All Resources
In this article:
minus iconplus icon
Share the Article

How Sentra Accurately Classifies Sensitive Data at Scale

July 30, 2024
5
 Min Read
Data Security

Background on Classifying Different Types of Data

It’s first helpful to review the primary types of data we need to classify - Structured and Unstructured Data and some of the historical challenges associated with analyzing and accurately classifying it.

What Is Structured Data?

Structured data has a standardized format that makes it easily accessible for both software and humans. Typically organized in tables with rows and/or columns, structured data allows for efficient data processing and insights. For instance, a customer data table with columns for name, address, customer-ID and phone number can quickly reveal the total number of customers and their most common localities.

Moreover, it is easier to conclude that the number under the phone number column is a phone number, while the number under the ID is a customer-ID. This contrasts with unstructured data, in which the context of each word is not straightforward. 

What Is Unstructured Data?

Unstructured data, on the other hand, refers to information that is not organized according to a preset model or schema, making it unsuitable for traditional relational databases (RDBMS). This type of data constitutes over 80% of all enterprise data, and 95% of businesses prioritize its management. The volume of unstructured data is growing rapidly, outpacing the growth rate of structured databases.

Examples of unstructured data include:

  • Various business documents
  • Text and multimedia files
  • Email messages
  • Videos and photos
  • Webpages
  • Audio files

While unstructured data stores contain valuable information that often is essential to the business and can guide business decisions, unstructured data classification has historically been challenging. However, AI and machine learning have led to better methods to understand the data content and uncover embedded sensitive data within them.

The division to structured and unstructured is not always a clear cut. For example, an unstructured object like a docx document can contain a table, while each structured data table can contain cells with a lot of text which on its own is unstructured. Moreover there are cases of semi-structured data. All of these considerations are part of Sentra’s data classification tool and beyond the scope of this blog.

Data Classification Methods & Models 

Applying the right data classification method is crucial for achieving optimal performance and meeting specific business needs. Sentra employs a versatile decision framework that automatically leverages different classification models depending on the nature of the data and the requirements of the task. 

We utilize two primary approaches: 

  1. Rule-Based Systems
  2. Large Language Models (LLMs)

Rule-Based Systems 

Rule-based systems are employed when the data contains entities that follow specific, predictable patterns, such as email addresses or checksum-validated numbers. This method is advantageous due to its fast computation, deterministic outcomes, and simplicity, often  providing the most accurate results for well-defined scenarios.

Due to their simplicity, efficiency, and deterministic nature, Sentra uses rule-based models whenever possible for data classification. These models are particularly effective in structured data environments, which possess invaluable characteristics such as inherent structure and repetitiveness.

For instance, a table named "Transactions" with a column labeled "Credit Card Number" allows for straightforward logic to achieve high accuracy in determining that the document contains credit card numbers. Similarly, the uniformity in column values can help classify a column named "Abbreviations" as 'Country Name Abbreviations' if all values correspond to country codes.

Sentra also uses rule-based labeling for document and entity detection in simple cases, where document properties provide enough information. Customer-specific rules and simple patterns with strong correlations to certain labels are also handled efficiently by rule-based models.

Large Language Models (LLMs)

Large Language Models (LLMs) such as BERT, GPT, and LLaMa represent significant advancements in natural language processing, each with distinct strengths and applications. BERT (Bidirectional Encoder Representations from Transformers) is designed for fine-grained understanding of text by processing it bidirectionally, making it highly effective for tasks like Named Entity Recognition (NER) when trained on large, labeled datasets.

In contrast, autoregressive models like the famous GPT (Generative Pre-trained Transformer) and Llama (Large Language Model Meta AI) excel in generating and understanding text with minimal additional training. These models leverage extensive pre-training on diverse data to perform new tasks in a few-shot or zero-shot manner. Their rich contextual understanding, ability to follow instructions, and generalization capabilities allow them to handle tasks with less dependency on large labeled datasets, making them versatile and powerful tools in the field of NLP. However, their great value comes with a cost of computational power, so they should be used with care and only when necessary.

Applications of LLMs at Sentra

Sentra uses LLMs for both Named Entity Recognition (NER) and document labeling tasks. The input to the models is similar, with minor adjustments, and the output varies depending on the task:

  • Named Entity Recognition (NER): The model labels each word or sentence in the text with its correct entity (which Sentra refers to as a data class).
  • Document Labels: The model labels the entire text with the appropriate label (which Sentra refers to as a data context).
  • Continuous Automatic Analysis: Sentra uses its LLMs to continuously analyze customer data, help our analysts find potential mistakes, and to suggest new entities and document labels to be added to our classification system.

Here you can see an example of how Sentra classifies personal information.
Note: Entity refers to data classes on our dashboard
Document labels refers to data context on our dashboard

Sentra’s Generative LLM Inference Approaches

An inference approach in the context of machine learning involves using a trained model to make predictions or decisions based on new data. This is crucial for practical applications where we need to classify or analyze data that wasn't part of the original training set. 

When working with complex or unstructured data, it's crucial to have effective methods for interpreting and classifying the information. Sentra employs Generative LLMs for classifying complex or unstructured data. Sentra’s main approaches to generative LLM inference are as follows:

Supervised Trained Models (e.g., BERT)

In-house trained models are used when there is a need for high precision in recognizing domain-specific entities and sufficient relevant data is available for training. These models offer customization to capture the subtle nuances of specific datasets, enhancing accuracy for specialized entity types. These models are transformer-based deep neural networks with a “classic” fixed-size input and a well-defined output size, in contrast to generative models. Sentra uses the BERT architecture, modified and trained on our in-house labeled data, to create a model well-suited for classifying specific data types. 

