Predictika · Financial data subscriptions

An AI sales agent's S&P Capital IQ Pro subscription order, checked before it goes out

An AI sales agent prepares a subscription order for S&P Capital IQ Pro for a fictional customer. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the subscription and returns a verdict, every reason and the figures it computed.

Scope: S&P Capital IQ Pro subscriptions: seats, term, desktop or API delivery, data sets, sector data, trial and discount. See everything this validator covers

This validator models subscription orders for S&P Capital IQ Pro, not just the scenarios shown below. The product, data set and sector names, the separate API licensing and the two week trial come from public S&P Global pages. The remaining limits (seat bounds, discount thresholds and approvals) are modeled from comparable data subscriptions and are not S&P Global published rules.

  • Seats and term — seat count, one, two or three year term, and very large seat counts (modeled).
  • Delivery — desktop or desktop plus API, with API licenses ordered separately from the desktop.
  • Data sets — public company financials plus estimates, ownership, transactions, Trucost and ESG, supply chain data (Panjiva), credit ratings (RatingsDirect) and private company data, and the count the agent states.
  • Sector data — financial institutions, insurance, energy, real estate, metals and mining, and TMT with Kagan research (modeled bundle limit).
  • Trial and discount — the two week trial and the discount allowed without sign-off for each term (modeled), and customer type (investment firm, corporate, government, university).

Any order for this subscription that touches these rules is in scope — not only the scenarios pictured here. Not covered: prices (S&P Global does not publish a price list, and none are modeled, so this validator produces no price build-up or line-item quote), other S&P Global products, and any change S&P Global has made to the subscription.

The AI sales agentBuilds the subscription 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.

8 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
API delivery ordered without API licensesOnly with care
The delivery choice and the license count are separate fields, so a rule has to tie them together.		Delivery and API licenses are in one request and compared on every call.
A discount above what the term allows without sign-offOnly with care
The limit depends on the term, so each term needs its own threshold rule and its own approval route.		The limit is computed from the term and compared with the discount on every call; the finding shows both numbers.
A data set count stated by the agent that does not match the selectionOnly with workarounds
The stated figure is usually free text or a separate field, so it takes an extra rule that adds up the selection and compares.		The stated figure is part of the request and is compared with the computed count on every call.
Items that need confirmation for a customer typeOnly with care
Each customer type has its own list of items needing a person, spread across many rule rows.		All customer type rules are checked together on every call, in whatever order the choices arrived.
Every problem returned in one answerOnly with workarounds
Rules tend to fire as 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 order has 0 Capital IQ Pro seats; a subscription needs at least one seat.
2501 seats is an enterprise sized order, and a person should confirm the licensing terms.
3The order is for desktop plus API delivery but has no API licenses; API access is licensed separately from the desktop.
4The order has 3 API licenses but names desktop only delivery.
5The discount of 6 percent is above the 5 percent the model allows without sign-off for this term.
6The discount of 11 percent is above the 10 percent the model allows without sign-off for this term.
7The discount of 16 percent is above the 15 percent the model allows without sign-off for this term.
8The discount of 30 percent is above the 15 percent the model allows without sign-off for this term.
9The model does not allow a discount of 31 percent; the maximum it considers is 30 percent.
10The order states 5 data sets in all, but the selection adds up to 3, counting public company financials.
11The order states 3 data sets in all, but the selection adds up to 4, counting public company financials.
12The order asks for a 4 week trial, and the public offer is a two week free trial.
13The order adds 4 sector data sets, and a person should confirm the bundle.
14API delivery for a university should be confirmed.
15Credit ratings data through RatingsDirect for a university should be confirmed.
16The order asks for a 6 week trial, and the public offer is a two week free trial.
17The model does not allow a discount of 40 percent; the maximum it considers is 30 percent.
18The order states 9 data sets in all, but the selection adds up to 3, counting public company financials.
19The discount of 20 percent is above the 5 percent the model allows without sign-off for this term.
20600 seats is an enterprise sized order, and a person should confirm the licensing terms.
21The order asks for a 3 week trial, and the public offer is a two week free trial.

  

  

    
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. Asset manager: 25 seats with API delivery, no API licenses and a 35 percent discount

    
The agent's request. Please quote S&P Capital IQ Pro for an investment firm: 25 seats on a 3 year term, desktop plus API with 0 API licenses, public company financials, estimates, ownership, transactions, and a 35 percent discount. The order states 5 data sets in all. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • API delivery without API licenses. The order is for desktop plus API delivery but has no API licenses; API access is licensed separately from the desktop.
    • 
      
  • Discount above the model maximum. The model does not allow a discount of 35 percent; the maximum it considers is 30 percent.
    • 
      
  • Stated data set count does not match. The order states 5 data sets in all, but the selection adds up to 4, counting public company financials.
    • 
    

    
2. Asset manager: The agent adds API licenses and brings the discount down

    
The agent's request. Please quote S&P Capital IQ Pro for an investment firm: 25 seats on a 3 year term, desktop plus API with 3 API licenses, public company financials, estimates, ownership, transactions, and a 15 percent discount. The order states 4 data sets in all. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Order summary. 25 seats, 4 data sets counting public company financials, 0 sector data sets; discount limit without sign-off 15 percent for this term.
    • 
    

