Using AI with sensitive data in ServiceNow workflows — how are you handling it?

[romc]

Hi everyone,

I’ve been exploring how AI is being used within ServiceNow workflows (ticket summaries, knowledge suggestions, automation, etc.), and I’m curious about one key challenge:

How are you handling sensitive data when using AI?

In many cases, tickets and records can contain:

• Personal data (PII)

• Internal business information

• Customer-specific details

Sending that directly to external AI models can raise compliance and privacy concerns.

I’m currently exploring an approach where data is anonymized before it’s processed by AI — essentially adding a layer that removes or masks sensitive information while still preserving usefulness.

A few questions for those working with ServiceNow AI:

• Are you using AI features on sensitive records today?

• Do you rely fully on built-in ServiceNow AI, or external models as well?

• Is data privacy a blocker in your use cases?

Would really appreciate hearing how others are approaching this in real environments.

[Mark Manders · Accepted solution]

ServiceNow provides Data Privacy plugins you can use for this.