This approach is advantageous because:

  • In multi-category classification, where a model needs to classify an object into one of many possible categories, the model outputs a vector the size of the number of categories, n. For example, when classifying a text document into categories like ["Financial," "Sports," "Politics," "Science," "None of the above"], the output vector will be of size n=5. Each coordinate of the output vector represents one of the categories, and the model's output can be interpreted as the likelihood of the input falling into one of these categories.
  • The BERT model is well-designed for fine-tuning specific classification tasks. Changing or adding computation layers is straightforward and effective.
  • The model size is relatively small, with around 110 million parameters requiring less than 500MB of memory, making it both possible to fine-tune the model’s weights for a wide range of tasks, and more importantly - run in production at small computation costs.
  • It has proven state-of-the-art performance on various NLP tasks like GLUE (General Language Understanding Evaluation), and Sentra’s experience with this model shows excellent results.

Zero-Shot Classification

One of the key techniques that Sentra has recently started to utilize is zero-shot classification, which excels in interpreting and classifying data without needing pre-trained models. This approach allows Sentra to efficiently and precisely understand the contents of various documents, ensuring high accuracy in identifying sensitive information.

The comprehensive understanding of English (and almost any language) enables us to classify objects customized to a customer's needs without creating a labeled data set. This not only saves time by eliminating the need for repetitive training but also proves crucial in situations where defining specific cases for detection is challenging. When handling sensitive or rare data, this zero-shot and few-shot capability is a significant advantage.

Our use of zero-shot classification within LLMs significantly enhances our data analysis capabilities. By leveraging this method, we achieve an accuracy rate with a false positive rate as low as three to five percent, eliminating the need for extensive pre-training.

Sentra’s Data Sensitivity Estimation Methodologies

Accurate classification is only a (very crucial) step to determine if a document is sensitive. At the end of the day, a customer must be able to also discern whether a document contains the addresses, phone numbers or emails of the company’s offices, or the company’s clients.

Accumulated Knowledge

Sentra has developed domain expertise to predict which objects are generally considered more sensitive. For example, documents with login information are more sensitive compared to documents containing random names. 

Sentra has developed the main expertise based on our collected AI analysis over time.

How does Sentra accumulate the knowledge? (is it via AI/ML?)

Sentra accumulates knowledge both from combining insights from our experience with current customers and their needs with machine learning models that continuously improve based on the data they are trained with over time.

Customer-Specific Needs

Sentra tailors sensitivity models to each customer’s specific needs, allowing feedback and examples to refine our models for optimal results. This customization ensures that sensitivity estimation models are precisely tuned to each customer’s requirements.

What is an example of a customer-specific need?

For instance, one of our customers required a particular combination of PII (personally identifiable information) and NPPI (nonpublic personal information). We tailored our solution by creating a composite classifier to meet their needs by designating documents containing these combinations as having a higher sensitivity level.

Sentra’s sensitivity assessment (that drives classification definition) can be based on detected data classes, document labels, and detection volumes, which triggers extra analysis from our system if needed.

Conclusion

In summary, Sentra’s comprehensive approach to data classification and sensitivity estimation ensures precise and adaptable handling of sensitive data, supporting robust data security at scale. With accurate, granular data classification, security teams can confidently proceed to remediation steps without need for further validation - saving time and streamlining processes.  Further, accurate tags allow for automation - by sharing contextual sensitivity data with upstream controls (ex. DLP systems) and remediation workflow tools (ex. ITSM or SOAR).

Additionally, our research and development teams stay abreast of the rapid advancements in Generative AI, particularly focusing on Large Language Models (LLMs). This proactive approach to data classification ensures our models not only meet but often exceed industry standards, delivering state-of-the-art performance while minimizing costs. Given the fast-evolving nature of LLMs, it is highly likely that the models we use today—BERT, GPT, Mistral, and Llama—will soon be replaced by even more advanced, yet-to-be-published technologies.

<blogcta-big>

After earning a BSc in Mathematics and a BSc in Industrial Engineering, followed by an MSc in Computer Science with a thesis in Machine Learning theory, Hanan has spent the last five years training models for feature-based and computer vision problems. Driven by the motivation to deliver real-world value through his expertise, he leverages his strong theoretical background and hands-on experience to explore and implement new methodologies and technologies in machine learning. At Sentra, one of his main focuses is leveraging large language models (LLMs) for advanced classification and analysis tasks.

Romi is the senior marketing manager at Sentra, bringing years of experience in various marketing roles in the cybersecurity field.

Subscribe

Latest Blog Posts

Mark Kiley
Mark Kiley
April 10, 2026
4
Min Read

South Carolina Insurance Data Security Act: Data Security Requirements and How to Prove “Reasonable Security”

South Carolina Insurance Data Security Act: Data Security Requirements and How to Prove “Reasonable Security”

South Carolina Insurance Data Security Act: Data Security Requirements and How to Prove “Reasonable Security”

If you’re an insurer or insurance licensee doing business in South Carolina, you now have two layers of data security law to worry about:

  • The general South Carolina data breach notification law (S.C. Code § 39‑1‑90), and
  • The South Carolina Insurance Data Security Act (Title 38, Chapter 99), which imposes sector‑specific cybersecurity and incident reporting requirements on insurance licensees.

