Today, we're excited to launch Selective Redaction — the industry’s first solution to redact sensitive customer information from all conversations with the highest accuracy using AI.

44% of data breaches in 2021 included customers’ personally identifiable information (PII), according to IBM’s Cost of Data Breach Report. With an average cost of $161M in penalties for every million unredacted records, PII and PCI-related compliance risk is expensive!

For today's digital-first businesses, protecting customer information is no more an option. Failing to redact PII and PCI data from your recorded customer interactions will put you at risk of data leaks resulting in lawsuits, compliance penalties, and loss of customer trust.

But here’s the wrinkle. Legacy redaction techniques like agent-initiated stop & resume, keyword-based redaction, and numeric redaction are prone to compliance misses and lack the flexibility to redact only the information you need to protect, not more and not less.

Let’s take the example of Insurance companies.

Insurance companies must protect customer privacy by redacting policy numbers and other PII and PCI information. But it’s equally important to keep other numeric information like enrollment fees and time mentions to later evaluate agents on their product understanding and customer interactions.

At Observe.AI, we believe you should be able to maintain compliance without losing important data due to a lack of accuracy in your redaction technologies.‍

Introducing Selective Redaction:
Redact what you need, not more and not less

Watch how it works:

Leveraging legacy redaction solutions to protect sensitive PCI/PII data from millions of customer conversations can be highly unscalable, inaccurate, and laborious.

With Selective Redaction, all of that changes. Now, you can protect sensitive customer data with high accuracy without losing any contextual information required for intelligent business decisions. All you need to do is define the sensitive entities you want to redact and let Observe.AI’s deep learning techniques do the work for you.

Using deep learning techniques, Observe.AI’s Selective Redaction analyzes the surrounding context of individual phrases to accurately identify PII and PCI entities even when there are transcription errors.

Our Selective Redaction technology is actively being used by 50+ Observe.AI customers across insurance, collections, financial services, and retail with a combined call volume of 60+ million. They are already using it to secure customer information and have reported less than 50 errors for missed redaction and less than 100 for over redaction, delivering the highest level of redaction accuracy in the industry.

Selective Redaction is a game-changer, especially for compliance and QA teams.

Now they can—

Protect customer information with ease

Instead of settling with subpar technology like agent-initiated stop & resume, keyword-based, or numeric redaction. Observe.AI automatically removes only the pre-defined, sensitive data needed from calls and transcripts— not too much and not too little. Redact credit card numbers, not refund amounts. Policy numbers but not enrollment fee.

Reduce compliance risk

Adapt to a broad spectrum of compliance use cases and mitigate risk by selectively securing sensitive PCI/PII entities with high confidence.‍

Boost QA accuracy and agent performance

When you get complete visibility into customer interactions, you also get improved QA accuracy and agent performance. Did the agent share the correct loan terms? Did the rep outline the policy enrollment fees? Access the insights you need for more accurate QA evaluations and targeted coaching.

“We rely on call transcripts and visibility into what’s being said on customer interactions for coaching our agents. Observe.AI’s Selective Redaction solves a big pain point for us. We can now confidently remove sensitive information shared by clients while still getting the full context of customer and account validation for training purposes.”