How findings are prioritised
Not every quantum-vulnerable setting is equally urgent. A key exchange that protects health records is a bigger problem than one that protects a session cookie, even if the cryptography is the same. QuantumShield ranks every finding by what is at stake in your org.
Harvest now, decrypt later
An attacker does not need a quantum computer today to steal today's data. They can record encrypted traffic now, store it, and decrypt it once a quantum computer can break the key exchange that protected it. So the question is not "when will quantum computers arrive?" but "will this data still matter when they do?"
The test
If the left side is bigger, you are already late: data sent today will still need protecting after it can be read. The difference in years is the finding's margin.
- Act now The margin is positive, or the setting is weak or non-compliant today regardless of quantum computing (plaintext HTTP, MD5, hardcoded keys).
- Plan Less than 3 years of slack: smaller than a typical budget and procurement cycle.
- Monitor Comfortable slack under the current assumptions.
Signatures and certificates are different: a recorded signature cannot be "decrypted later". For them only the fix time counts, because the replacement must be in place before forgery becomes possible.
Where the numbers come from
How long the data must stay secret
The scan reads, from metadata only and never from records:
- Your field classification. Salesforce's Data Sensitivity Level (Restricted and Mission Critical count as long-lived regulated data) and Compliance Categorization (PII, GDPR and CCPA count as personal data; HIPAA, PCI and GLBA as long-lived regulated data).
- Shield Platform Encryption. Fields you chose to encrypt count as at least confidential, and decide what each Shield key protects.
- Best-practice defaults for standard and industry objects that exist in the org, when nothing is classified (table below).
An integration is assumed to carry the most sensitive data found in the org. You know better: open a finding in the app and say what that integration really carries. Salesforce only shows fields the scanning user can see, so treat the profile as a lower bound.
How long the fix takes
Each rule states the effort of its fix. By default low effort takes 1 year, medium 2 years and high 4 years, including testing and coordination with partners.
When a quantum computer can break it
Nobody knows. QuantumShield assumes 2033 by default and offers 2030 (conservative), 2033 (moderate) and 2035 (optimistic) presets. NIST plans to deprecate today's public-key algorithms after 2030 and disallow them after 2035 (NIST IR 8547). Change the year in the app and the ranking updates instantly.
Default assumptions
All of these can be changed in the findings view. Your changes stay in your browser.
| Data | Stays secret for | Examples |
|---|---|---|
| Long-lived regulated | 25 years | Health, national identity and financial-account data that stays sensitive for decades. |
| Personal data | 10 years | Names, contact details and other personal data covered by privacy law. |
| Confidential business | 7 years | Contracts, pricing, pipeline and financial records kept for audit. |
| Internal | 3 years | Operational data with little value to an outsider after a few years. |
| Short-lived | 0 years | Session tokens, one-time codes and data that is worthless within months. |
Worked example
A named credential sends Contact records, which include a national ID number classified as Restricted, to a partner over TLS. Salesforce callouts still agree their keys with classical cryptography. The data must stay secret for 25 years and the fix is high effort (4 years). Scanned in 2026 with 2033 as the quantum year: 25 + 4 = 29 years, against 7 years left. The margin is 22 years, so it's Act now: trim what the integration sends. If you mark that integration as carrying only short-lived data, the same finding drops to Monitor.
Best-practice defaults for standard objects
Used only when the object exists in the scanned org. Objects that exist in every org whether used or not (such as PaymentMethod) are left out, because their presence says nothing about your data.
| Object | Assumed data | Why |
|---|---|---|
| Contact | Personal data | people’s names and contact details |
| Lead | Personal data | prospects’ names and contact details |
| Individual | Personal data | consent and privacy preferences |
| ContactPointEmail | Personal data | email addresses |
| ContactPointPhone | Personal data | phone numbers |
| ContactPointAddress | Personal data | postal addresses |
| Case | Personal data | free-text customer conversations |
| Account | Confidential business | customer relationships (personal data with person accounts) |
| Opportunity | Confidential business | pipeline and pricing |
| Contract | Confidential business | contract terms |
| Order | Confidential business | order history |
| Quote | Confidential business | pricing |
| IdentityDocument | Long-lived regulated | identity documents |
| FinancialAccount | Long-lived regulated | financial accounts |
| FinServ__FinancialAccount__c | Long-lived regulated | financial accounts |
| FinServ__Card__c | Long-lived regulated | card data |
| HealthCondition | Long-lived regulated | health records |
| ClinicalEncounter | Long-lived regulated | health records |
| PatientMedicalProcedure | Long-lived regulated | health records |
| MedicationStatement | Long-lived regulated | health records |
| CarePlan | Long-lived regulated | health records |
| HealthCloudGA__EhrPatient__c | Long-lived regulated | health records |
What this does not do
- It does not read record data. Field classification and encryption are metadata.
- It does not change severity, which stays as written in the rule catalog. Priority is a separate ranking on top.
- It does not predict when quantum computers will arrive. It shows how your priorities change with that one assumption.