Directors, CISOs, and compliance leaders keep asking the same core questions:

  • Exactly who does the Insurance Data Security Act apply to?
  • What does “reasonable security” actually mean under this law?
  • How does it interact with the general breach notification statute?
  • And how can we prove compliance when an exam or incident happens?

This post walks through the law in plain language and, more importantly, shows how data‑centric security and DSPM make it far easier to demonstrate that you’re doing the right things.

What is the South Carolina Insurance Data Security Act?

The South Carolina Insurance Data Security Act (SCIDSA) is codified at Title 38, Chapter 99 of the South Carolina Code of Laws. It was enacted in 2018 and is modeled closely on the NAIC Insurance Data Security Model Law.

In short, it:

  • Requires insurance licensees to develop, implement, and maintain a comprehensive written information security program based on a risk assessment.
  • Imposes specific obligations around incident response, investigation, and reporting of cybersecurity events to the SC Department of Insurance.
  • Mandates oversight of third‑party service providers that access nonpublic information.
  • Requires annual certification of compliance to the Director of Insurance (for domestic insurers).

Think of it as the insurance‑specific overlay on top of South Carolina’s general breach law and any federal obligations you may have (GLBA, HIPAA for certain products, etc.).

Who does the Act apply to?

The Act applies to “licensees” that are licensed, authorized, or registered (or required to be) under South Carolina’s insurance laws, with certain exceptions.

That includes, for example:

  • Insurers (life, P&C, health, specialty carriers)
  • HMOs and many health plans regulated as insurers
  • Producers, agencies, and certain intermediaries
  • Other entities holding a license from the SC Department of Insurance

There are some exemptions. For instance, licensees with fewer than a specified number of employees or licensees already compliant with equivalent requirements in another state, but they are narrow, and you should confirm applicability with counsel.

If you’re writing policies in South Carolina or handling SC policyholder/insured data, you should assume SCIDSA applies until you’ve proven otherwise.

What does the law require? (High‑level overview)

At a high level, the Insurance Data Security Act requires licensees to do five big things:

  1. Build and maintain a written information security program that is risk‑based and appropriate to your size, complexity, and the sensitivity of your data.
  2. Conduct regular risk assessments to identify reasonably foreseeable threats to nonpublic information and your information systems.
  3. Implement controls across administrative, technical, and physical domains—covering access control, encryption, monitoring, incident response, and more.
  4. Oversee third‑party service providers that handle nonpublic information on your behalf, ensuring they maintain appropriate security.
  5. Investigate and report cybersecurity events to the SC Department of Insurance within defined timelines, and maintain documentation and records for exam and enforcement purposes.

What follows is a closer look at the pieces that matter most from a data‑security and breach‑readiness perspective.

Nonpublic information: what are you actually protecting?

The Act uses the term “nonpublic information” rather than just PII. That typically includes:

  • Business‑related information that, if tampered with or disclosed, would materially impact your operations or security.
  • Consumer information that can identify a person when combined with data elements like SSNs, financial account numbers, or driver’s license numbers—similar to but often broader than the general breach law’s definition.
  • Certain health or medical information associated with insurance products.

For insurers, this often spans:

  • Policyholder and applicant records
  • Claims data and adjuster notes
  • Payment and bank details
  • Underwriting models and internal risk scoring data
  • Credentials and MFA secrets granting access to core systems

This “nonpublic information” is what your information security program, and your incident response obligations, are ultimately organized around.

“Reasonable security” under SCIDSA: more than a buzzword

Many laws talk about “reasonable” or “appropriate” safeguards. The Insurance Data Security Act goes further by spelling out what a risk‑based program must include, such as:

  • Designating one or more responsible individuals (or an outside vendor) to oversee the information security program.
  • Conducting and documenting periodic risk assessments of threats to nonpublic information and information systems.
  • Implementing controls for:
    • Access controls and authentication
    • Physical and environmental security
    • Encryption or equivalent protections for data in transit and at rest
    • Secure development practices for internally built applications
    • Monitoring, logging, and testing for unauthorized access or changes
    • Incident response planning and disaster recovery

For licensees with a board of directors, executive management must regularly report to the board (or a committee) on the overall status of the information security program, material risks, and recommended changes.

This is the statutory backbone behind a regulator or plaintiff saying, “Did you have reasonable security in place?”

Cybersecurity events and reporting: what triggers notice?

The Insurance Data Security Act introduces the defined concept of a “cybersecurity event”—broadly, an event resulting in unauthorized access to or disruption of an information system or nonpublic information, with some carve‑outs (such as access that is determined not to have been used or released and has been returned or destroyed).

When a licensee learns that a cybersecurity event has occurred or may have occurred, it must conduct a prompt investigation to determine:

  • Whether a cybersecurity event actually occurred
  • The nature and scope of the event
  • What nonpublic information may have been involved
  • What measures are needed to restore system security and prevent recurrence

Department of Insurance notification

A licensee must notify the Director of the Department of Insurance of a cybersecurity event no later than 72 hours after determining that such an event has occurred, when either:

  • South Carolina is the licensee’s state of domicile (for insurers) or home state (for producers), or
  • The event involves the nonpublic information of at least 250 South Carolina consumers and either:
    • The event must be reported to another governmental body or regulator, or
    • The event is reasonably likely to materially harm a South Carolina consumer or a material part of the licensee’s normal operations.

