
How to Assess Third-Party Risk: A Practical Framework
Third-Party Risk Risk Management
What Is Third-Party Data Risk? Third-party data risk is the chance that an external vendor will cause a security …

Learn eight strategies to stop insiders from stealing sensitive data before it leaves your network.
• 59% of departing employees take confidential business data with them, according to a Ponemon Institute study. Most IP theft happens within 30 days of resignation. Tighten monitoring and limit access as soon as someone gives notice.
• Malicious insider attacks cost $4.92 million per incident on average. But negligent insiders cause more total damage because there are so many more of them. Training and credential hygiene prevent most negligent incidents.
• When employee credentials show up in stealer logs, it means a device they used is infected with malware. That’s both an insider threat and an external attack vector. Credential monitoring catches this before attackers exploit the access.
• No single control stops insider threats. Least privilege access limits the damage, DLP tools catch exfiltration in progress, and offboarding procedures close the departure window. You need all three.
A Google engineer stole thousands of self-driving car files before joining Uber. A Tesla engineer took Autopilot source code to a Chinese competitor. Both were caught, but the damage was already done.
Insiders don’t need to break through your perimeter. They already have valid credentials and authorized access to sensitive systems.
Defending against employees requires a different playbook than defending against external attackers. You can’t just block access. You need to detect when it’s being misused.
This guide covers eight strategies to prevent employee data theft, the warning signs to watch for, and what to do when you suspect someone is stealing data.
The term covers more than you might think. It’s not just a disgruntled employee walking out with files on a USB drive.
Employee data theft is the unauthorized taking of company data by current or former employees. It covers deliberate theft by malicious insiders and accidental exposure by negligent ones. It also includes cases where attackers use compromised employee credentials to access and steal data without the employee’s knowledge.
The numbers are worse than most companies realize. The Ponemon Institute found that 59% of departing employees take confidential business data with them. The DTEX/Ponemon 2026 insider risk report puts the average annual cost of insider risk at $19.5 million per organization.
Not all insider threats look the same. There are five types to watch for.
Negligent insiders accidentally create security risks. They fall for phishing and reuse passwords. This is the most common type and accounts for 53% of total insider risk costs.
Malicious insiders deliberately steal or leak data. Their motivations range from financial gain to revenge. They’re dangerous because they know your systems from the inside.
Compromised insiders have had their credentials stolen by external attackers through phishing or infostealer malware. The attacker uses the employee’s access without their knowledge.
Departing employees pose unique risks during their transition out. The next section covers why this group needs its own strategy.
Third-party insiders include contractors and vendors with privileged access. The Verizon 2025 DBIR shows that internal actors are involved in 30% of breaches globally, with partner actors showing increased privilege misuse.
This is where most employee data theft happens. According to Carnegie Mellon’s CERT program, 70% of intellectual property theft by insiders occurs within 30 days of their resignation announcement.
Why? Some employees take data as an “insurance policy” for their next role. Others want a competitive edge at their new employer. Some don’t even think of it as theft. They consider the client list they built or the code they wrote as “theirs.”
Anthony Levandowski (Google/Waymo to Uber). Before leaving Google, Levandowski downloaded 14,000 confidential files related to self-driving car technology. He founded Otto, a self-driving truck startup that Uber acquired. Waymo sued, and Uber settled for approximately $245 million in equity. Levandowski was sentenced to 18 months in prison and ordered to pay $756,499 in restitution.
Guangzhi Cao (Tesla to Xpeng). A Tesla engineer copied over 300,000 files of Autopilot source code to his personal iCloud and a thumb drive before leaving to join Chinese EV competitor Xpeng. Tesla sued under the Defend Trade Secrets Act. The case settled with Cao making a monetary payment.
These aren’t edge cases. They’re the pattern. Senior employees with access to IP leave for competitors and take what they can carry.
The period between when an employee gives notice and their last day is the highest-risk window. Access is still active, but loyalty has shifted. Some employees download data the same day they resign.
Your offboarding process needs to account for this. We’ll cover specific controls in the prevention section below.
Catching insider threats early requires monitoring for both behavioral and technical indicators.
Unusual work hours. Employees accessing systems late at night or on weekends when they normally don’t may be trying to avoid detection.
Job dissatisfaction. Employees passed over for promotion or given negative performance reviews are at higher risk.
Resignation announcement. The notice period is high-risk. Watch for changes in data access patterns immediately after someone gives notice.
Financial stress. Sudden financial problems can motivate employees to sell company data.
Accessing files outside their job function. An accountant downloading engineering documents should trigger an alert.
Unusual data volumes. Large downloads or bulk file transfers, especially to external drives or cloud storage, need investigation.
Unauthorized USB devices. Personal USB drives connected to corporate systems are a classic exfiltration method.
Disabled security tools. Employees turning off endpoint protection or DLP tools may be preparing to move data.
Credential anomalies. Credentials used from unusual locations or at unusual times may indicate compromise rather than malicious intent. Check for insider threat indicators to know what to look for.
No single control stops all insider threats. But layering these eight strategies covers most of the ways data walks out the door.
Limit each employee’s access to only what they need. This constrains the damage any single insider can cause.
Use role-based access controls that adjust automatically when employees change roles. Run quarterly access reviews to catch privilege creep. For administrative access, use just-in-time provisioning that grants rights only when needed and revokes them automatically.
DLP tools monitor and control data movement. They catch both intentional theft and accidental leakage.
Content inspection examines files for sensitive data before they leave your network. Policy enforcement blocks transfers that violate your rules, like uploading customer lists to personal cloud storage. The goal is catching exfiltration in progress, not just logging it after the fact.
