Predictika · Cyber insurance

An AI sales agent's cyber insurance quote, checked before it goes out

An AI sales agent prepares a Berkley Cyber Risk Solutions quote for an applicant: the industry class and revenue, the aggregate limit, the retention, the social engineering sublimit, the Technology E&O endorsement, the applicant's security controls and the premium the quote states. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of how those choices fit together and returns a verdict, every reason and the premium it computed.

Scope: One commercial cyber quote: industry class, revenue, limit, retention, social engineering sublimit, endorsement, controls and the stated premium. See everything this validator covers

This validator models Berkley Cyber Risk Solutions quotes, not just the scenarios shown below.

  • Limit and retention — the aggregate limits offered, the largest limit and the smallest retention that each revenue size allows, and the retentions offered.
  • Industry class and revenue — restricted classes that are not written, the range of revenue the model accepts, and the revenue and class combinations that go to an underwriter.
  • Sublimit, records and controls — the social engineering sublimit against the aggregate limit, the number of stored personal records, multi-factor authentication, offline backups and a prior incident.
  • Premium — the premium computed from revenue, class, limit, retention, sublimit and the Technology E&O endorsement, and the stated premium against the one computed.

Any commercial cyber quote that touches these rules is in scope — not only the scenarios pictured here. Not covered: other Berkley products and coverage parts, taxes and fees. The product name and coverage areas follow Berkley's public cyber page; the industry classes, limits, retentions, referral thresholds and every premium rate are modeled from comparable cyber products and are illustrative, not published by Berkley.

The AI sales agentBuilds the cyber quote from the conversation with the broker and says it is ready to send.
Predictika validatorChecks every rule at once and returns the verdict, each finding, and the computed figures.

6 scenarios in 2 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
Limits that depend on the revenue sizeOnly with care
The largest limit and the smallest retention both change with the applicant's revenue, so each size needs its own pair of rules, and a new size means new rules.		The size is looked up from the revenue and both bounds are checked in the same pass.
A sublimit that depends on the aggregate limitOnly with workarounds
The social engineering sublimit must be one of the offered amounts and no more than half the aggregate limit, so its rule reads another field the quote chooses.		The sublimit is checked against the aggregate limit chosen in the same call.
Controls that stop or refer a quoteOnly with workarounds
Missing multi-factor authentication stops the quote, while no offline backups, a prior incident or a large record count only refer it, and the rules for each fire at different times.		Every control is evaluated together, and the answer says which stop the quote and which go to an underwriter.
A stated premium that must match the computed oneOnly with workarounds
The premium depends on revenue, class, limit, retention, sublimit and the endorsement, so a separate rule has to compute it and compare it with the premium the quote states.		The premium is computed inside the model and compared with the stated one.
Every problem returned in one answerOnly with workarounds
Rules tend to fire as the quote is saved or a step is completed, so a broker often fixes one error and meets the next.		All findings come back together, so the agent can fix them in a single pass.

  

  

    
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.

    
1A limit of 5 million dollars exceeds the 3 million maximum for this revenue size.
2A 10 million dollar limit needs capacity confirmed by an underwriter.
3An aggregate limit of 4 million dollars is not one of the 1, 2, 3, 5 or 10 million options.
4A retention of 15 thousand dollars is not one of the 10, 25, 50, 100 or 250 thousand options.
5A retention of 10 thousand dollars is below the 25 thousand minimum for this revenue size.
6Annual revenue of 101 million dollars is above 100 million; an underwriter must review the account.
7Annual revenue of 251 million dollars is outside the 1 to 250 million range of this quote model.
8The applicant is in a restricted industry class that this quote model does not write.
9Multi-factor authentication for remote access, email and privileged accounts is not in place.
10No offline or immutable backups; an underwriter must review the ransomware terms.
11A cyber incident or claim in the last three years; an underwriter must review it.
12A social engineering sublimit of 500 thousand dollars needs an underwriter to confirm.
13A social engineering sublimit of 300 thousand dollars is not offered or exceeds half the aggregate limit.
141,001 thousand stored personal records is above 1,000 thousand; an underwriter must review the data exposure.
155,001 thousand stored personal records is above the 5,000 thousand this quote model accepts.
16An elevated-risk industry with a limit of 5 million dollars or more needs an underwriter to confirm.
17The quote states an annual premium of 20,399 dollars; the rating gives 20,400 dollars.