The notice to the Director must include, as much as possible at the time:

  • Date and nature of the event
  • How information was exposed, lost, or accessed
  • Types of nonpublic information involved
  • Number of SC consumers affected
  • Law enforcement involvement
  • Steps taken to investigate, remediate, and notify consumers

The Constangy Cyber and other practitioner summaries emphasize that these requirements layer on top of, not in place of, your obligations to consumers under § 39‑1‑90 and any federal laws.

Interaction with the general breach notification statute

SCIDSA is explicit that licensees must still comply with S.C. Code § 39‑1‑90 for consumer notice, where applicable. In other words:

  • SCIDSA: Notice to the Director of Insurance (regulator) within ~72 hours in certain thresholds and conditions.
  • § 39‑1‑90: Notice to South Carolina residents, the Department of Consumer Affairs, and credit bureaus based on the scope of the breach and harm threshold.

For insurance CISOs and compliance officers, the practical takeaway is that you must design an incident response program that simultaneously satisfies both statutes, plus any other state or federal obligations in the jurisdictions where you operate.

Why this is hard in 2026: the data sprawl reality

On paper, the Insurance Data Security Act reads like a classic security program checklist. In practice, the hardest parts are:

  • Knowing where nonpublic information actually resides across policy admin platforms, claims systems, data warehouses, email, collaboration tools, and third‑party SaaS.
  • Understanding real‑world access—not just IAM roles on paper, but which users, service accounts, vendors, and AI tools can actually touch sensitive data.
  • Quantifying impact quickly during an incident so you can meet the 72‑hour Department of Insurance clock and the “most expedient time possible” obligation for consumer notice.

Most insurers have grown through acquisition, product launches, and third‑party integrations. That means policyholder and claimant data often exists in:

  • Multiple core systems and data lakes
  • Historical archives and backups nobody has looked at for years
  • Ad‑hoc exports shared with vendors or internal analytics teams
  • Email and collaboration workspaces, often with full Excel dumps attached

Without continuous, accurate visibility into that sprawl, your ability to prove “reasonable security” and respond within statutory timelines depends more on luck than design.

How data‑centric security and DSPM help prove compliance

This is where Data Security Posture Management (DSPM) and data‑centric security come into play.

A modern DSPM platform like Sentra is designed to answer, continuously and at scale, the core questions baked into the Insurance Data Security Act:

  • What nonpublic information do we actually hold?
  • Where does it reside across cloud, SaaS, and on‑prem?
  • How is it protected (encryption, masking, labeling, access controls)?
  • Who or what can access it—humans, APIs, AI agents, third‑party vendors?

Sentra does this by:

  • Discovering and classifying sensitive data across your clouds, SaaS, and on‑prem data stores, including policy/claims systems, warehouses, and collaboration tools.
  • Building a live, context‑rich inventory of regulated data (PII, PCI, health, credentials, etc.) and mapping it to your environments and business units.
  • Continuously analyzing exposure and effective access, so you see where nonpublic information is overshared, unencrypted, or accessible from risky identities.
  • Integrating with your incident response workflows, so that when a security event occurs, you can quickly scope which data sets and how many consumers are truly in play.

A real‑world parallel: SoFi’s DSPM story

While SoFi is a financial services leader rather than an insurer, their story closely mirrors the challenges SC insurers face.

In our webinar and case study with SoFi, their security leaders describe how they used Sentra to:

  • Create a centralized data catalog of sensitive customer data across a complex cloud environment.
  • Improve compliance mapping by aligning data classes with regulatory frameworks and internal policies.
  • Tighten data access governance, reducing false positives and focusing on the exposures that matter most.

Their experience moving from scattered, manual efforts to automated, high‑confidence visibility and classification is exactly what SC insurers need to align day‑to‑day operations with SCIDSA’s “reasonable security” and incident reporting expectations.

Making the Insurance Data Security Act manageable

If you’re an insurance licensee in South Carolina, the path forward looks something like this:

  1. Confirm applicability and scope with legal counsel, but assume SCIDSA and § 39‑1‑90 both matter if you hold South Carolina policyholder or claimant data.
  2. Inventory your nonpublic information using automated discovery and classification—it’s no longer realistic to do this with spreadsheets.
  3. Align your risk assessment and controls with what the Act explicitly requires: access control, encryption, monitoring, incident response, and third‑party oversight.
  4. Instrument your incident response so that every suspected cybersecurity event immediately pulls in DSPM context: data types, consumers affected, exposure level, and cross‑jurisdiction triggers.
  5. Document everything: risk assessments, board reports, incident investigations, and rationales for notification or non‑notification decisions. This is what “reasonable security” looks like under exam.

With that in place, the next early‑morning call from the SOC won’t feel like a blind scramble against a 72‑hour clock. You’ll be operating from a place of data‑driven confidence, not guesswork.

Call to action

If you’re responsible for data security or compliance under the South Carolina Insurance Data Security Act, now is the time to get ahead of the next exam—or the next incident.

See how Sentra helps insurers continuously map nonpublic information, reduce exposure, and accelerate investigations so you can meet SCIDSA and § 39‑1‑90 obligations with confidence.