    
3. University: 40 seats with ratings data, API and a four week trial

    
The agent's request. Please quote S&P Capital IQ Pro for a university: 40 seats on a 1 year term, desktop plus API with 0 API licenses, public company financials, estimates, credit ratings (RatingsDirect), a 4 week trial, and a 12 percent discount. The order states 2 data sets in all. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • API delivery without API licenses. The order is for desktop plus API delivery but has no API licenses; API access is licensed separately from the desktop.
    • 
      
  • Stated data set count does not match. The order states 2 data sets in all, but the selection adds up to 3, counting public company financials.
    • 
      
  • Discount above the automatic limit. The discount of 12 percent is above the 5 percent the model allows without sign-off for this term.
    • 
      
  • Trial longer than two weeks. The order asks for a 4 week trial, and the public offer is a two week free trial.
    • 
      
  • API for a university. API delivery for a university should be confirmed.
    • 
      
  • Ratings data for a university. Credit ratings data through RatingsDirect for a university should be confirmed.
    • 
    

    
4. University: The agent resubmits after fixing the errors

    
The agent's request. Please quote S&P Capital IQ Pro for a university: 40 seats on a 1 year term, desktop plus API with 2 API licenses, public company financials, estimates, credit ratings (RatingsDirect), a 4 week trial, and a 12 percent discount. The order states 3 data sets in all. It is ready to send.

    
Verdict: Needs approval.

    
    
      
  • Discount above the automatic limit. The discount of 12 percent is above the 5 percent the model allows without sign-off for this term.
    • 
      
  • Trial longer than two weeks. The order asks for a 4 week trial, and the public offer is a two week free trial.
    • 
      
  • API for a university. API delivery for a university should be confirmed.
    • 
      
  • Ratings data for a university. Credit ratings data through RatingsDirect for a university should be confirmed.
    • 
      
  • Order summary. 40 seats, 3 data sets counting public company financials, 0 sector data sets; discount limit without sign-off 5 percent for this term.
    • 
    

    
5. University: The agent drops the API and ratings data and shortens the trial

    
The agent's request. Please quote S&P Capital IQ Pro for a university: 40 seats on a 1 year term, desktop delivery, public company financials, estimates, a 2 week trial, and a 5 percent discount. The order states 2 data sets in all. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Order summary. 40 seats, 2 data sets counting public company financials, 0 sector data sets; discount limit without sign-off 5 percent for this term.
    • 
    

    
6. Corporate: 600 seats with four sector data sets and a wrong data set count

    
The agent's request. Please quote S&P Capital IQ Pro for a corporate: 600 seats on a 2 year term, desktop delivery, public company financials, Trucost and ESG, supply chain data (Panjiva), sector data for energy, real estate, metals and mining, TMT with Kagan, and a 10 percent discount. The order states 2 data sets in all. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Stated data set count does not match. The order states 2 data sets in all, but the selection adds up to 3, counting public company financials.
    • 
      
  • Very large seat count. 600 seats is an enterprise sized order, and a person should confirm the licensing terms.
    • 
      
  • More than three sector data sets. The order adds 4 sector data sets, and a person should confirm the bundle.
    • 
    

    
7. Corporate: The agent corrects the data set count

    
The agent's request. Please quote S&P Capital IQ Pro for a corporate: 600 seats on a 2 year term, desktop delivery, public company financials, Trucost and ESG, supply chain data (Panjiva), sector data for energy, real estate, metals and mining, TMT with Kagan, and a 10 percent discount. The order states 3 data sets in all. It is ready to send.

    
Verdict: Needs approval.

    
    
      
  • Very large seat count. 600 seats is an enterprise sized order, and a person should confirm the licensing terms.
    • 
      
  • More than three sector data sets. The order adds 4 sector data sets, and a person should confirm the bundle.
    • 
      
  • Order summary. 600 seats, 3 data sets counting public company financials, 4 sector data sets; discount limit without sign-off 10 percent for this term.
    • 
    

    
8. Corporate: The agent reduces the seats and sector data sets

    
The agent's request. Please quote S&P Capital IQ Pro for a corporate: 450 seats on a 2 year term, desktop delivery, public company financials, Trucost and ESG, supply chain data (Panjiva), sector data for energy, metals and mining, and a 10 percent discount. The order states 3 data sets in all. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Order summary. 450 seats, 3 data sets counting public company financials, 2 sector data sets; discount limit without sign-off 10 percent for this term.
    • 
    

  

  

    
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 subscription order for S&P Capital IQ Pro for a fictional customer. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the subscription 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 S&P Capital IQ Pro for an investment firm: 25 seats on a 3 year term, desktop plus API with 0 API licenses. The order states 5 data sets in all. The rest of the request is on screen.
    4. 
      
  4. Before it goes out, the agent calls the validator.
    5. 
      
  5. Blocked, with three findings, all in one answer. The order is for desktop plus API delivery but has no API licenses; API access is licensed separately from the desktop. The model does not allow a discount of 35 percent; the maximum it considers is 30 percent. The order states 5 data sets in all, but the selection adds up to 4, counting public company financials.
    6. 
      
  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. 
      
  7. What this shows: for a data set count stated by the agent that does not match the selection, a rules-based CPQ can do this only with workarounds.
    8. 
      
  8. Now the agent resubmits.
    9. 
      
  9. The corrected request. It changes the delivery to Desktop plus API (3 API licenses), the discount to 15 percent and the stated data sets, as to 4.
    10. 
      
  10. It calls the validator again.
    11. 
      
  11. This time the validator clears it.
    12. 
      
  12. All three earlier findings are cleared, and the validator's figure matches the quote.
    13. 
      
  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.
    14.