  

  

    
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. Elevated-risk applicant quote: An elevated-risk quote with a retention below the minimum for the revenue size, no multi-factor authentication and a wrong premium, for an applicant with no offline backups, a large record count, a 10 million dollar limit and a large social engineering sublimit

    
The agent's request. Please quote Berkley Cyber Risk Solutions for a elevated-risk applicant with 60 million dollars of revenue and 1,500,000 personal records. Aggregate limit: 10 million dollars; retention: 25,000 dollars; social engineering sublimit: 1,000,000 dollars. Multi-factor authentication: no; offline backups: no. The quote states an annual premium of 15,000 dollars. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Retention below minimum for revenue. A retention of 25 thousand dollars is below the 50 thousand minimum for this revenue size.
    • 
      
  • No multi-factor authentication. Multi-factor authentication for remote access, email and privileged accounts is not in place.
    • 
      
  • Stated premium does not match. The quote states an annual premium of 15,000 dollars; the rating gives 355,680 dollars.
    • 
      
  • No offline backups. No offline or immutable backups; an underwriter must review the ransomware terms.
    • 
      
  • 10 million limit. A 10 million dollar limit needs capacity confirmed by an underwriter.
    • 
      
  • Large social engineering sublimit. A social engineering sublimit of 1,000 thousand dollars needs an underwriter to confirm.
    • 
      
  • Many records. 1,500 thousand stored personal records is above 1,000 thousand; an underwriter must review the data exposure.
    • 
      
  • Elevated class with a large limit. An elevated-risk industry with a limit of 5 million dollars or more needs an underwriter to confirm.
    • 
    

    
2. Elevated-risk applicant quote: The broker resubmits after fixing the errors

    
The agent's request. Please quote Berkley Cyber Risk Solutions for a elevated-risk applicant with 60 million dollars of revenue and 1,500,000 personal records. Aggregate limit: 5 million dollars; retention: 50,000 dollars; social engineering sublimit: 250,000 dollars. Multi-factor authentication: yes; offline backups: no. The quote states an annual premium of 183,435 dollars. It is ready to send.

    
Verdict: Needs approval.

    
    
      
  • No offline backups. No offline or immutable backups; an underwriter must review the ransomware terms.
    • 
      
  • Many records. 1,500 thousand stored personal records is above 1,000 thousand; an underwriter must review the data exposure.
    • 
      
  • Elevated class with a large limit. An elevated-risk industry with a limit of 5 million dollars or more needs an underwriter to confirm.
    • 
      
  • Quote summary. Estimated annual premium of 183,435 dollars for a 5 million dollar limit with a 50 thousand dollar retention.
    • 
    

    
3. Elevated-risk applicant quote: The broker resubmits after fixing the errors

    
The agent's request. Please quote Berkley Cyber Risk Solutions for a standard applicant with 60 million dollars of revenue and 800,000 personal records. Aggregate limit: 5 million dollars; retention: 50,000 dollars; social engineering sublimit: 250,000 dollars. Multi-factor authentication: yes; offline backups: yes. The quote states an annual premium of 115,853 dollars. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Quote summary. Estimated annual premium of 115,853 dollars for a 5 million dollar limit with a 50 thousand dollar retention.
    • 
    

    
4. Small applicant quote with a 5 million dollar limit: A quote for a small applicant with a limit above what its revenue size allows, a social engineering sublimit that is not offered, more stored records than the model accepts and a wrong premium

    
The agent's request. Please quote Berkley Cyber Risk Solutions for a standard applicant with 8 million dollars of revenue and 6,000,000 personal records. Aggregate limit: 5 million dollars; retention: 10,000 dollars; social engineering sublimit: 300,000 dollars. The Technology E&O endorsement is included. Multi-factor authentication: yes; offline backups: yes. The quote states an annual premium of 3,000 dollars. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Limit too high for revenue. A limit of 5 million dollars exceeds the 3 million maximum for this revenue size.
    • 
      