Request a Sentra demo

Read More
Mark Kiley
Mark Kiley
April 10, 2026
4
Min Read

South Carolina Data Breach Notification Requirements: What CISOs Need to Know About SC Code § 39‑1‑90

South Carolina Data Breach Notification Requirements: What CISOs Need to Know About SC Code § 39‑1‑90

South Carolina Data Breach Notification Requirements: What CISOs Need to Know About SC Code § 39‑1‑90

Imagine you’re the CISO of a fast‑growing company in Charleston. It’s 6:30 a.m., and your phone lights up with a message from the SOC: suspicious activity in a cloud database that holds customer information.

The good news: logs show you caught it quickly. The bad news: you’re doing business in South Carolina, and that means SC Code § 39‑1‑90, the state’s data breach notification law, just became very real.

Within hours, your executive team will want answers:

  • Is this a “breach of the security of the system” under South Carolina law?
  • Do we have to notify South Carolina residents?
  • What about the Consumer Protection Division and credit bureaus?
  • How fast do we need to move—and what happens if we get it wrong?

To answer those questions confidently, you need more than a legal summary. You need a clear view of your data: where South Carolina residents’ information actually lives, how it’s protected, and how many people could be affected in a worst‑case scenario.

This article walks through the law and the operational reality behind it.

Who South Carolina’s breach law applies to

South Carolina’s general breach notification statute, S.C. Code § 39‑1‑90, applies broadly to any “person conducting business in this State” that owns or licenses computerized data (or other data) including personal identifying information of South Carolina residents.

It also covers organizations that maintain such data on behalf of someone else, even if they don’t own it.

In practice, that means:

  • South Carolina–headquartered companies
  • Out‑of‑state companies doing business in SC and holding SC residents’ data
  • Many types of commercial entities, and even some governmental subdivisions

There are limited carve‑outs for example, financial institutions already in compliance with Gramm‑Leach‑Bliley Act (GLBA) privacy and security provisions can be deemed compliant under certain circumstances. But most organizations handling consumer data should assume they’re squarely in scope.

What “personal identifying information” means in South Carolina

Under § 39‑1‑90(D)(3), “personal identifying information” is defined as a South Carolina resident’s first name or first initial and last name in combination with one or more of the following, when not encrypted or redacted:

  • Social Security number
  • Driver’s license number or state identification card number
  • Financial account number, credit card number, or debit card number plus any required security code, access code, or password that would permit access to a financial account
  • Other numbers or information that may be used to access a person’s financial accounts, or numbers or information issued by a governmental or regulatory entity that uniquely identify an individual

Information that is lawfully available from public records, or already public, is excluded.

The key is that South Carolina is focused on financial and identity‑enabling data elements, not every piece of PII you might hold. But in a modern environment with backups, exports, analytics, and AI pipelines, those elements often end up in more places than you expect.

What counts as a “breach of the security of the system”

South Carolina law defines a “breach of the security of the system” as unauthorized access to and acquisition of computerized data (not rendered unusable through encryption, redaction, or other methods) that compromises the security, confidentiality, or integrity of personal identifying information when:

  • Illegal use of the information has occurred, or
  • Illegal use is reasonably likely to occur, or
  • Use of the information creates a material risk of harm to a resident

There are a few important qualifiers:

  • Good‑faith access by employees or agents for legitimate business purposes is not considered a breach, provided the data is not used improperly or further disclosed.
  • If data is properly encrypted, redacted, or otherwise rendered unusable, the statute does not apply.

External analyses, such as Davis Wright Tremaine’s summary, emphasize that the risk‑of‑harm threshold is real: notification is generally required only if illegal use has occurred or is likely, or if there’s a material risk of harm.

That sounds flexible. In practice, it means you need facts—not guesses—about the data involved before you decide whether to notify.

Notification obligations and timelines

Once you determine that a breach of the security of the system has occurred, SC Code § 39‑1‑90 imposes several notification duties.

Notification to South Carolina residents

If you own or license the data, you must notify affected South Carolina residents “in the most expedient time possible and without unreasonable delay” after discovery, taking into account law enforcement needs and efforts to determine the scope of the breach and restore system integrity.

The statute doesn’t prescribe a specific number of days, but third‑party guides like the Davis Wright Tremaine and Mintz matrices reinforce that regulators will look closely at whether your investigation and response were reasonable in context.

Notice can be delivered by:

  • Written letter
  • Telephone
  • Electronic notice, where appropriate under E‑SIGN and state law

If the cost or scale is prohibitive (e.g., more than 500,000 people affected or notification would cost over $250,000), substitute notice is permitted, typically combining email (if available), website posting, and major statewide media.

Unlike some states, South Carolina does not prescribe detailed content requirements for consumer notices, but best practice—reflected in many AG and regulator expectations—is to describe the incident, the type of information involved, and what you’re doing about it.

Notification to the Department of Consumer Affairs

If you notify more than 1,000 South Carolina residents about a breach, you must also notify, without unreasonable delay:

  • The Consumer Protection Division of the South Carolina Department of Consumer Affairs, and
  • All nationwide consumer reporting agencies (credit bureaus) about the timing, distribution, and content of your consumer notice

Note that this is not the Attorney General; South Carolina’s primary regulator for breach notices is the Department of Consumer Affairs.

Notification by third‑party data custodians

If you maintain personal identifying information on behalf of another entity, but don’t own it, you must notify the owner or licensee of any breach immediately following discovery.