When employee credentials appear on the dark web, it usually means something has gone wrong that you don’t know about yet.
Credential monitoring continuously scans third-party breach data and infostealer logs for your employees’ exposed credentials. When passwords appear in stealer output, it often means the employee’s device is infected with malware that’s harvesting everything they type or have saved in their browser.
Credentials leak through two main channels. Third-party breaches expose passwords when employees reuse them across services. Infostealer malware harvests credentials directly from infected devices.
Compromised credential monitoring detects this exposure early. When an employee’s credentials appear in stealer logs, it signals that their device may be infected. That’s both an insider threat and a potential external attack vector.
The departure window is where most data theft happens. Your offboarding process needs to close that gap.
Revoke access the moment termination is decided, not after the employee’s last day. For resignations, limit access during the notice period to only what they need to finish their work. Collect all devices and wipe corporate data from personal devices.
Run exit interviews that remind employees of their confidentiality obligations. Most won’t steal data if they know you’re watching and that there are legal consequences.
Monitor activity during the notice period specifically. Watch for unusual downloads and bulk file transfers. Emails to personal accounts are another red flag.
Most insider incidents stem from negligence, not malice. Training addresses the root cause.
Run regular phishing simulations. Require password managers so employees use strong, unique passwords everywhere. Build a reporting culture where employees flag suspicious behavior from coworkers without fear of retaliation.
For more on the human errors that lead to breaches, see our guide on how human error causes data breaches.
UEBA tools establish what normal looks like for each employee, then alert when behavior deviates.
An employee accessing unusual files after announcing their resignation is more concerning than either signal alone. UEBA connects these weak signals into stronger indicators. Risk scoring helps your security team focus on the highest-priority alerts instead of drowning in noise.
Without clear policies, you can’t hold employees accountable for data theft.
Define what employees can and can’t do with corporate data. Be specific about personal devices and cloud storage. Classify your data so employees know what’s sensitive. Document consequences for violations and enforce them consistently. Annual policy acknowledgment creates an audit trail that matters if things end up in court.
Dark web monitoring catches when your data or credentials show up on criminal marketplaces. This detects both compromised insiders (employees whose credentials were stolen) and the aftermath of malicious insider activity.
If an employee sells or leaks data, it often surfaces on dark web forums or marketplaces. Internal tools like DLP and UEBA catch exfiltration in progress. Dark web monitoring catches what already got out.
When you suspect an employee of stealing data, loop in legal and HR before you do anything else.
Gather evidence discreetly. Review access logs and email records. Document everything. Premature confrontation can lead to evidence destruction.
Involve legal counsel and HR early. Employment law varies by jurisdiction. You need guidance on permissible investigation methods and required documentation.
Preserve forensic data. Don’t power off suspect devices. Create forensic images before any analysis. Ensure logs aren’t overwritten. Maintain chain of custody documentation for everything you collect.
Assess the scope. Determine what data was accessed and whether it left the network. Track where it went. This informs remediation and any notification obligations.
Revoke access if the risk is high. Disable their accounts and reset passwords. For ongoing investigations, weigh this against tipping off the employee before you’ve collected enough evidence.
Consider law enforcement. For trade secret theft or large-scale data exfiltration, early engagement makes prosecution more viable. Levandowski got 18 months in prison. Prosecutors do pursue these cases when the evidence is clear-cut.
For a broader view of breach response, see our data theft prevention guide.
Employee data theft requires a different defense than external attacks. Insiders already have access. Your job is making sure they use it appropriately and detecting when they don’t.
The eight strategies above aren’t theoretical. The companies that catch departing employee data theft early are the ones with least privilege access and active monitoring during notice periods. Credential exposure detection running in the background catches the rest.
Start with the highest-impact controls: lock down access and monitor the departure window. Add credential monitoring to catch compromised insiders. Those cover the scenarios behind most employee data theft incidents.
Check your exposure to see if employee credentials from your organization are already circulating on criminal markets.
Negligent insiders cause more incidents than malicious ones. The Verizon 2025 DBIR found that unintentional errors happen roughly twice as often as deliberate insider schemes. However, malicious insiders cause more damage per incident because they target high-value data intentionally.
Watch for unusual access patterns: large downloads and after-hours activity. Accessing files outside their job function is another red flag. DLP and UEBA tools automate this. Data breach detection tools flag anomalies before data leaves the network.
Gather evidence discreetly before confronting them. Involve legal counsel and HR early. Preserve forensic data including access logs and email records. Revoke access immediately if the risk is high, but balance this against tipping them off before you’ve collected evidence.
Yes. In the US, the Computer Fraud and Abuse Act (CFAA) criminalizes unauthorized access. The Defend Trade Secrets Act covers IP theft specifically. Other countries have similar laws. Criminal charges are real. One engineer got 18 months in prison for stealing self-driving car files.
Common methods include emailing files to personal accounts and copying data to USB drives. Cloud uploads and screenshots are popular too. 70% of IP theft happens within 30 days of resignation. Strong data leak prevention and monitoring during the notice period catch most exfiltration attempts.
DLP software blocks unauthorized transfers. UEBA detects anomalous behavior patterns. Credential monitoring catches when employee passwords appear in stealer logs, which often means an infected endpoint. No single tool covers everything. Layer them.

Third-Party Risk Risk Management
What Is Third-Party Data Risk? Third-party data risk is the chance that an external vendor will cause a security …

What Is Business Email Compromise? It goes by several names, but they all describe the same scam. Business email …