  • Social engineering sublimit invalid. A social engineering sublimit of 300 thousand dollars is not offered or exceeds half the aggregate limit.
    • 
      
  • Too many records. 6,000 thousand stored personal records is above the 5,000 thousand this quote model accepts.
    • 
      
  • Stated premium does not match. The quote states an annual premium of 3,000 dollars; the rating gives 19,166 dollars.
    • 
    

    
5. Small applicant quote with a 5 million dollar limit: The broker resubmits after fixing the errors

    
The agent's request. Please quote Berkley Cyber Risk Solutions for a standard applicant with 8 million dollars of revenue and 900,000 personal records. Aggregate limit: 3 million dollars; retention: 10,000 dollars; social engineering sublimit: 500,000 dollars. The Technology E&O endorsement is included. Multi-factor authentication: yes; offline backups: yes. The quote states an annual premium of 14,815 dollars. It is ready to send.

    
Verdict: Needs approval.

    
    
      
  • Large social engineering sublimit. A social engineering sublimit of 500 thousand dollars needs an underwriter to confirm.
    • 
      
  • Quote summary. Estimated annual premium of 14,815 dollars for a 3 million dollar limit with a 10 thousand dollar retention.
    • 
    

    
6. Small applicant quote with a 5 million dollar limit: The broker resubmits after fixing the errors

    
The agent's request. Please quote Berkley Cyber Risk Solutions for a standard applicant with 8 million dollars of revenue and 900,000 personal records. Aggregate limit: 3 million dollars; retention: 10,000 dollars; social engineering sublimit: 100,000 dollars. The Technology E&O endorsement is included. Multi-factor authentication: yes; offline backups: yes. The quote states an annual premium of 13,358 dollars. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Quote summary. Estimated annual premium of 13,358 dollars for a 3 million dollar limit with a 10 thousand dollar retention.
    • 
    

  

  

    
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 Berkley Cyber Risk Solutions quote for an applicant: the industry class and revenue, the aggregate limit, the retention, and more. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of how those choices fit together and returns a verdict, every reason and the premium it computed.
    2. 
      
  2. I will start with the first request by the agent.
    3. 
      
  3. The agent says. Please quote Berkley Cyber Risk Solutions for a elevated-risk applicant with 60 million dollars of revenue and 1,500,000 personal records. Aggregate limit: 10 million dollars; retention: 25,000 dollars; social engineering sublimit: 1,000,000 dollars. The rest of the request is on screen. The quote states an annual premium of 15,000 dollars.
    4. 
      
  4. Before it goes out, the agent calls the validator.
    5. 
      
  5. Blocked, with three findings and five approvals, all in one answer. A retention of 25 thousand dollars is below the 50 thousand minimum for this revenue size. Multi-factor authentication for remote access, email and privileged accounts is not in place. The quote states an annual premium of 15,000 dollars; the rating gives 355,680 dollars.
    6. 
      
  6. The validator returned three blocking findings and five approvals 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 stated premium that must match the computed one, a rules-based CPQ can do this only with workarounds.
    8. 
      
  8. Now the agent resubmits.
    9. 
      
  9. The corrected request. It changes the aggregate limit to 5,000,000 dollars, the retention to 50,000 dollars, the social engineering sublimit to 250,000 dollars, the multi-factor authentication to yes and the premium stated on the quote to 183,435 dollars.
    10. 
      
  10. It calls the validator again.
    11. 
      
  11. Not blocked, but held for a person.
    12. 
      
  12. The three blocking findings are cleared; three approvals remain for a person: no offline backups, many records and elevated class with a large limit.
    13. 
      
  13. The agent resubmits once more.
    14. 
      
  14. The revised request. It changes the industry class to standard, the records held to 800,000, the offline backups to yes and the premium stated on the quote to 115,853 dollars.
    15. 
      
  15. It calls the validator once more.
    16. 
      
  16. This time the validator clears it.
    17. 
      
  17. All three earlier findings are cleared.
    18. 
      
  18. Eight 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.