For service providers, this means notification clauses and SLAs in customer contracts need to line up with statutory expectations.

Enforcement, penalties, and private lawsuits

South Carolina’s statute has real teeth.

A resident injured by a violation may bring a civil action to:

  • Recover damages for willful and knowing violations, or
  • Recover actual damages for negligent violations
  • Seek an injunction to enforce compliance
  • Recover attorneys’ fees and court costs if successful

Separately, a person who knowingly and willfully violates the statute is subject to an administrative fine of $1,000 per South Carolina resident whose information was accessible because of the breach, as determined by the Department of Consumer Affairs.

Legal summaries from firms like Davis Wright Tremaine, Constangy, and Mintz all underscore that South Carolina offers both regulatory enforcement and a private right of action, making it riskier to treat breach obligations as a check‑the‑box exercise.

Where CISOs get stuck: the “material risk of harm” problem

On paper, South Carolina’s harm threshold sounds friendly to businesses: notification isn’t required if you reasonably believe illegal use has not occurred and isn’t likely, or if the incident doesn’t create a material risk of harm.

In real incidents, that’s exactly where CISOs and legal teams get stuck.

To make that judgment call—especially one you’re comfortable defending to regulators, plaintiffs’ lawyers, and your own board—you need to answer a few hard questions fast:

  • What exact data did the attacker access, copy, or have the opportunity to exfiltrate?
  • Was it really rendered unusable by encryption or tokenization, or were keys and secrets in the same blast radius?
  • How many South Carolina residents had personal identifying information in those datasets?
  • Was access limited to a tightly scoped subset, or were you dealing with a broad analytics cluster or data lake that’s been accreting identity and financial data for years?

In many organizations, getting those answers requires days or weeks of manual work: exporting schemas, sampling records, cross‑referencing customer locations, and hoping you didn’t miss the one legacy bucket with a full set of unencrypted account numbers.

That delay is exactly what “most expedient time possible and without unreasonable delay” is designed to avoid.

Why DSPM is becoming a necessity for South Carolina breach readiness

This is where Data Security Posture Management (DSPM) shifts from buzzword to practical foundation.

DSPM focuses on continuously discovering, classifying, and assessing the risk posture of sensitive data across cloud, SaaS, and on‑prem environments—so you can answer, on demand, the questions South Carolina’s law implicitly asks: what data, whose data, how exposed, and how bad is the harm?

Continuous visibility into SC‑regulated data

A modern DSPM platform like Sentra connects agentlessly to cloud providers, data warehouses, SaaS apps, and on‑prem data stores, building a live map of:

  • Where sensitive data lives
  • Which datasets contain personal identifying information that fits South Carolina’s definition
  • How that data is protected (encryption, access controls, masking)
  • Who can actually access it, including service accounts and AI assistants

That means when an incident hits an S3 bucket, a Snowflake database, or a collaboration platform, you’re not starting from zero—you already know which assets in that blast radius contain South Carolina residents’ identity and financial data.

From law-on-paper to incident decisions

The value becomes clear in the heat of an investigation. Instead of arguing in the abstract about “material risk of harm,” your security and legal teams can look at concrete facts:

  • The compromised system contained N records with name + account numbers for South Carolina residents, unencrypted.
  • The attacker had read access for T minutes, during which exfiltration events were/weren’t observed.
  • Keys for encrypted datasets were in a separate KMS environment and not accessed.

Those details don’t just drive whether notice is required; they also shape the narrative with regulators and consumers when you do notify.

A real‑world example: SoFi’s DSPM journey with Sentra

While SoFi isn’t a South Carolina case, their story illustrates what this looks like in practice.

The SoFi security team operates in a heavily regulated financial environment, with a complex cloud‑native data landscape. In a recent webinar and blog, SoFi’s Director of Product Security and Senior Staff Application Security Engineer described how they used Sentra’s DSPM to:

  • Build a centralized data catalog with accurate, automated discovery and classification of sensitive data
  • Map data assets to regulatory requirements and internal security policies
  • Strengthen data access governance, reducing over‑permissioning and improving their ability to answer “who can see what, and why?”

Their challenge—data sprawl, lack of visibility, compliance pressure—is the same pattern CISOs in South Carolina see every day. The difference is that, with DSPM, SoFi moved from reactive fire drills to a proactive, continuously updated view of risk.

That’s exactly the posture you want when you’re making high‑stakes decisions under a state law that hinges on “reasonable belief” and “material risk of harm.”

Bringing it all together for South Carolina

If you operate in South Carolina, breach readiness isn’t just about having a playbook and a PR statement. It’s about being able to prove, under pressure, that you:

  • Know where South Carolina residents’ personal identifying information actually lives
  • Can quickly determine whether a given incident meets SC Code § 39‑1‑90’s definition of a breach
  • Can back up your harm assessments with data, not intuition
  • Can notify residents, regulators, and credit bureaus without unreasonable delay—and without having to issue embarrassing corrections later

DSPM gives you the substrate for those decisions. From there, you can integrate it into:

  • Your incident response plans, so every major incident automatically pulls in data classification and exposure context
  • Your governance processes, so you reduce breach blast radius over time by cleaning up shadow data and tightening access
  • Your board and regulator communications, with concrete, continuously updated metrics instead of static spreadsheets

South Carolina’s law is not unique in the US—but it’s a clear example of where data‑centric security and legal obligations now meet.

