version of the chosen robot on every call.
A controller listed for only some robotsOnly with care
The compatible controllers differ by variant, so a lookup table has to be kept in step with the catalog.		The listed controllers are derived from the variant and compared with the one chosen.
A stated load and reach compared with the listed figuresOnly with workarounds
The figures typed into the order are free text that has to be compared with a lookup of the chosen variant.		The stated figures are part of the request and are compared with the derived ones on every call.
Every problem returned in one answerOnly with workarounds
Rules tend to fire as the order is saved or a step is completed, 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: 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 tool and workpiece weigh 12 kg; the selected robot is listed with a total load of 11 kg.
2The application needs 902 mm of reach; the selected robot is listed with a maximum reach of 901 mm.
3The tool and workpiece weigh 7 kg; the selected robot is listed with a total load of 6 kg.
4The order states a total load of 10 kg; the selected robot is listed with 11 kg.
5The order states a maximum reach of 1,101 mm; the selected robot is listed with 901 mm.
6A cleanroom application needs a cleanroom (CR) robot version; the selected robot is not one.
7The robot is for a potentially explosive area; a person should confirm the zone classification and the approvals needed before the order goes out.
8A potentially explosive area needs an explosion-protected (EX) robot version; the selected robot is not one.
9A food or pharmaceutical contact application needs a hygienic (HO) robot version; the selected robot is not one.
10An application that needs electrostatic protection needs an electrostatic protected (EP) robot version; the selected robot is not one.
11A wet area needs a robot listed with protection class IP 67; the selected robot is listed with IP 54 only.
12The selected robot is listed with the KR C5 micro controller only, not the KR C5 micro-2.
13The order has no robots.
14The tool and workpiece weigh 9 kg; the selected robot is listed with a total load of 6 kg.
15The application needs 1,000 mm of reach; the selected robot is listed with a maximum reach of 901 mm.
16The tool and workpiece weigh 11 kg; the selected robot is listed with a total load of 10 kg.

  

  

    
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. Load and reach: Six KR 6 R700-2 robots for an 8 kg load at 800 mm

    
The agent's request. Please quote 6 KR 6 R700-2 robots on a KR C5 micro controller, working in a dry, standard environment, for a tool and workpiece of 8 kg at 800 mm of reach. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Total load above the robot rating. The tool and workpiece weigh 8 kg; the selected robot is listed with a total load of 6 kg.
    • 
      
  • Reach shorter than needed. The application needs 800 mm of reach; the selected robot is listed with a maximum reach of 726 mm.
    • 
    

    
2. Load and reach: The agent resubmits after fixing the errors

    
The agent's request. Please quote 6 KR 10 R900-2 robots on a KR C5 micro controller, working in a dry, standard environment, for a tool and workpiece of 8 kg at 800 mm of reach. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Order summary. 6 robots, each with a total load of 11 kg and a maximum reach of 901 mm.
    • 
    

    
3. Environment and controller: Three KR 10 R1100 HM-SC robots for food contact on a micro-2 controller

    
The agent's request. Please quote 3 KR 10 R1100 HM-SC robots on a KR C5 micro-2 controller, working in food contact, for a tool and workpiece of 9 kg at 1,000 mm of reach. The order states a total load of 10 kg and a maximum reach of 1,101 mm. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Food contact needs the hygienic version. A food or pharmaceutical contact application needs a hygienic (HO) robot version; the selected robot is not one.
    • 
      
  • Controller not listed for this robot. The selected robot is listed with the KR C5 micro controller only, not the KR C5 micro-2.
    • 
    

    
4. Environment and controller: The agent resubmits after fixing the errors

    
The agent's request. Please quote 3 KR 10 R1100-2 HO robots on a KR C5 micro-2 controller, working in food contact, for a tool and workpiece of 9 kg at 1,000 mm of reach. The order states a total load of 11 kg and a maximum reach of 1,101 mm. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Order summary. 3 robots, each with a total load of 11 kg and a maximum reach of 1,101 mm.
    • 
    

    
5. Wet area: Four KR 10 R900 CR robots in a wet area

    
The agent's request. Please quote 4 KR 10 R900 CR robots on a KR C5 micro controller, working in a wet area, for a tool and workpiece of 9 kg at 850 mm of reach. The order states a total load of 11 kg and a maximum reach of 901 mm. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Stated load rating does not match the robot. The order states a total load of 11 kg; the selected robot is listed with 10 kg.
    • 
      
  • Wet area needs IP 67. A wet area needs a robot listed with protection class IP 67; the selected robot is listed with IP 54 only.
    • 
    

    
6. Wet area: The agent resubmits after fixing the errors

    
The agent's request. Please quote 4 KR 10 R900 WP robots on a KR C5 micro controller, working in a wet area, for a tool and workpiece of 9 kg at 850 mm of reach. The order states a total load of 11 kg and a maximum reach of 901 mm. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Stated load rating does not match the robot. The order states a total load of 11 kg; the selected robot is listed with 10 kg.
    • 
    

    
7. Wet area: The agent resubmits after fixing the errors

    
The agent's request. Please quote 4 KR 10 R900 WP robots on a KR C5 micro controller, working in a wet area, for a tool and workpiece of 9 kg at 850 mm of reach. The order states a total load of 10 kg and a maximum reach of 901 mm. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Order summary. 4 robots, each with a total load of 10 kg and a maximum reach of 901 mm.
    • 
    

  

  

    
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 KR AGILUS small robots from the public KUKA product page: which of the 17 listed variants, which controller, what environment the robot works in, and more. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of that page and returns a verdict, every reason and the figures it computed.
    2. 
      
  2. I will start with the first request by the agent.
    3. 
      
  3. The agent says. Please quote 4 KR 10 R900 CR robots on a KR C5 micro controller, working in a wet area, for a tool and workpiece of 9 kg at 850 mm of reach. The order states a total load of 11 kg and a maximum reach of 901 mm. The rest of the request is on screen.
    4. 
      
  4. Before it goes out, the agent calls the validator.
    5. 
      
  5. Blocked, with two findings, all in one answer. The order states a total load of 11 kg; the selected robot is listed with 10 kg. A wet area needs a robot listed with protection class IP 67; the selected robot is listed with IP 54 only.
    6. 
      
  6. The validator returned two 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. 
      
  7. What this shows: for a robot version that must match the site, a rules-based CPQ is error-prone for this.
    8. 
      
  8. Now the agent resubmits.
    9. 
      
  9. The corrected request. It changes the robot to KR 10 R900 WP.
    10. 
      
  10. It calls the validator again.
    11. 
      
  11. Still blocked, with one finding, all in one answer. The order states a total load of 11 kg; the selected robot is listed with 10 kg.
    12. 
      
  12. One finding cleared. Still blocking: stated load rating does not match the robot.
    13. 
      
  13. The agent resubmits once more.
    14. 
      
  14. The revised request. It changes the stated total load to 10 kg.
    15. 
      
  15. It calls the validator once more.
    16. 
      
  16. This time the validator clears it.
    17. 
      
  17. The earlier finding is cleared.
    18. 
      
  18. Two problems caught in one call, fixed in two passes, 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.
    19.