If you’d like to see what South Carolina breach readiness looks like with real‑time data intelligence, we can show you.

See how Sentra maps sensitive data, exposure, and risk so you can make faster, defensible decisions under SC Code § 39‑1‑90. Request a Sentra demo

Read More
Mark Kiley
Mark Kiley
April 6, 2026
3
Min Read

North Carolina Data Breach Notification Law: Requirements, Timelines, and Checklist for 2026

North Carolina Data Breach Notification Law: Requirements, Timelines, and Checklist for 2026

North Carolina has been ahead of the curve on breach notification. Its Identity Theft Protection Act (N.C. Gen. Stat. Chapter 75, Article 2A) sets clear requirements for how quickly organizations must notify residents and the Attorney General when personal information is exposed in a security incident.

For security and compliance leaders operating in or with NC, the big challenge isn’t just understanding the law on paper, it’s being able to answer, with evidence, exactly what data was exposed, where it lived, and who was affected when an incident hits.

This guide breaks down:

  • Who the NC breach law applies to
  • What “personal information” means under NC law
  • What counts as a security breach
  • Notification requirements and timelines
  • A practical checklist to operationalize NC breach readiness
  • How Data Security Posture Management (DSPM) makes this manageable at cloud scale

Who the North Carolina breach law applies to

North Carolina’s Identity Theft Protection Act applies broadly to any business that owns or licenses personal information of NC residents or conducts business in NC and holds personal information, whether computerized or not.

That includes:

  • NC‑headquartered organizations
  • Out‑of‑state organizations holding NC residents’ personal data
  • Both private sector and, for certain provisions, state and local agencies

If your organization stores customer, employee, or patient data for NC residents—especially in healthcare, financial services, insurance, education, retail, or SaaS—you should assume the law applies.

What “personal information” means in North Carolina

Under N.C. Gen. Stat. § 75‑61 and § 75‑65, “personal information” (PI) is defined as a person’s first name or first initial and last name in combination with any one of several sensitive data elements, when that data is not encrypted or redacted.

Common examples include:

  • Social Security numbers
  • Driver’s license, state ID, or passport numbers
  • Financial account numbers, credit or debit card numbers, plus any required security code, access code, or password for the account
  • Biometric data and other unique identifiers that can be used to access financial resources or uniquely identify an individual

Certain electronic identifiers (like usernames, email addresses, or internet account numbers) can also qualify as PI if they would permit access to a financial account or resources when combined with a password or other credentials.

For security teams, the takeaway is straightforward but difficult in practice: anything that can be used to impersonate, financially exploit, or uniquely identify an NC resident must be treated as regulated data.

What counts as a “security breach” in NC?

North Carolina defines a “security breach” as an incident of unauthorized access to and acquisition of unencrypted and unredacted records or data containing personal information where illegal use has occurred or is reasonably likely to occur, or that creates a material risk of harm to a consumer.

A few important nuances:

  • Good‑faith access by employees or agents is not a breach, as long as the information is used only for legitimate business purposes and is not subject to further unauthorized disclosure.
  • Encrypted data is generally not considered breached unless the encryption keys or confidential process needed to unlock the data are also compromised.
  • North Carolina guidance explicitly recognizes identity theft and financial harm as key risk factors when determining whether notice is required.

In practice, many organizations err on the side of treating any credible unauthorized access to PI as a potential breach until a risk assessment proves otherwise.

Notification requirements and timelines

Once your organization discovers or is notified of a breach involving NC residents’ PI, several notification obligations may apply.

1. Notice to affected individuals

Businesses must notify affected NC residents “without unreasonable delay” after discovery of the breach, taking into account law enforcement needs and time to determine the scope of the breach and restore system integrity.

The notice must be clear and conspicuous and include at least:

  • A general description of the incident
  • The type of personal information involved
  • A description of measures taken to protect the information from further unauthorized access
  • A contact telephone number for more information
  • Advice to review account statements and monitor free credit reports
  • Contact details for the major consumer reporting agencies, the Federal Trade Commission, and the NC Attorney General’s Office

Notice can be provided by:

  • Written notice
  • Electronic notice (if the consumer has agreed to electronic communications)
  • Telephonic notice
  • Substitute notice (email + prominent website posting + statewide media) if costs or scale exceed statutory thresholds.

2. Notice to the North Carolina Attorney General

If a business provides notice to affected individuals, it must also notify the Consumer Protection Division of the NC Attorney General’s Office without unreasonable delay.

That notice must describe:

  • The nature of the breach
  • The number of NC consumers affected
  • Steps taken to investigate the breach and prevent future incidents
  • Timing, distribution, and content of consumer notices

The NC Department of Justice maintains guidance and contact information here:

3. Notice to consumer reporting agencies

If you notify more than 1,000 individuals at one time, you must also notify all nationwide consumer reporting agencies of the timing, distribution, and content of the consumer notice, without unreasonable delay.

Penalties and enforcement

A violation of North Carolina’s breach notification requirements is considered an unfair or deceptive trade practice under N.C. Gen. Stat. § 75‑1.1, enforced by the Attorney General.

Key points:

  • The AG can seek injunctive relief, civil penalties, and other remedies.
  • Individuals may have a private right of action if they are injured as a result of the violation.
  • Repeated or willful noncompliance can significantly increase exposure, especially if regulators view your security practices as unreasonable given your size and risk profile.

For many boards and CISOs, the reputational damage and downstream regulatory scrutiny from a mishandled NC breach can matter as much as direct financial penalties.

Why NC breach readiness is hard in 2026

On paper, NC’s requirements look straightforward: discover breach → determine scope → notify affected people and regulators promptly. The complexity comes from the “determine scope” step:

  • Cloud sprawl: Sensitive data sprawls across object storage (e.g., S3, GCS, Azure Blob), data warehouses, SaaS apps, and backups.
  • Shadow and legacy data: Old exports, test copies, and forgotten file shares often have the most complete—and poorly protected—PII sets.
  • Multi‑cloud and hybrid: Different platforms expose different telemetry; correlating it to “which NC residents were affected?” can take weeks.
  • AI and unstructured data: Chat logs, support transcripts, and AI training sets now routinely contain PI but are rarely tracked like systems of record.

Without always‑on, accurate visibility into where personal data lives and how it’s exposed, NC’s expectation of “without unreasonable delay” can collide with your ability to answer basic questions:

  • Which datasets in the affected environment contained NC residents’ PI?
  • Exactly what types of PI were present (SSNs, account numbers, health data)?
  • Who had access, and were they over‑permissioned?

This is where Data Security Posture Management (DSPM) becomes a practical foundation rather than a buzzword.

How DSPM helps operationalize North Carolina breach requirements

Data Security Posture Management (DSPM) focuses on continuously discovering, classifying, and assessing the risk posture of sensitive data—wherever it lives across cloud, SaaS, and hybrid environments.

A mature DSPM program gives NC‑regulated organizations the ability to:

  1. Maintain a live inventory of NC residents’ PI


    • Automatically discover data stores containing PI across cloud, SaaS, and on‑prem.
    • Classify data as PII/PHI/PCI, tagged by geography or residency where possible.
    • See at a glance which systems hold North Carolina‑resident data and what types.
  2. → Learn more: Data Security Posture Management (DSPM)

  3. Assess exposure and “material risk of harm” quickly


    • Understand whether affected datasets were encrypted and how keys are managed (critical under NC’s definition of breach).
    • See who had effective access to PI (including service accounts and AI agents), not just theoretical permissions.
    • Identify misconfigurations like public buckets, overly broad access policies, or data in high‑risk regions.
  4. → Related reading: Cloud Data Security Means Shrinking the Data Attack Surface

  5. Accelerate incident scoping and notification decisions


    • When a storage location, SaaS tenant, or account is compromised, instantly surface:
      • Which tables/buckets/files contained NC‑defined personal information
      • How many unique NC residents were likely impacted
      • Whether encryption, masking, or tokenization meaningfully reduced risk
    • Use this as the factual backbone of your AG and consumer notifications.
  6. → Related use case: Keep Your Cloud Data Compliant

  7. Continuously reduce breach blast radius in NC and beyond


    • Proactively remove ROT (redundant, obsolete, trivial) data and risky legacy copies that amplify NC breach scope.
    • Automate remediation workflows—tightening access, encrypting high‑risk data stores, and enforcing retention policies.
    • Generate evidence for audits and regulator inquiries about your ongoing data protection program.
  8. → Deep dive: Manage Data Security and Compliance Risks with DSPM

A practical NC breach‑readiness checklist

To align with North Carolina’s data breach law and make incident response defensible:

  1. Map your NC footprint

    • Identify which systems hold NC residents’ PI and tag them accordingly in your asset inventory/DSPM.
  2. Deploy continuous discovery and classification

    • Move from annual spreadsheets to ongoing, automated detection of PI across cloud, SaaS, and on‑prem data stores.
  3. Define “material risk of harm” criteria

    • Involve legal, compliance, and security to define when access to PI triggers NC notification duties, incorporating encryption and key‑management posture.
  4. Pre‑draft NC‑specific notification templates

    • Include all NC‑required content elements and keep them updated with current AG and FTC contact information.
  5. Integrate DSPM findings with IR playbooks

    • Ensure your incident response runbooks explicitly call DSPM to:
      • Enumerate affected data stores
      • Classify PI types
      • Estimate affected NC residents
  6. Test NC‑specific tabletop exercises

    • Run at least one scenario per year involving NC residents’ data, including simulated AG notification, CRA notification, and evidence collection.
  7. Document your “no‑notice” decisions

    • If you determine a particular incident does not require notice under NC law, document the risk assessment and supporting data (encryption status, access logs, etc.) and retain it.
  8. → NC DOJ guidance: Security Breach Information – NC DOJ

Ready to see your North Carolina breach exposure in real time?Request a Sentra demo

Read More
Expert Data Security Insights Straight to Your Inbox
What Should I Do Now:
1

Get the latest GigaOm DSPM Radar report - see why Sentra was named a Leader and Fast Mover in data security. Download now and stay ahead on securing sensitive data.

2

Sign up for a demo and learn how Sentra’s data security platform can uncover hidden risks, simplify compliance, and safeguard your sensitive data.

3

Follow us on LinkedIn, X (Twitter), and YouTube for actionable expert insights on how to strengthen your data security, build a successful DSPM program, and more!

Before you go...

Get the Gartner Customers' Choice for DSPM Report

Read why 98% of users recommend Sentra.

White Gartner Peer Insights Customers' Choice 2025 badge with laurel leaves inside a speech bubble.