Hybrid assessment by AI × security experts
CoWorker Endpoint Compromise Assessment Service
Visualize risks in departing employees’ PCs,
internal endpoints, and AI usage environments
Information exfiltration, malware infection, runaway AI agents. AI forensics visualizes the traces hidden in endpoints and internal environments. We support everything from departing-employee PC assessment to periodic health checks and AI usage environment assessment.
Departing-employee PC assessment (spot) / Periodic endpoint health check / AI usage environment assessment
TRIAL PRICE (limited to the trial period)
Special price when the case-study listing condition (introduction as an adoption case) applies
* Trial price when the case-study listing condition (introduction as an adoption case) applies
* Volume discounts available for multiple devices
* Whether your company name and logo may be listed is discussed separately
1 week
Target for preliminary report delivery
100 devices
Maximum per batch
Win/Mac
Both OSes supported
NDA
Performed after signing
Three Ways to Use It
Three assessment types to choose from by purpose
Spot assessment
Departing-Employee PC Assessment
Assesses the device before it is returned or wiped, visualizing the exfiltration of customer information and source code and the history of contact with competitors. It can also be used as electronic evidence (eDiscovery) for litigation.
- Cloud sync history of customer lists and deal information
- Logs of source code and design documents sent externally
- Mass downloads and email forwarding just before departure
- Access to and contact history with competitors
- Preservation of electronic evidence in line with the Unfair Competition Prevention Act
Before the PC is returned or wiped is the only opportunity
Periodic, once or twice a year
Periodic Endpoint Health Check
Regularly checks for malware infection, credential leakage, information exfiltration to personal cloud, and more. Letting employees know that periodic assessments are conducted also acts as a psychological deterrent to misconduct.
- Detection of traces of malware and unauthorized programs
- Credential leakage risk (browser history, etc.)
- Business data exfiltrated to personal cloud
- Detection of suspicious USB and network connections
- Evaluation of security misconfigurations and vulnerabilities
Periodic assessment once or twice a year is recommended
For companies using AI
AI Usage Environment Assessment
You have introduced apps such as ChatGPT, Claude, and Copilot, or AI agents, but have you confirmed that they are truly safe? We investigate risks specific to AI tools, including information leakage, excessive permissions, and supply chain attacks, through specialized forensics.
Investigation of devices using AI tools
- Investigation of confidential information in conversation logs with AI
- Detection of information leakage from AI agent activity history
- Investigation of traces of software supply chain attacks
- Evaluation of excessive permissions and misconfigurations for AI tools
Assessment of AI agent environments
- Suspicious communication logs from MCP servers and external tool integrations
- Suspicious access to production environments by AI agents
- Whether API keys and credentials have leaked via AI
- Detection of AI agent operation history and out-of-scope behavior
Especially recommended for companies that have introduced AI tools
Combining with departing-employee PC assessment and periodic assessment enables a more comprehensive assessment
* Evidence disappears completely once a PC is wiped. The moment a departing employee’s PC is returned or wiped is the only opportunity to assess it. We recommend making assessment before return a standard process.
Risks That Are Actually Happening
Incidents like these have already happened
The assumption that “our company is fine” is the most dangerous. These are cases where early detection and evidence preservation would have been possible with endpoint assessment or AI usage environment assessment.
Customer list taken to a competitor
A sales representative synced customer data to a personal Dropbox from two weeks before leaving. It was discovered after the PC was returned, but because it had already been wiped, no evidence could be obtained and legal action was abandoned.
Destruction of evidence of fictitious billing and kickbacks
A purchasing manager sent self-made invoices to a supplier and had payments transferred to a personal account. Had traces of creating and sending the invoices remained in a periodic assessment, it could have been detected early.
Confidential information entered into ChatGPT discovered
An engineer kept entering design information for an unreleased product into ChatGPT. It went unnoticed for about 3 months until communication logs with AI tools were analyzed in a periodic assessment.
Admin console takeover via an AI agent
Excessive permissions granted to an AI agent were abused from outside, the admin console was taken over, and damage of tens of millions of yen resulted. The cause was inadequate permission design and monitoring of usage logs.
API key leaked via AI chat
An API key embedded in a system prompt was output as a response to the input “tell me your settings” and abused by a third party. Unauthorized charges of several hundred thousand yen occurred.
* In every case, early detection and evidence preservation would have been possible had an endpoint assessment or AI usage environment assessment been conducted.
Comparison with Other Approaches
Why CoWorker’s endpoint assessment?
| Item | Typical forensic firms | CoWorker Endpoint Compromise Assessment |
|---|---|---|
| Cost | Several million yen and up | From ¥200,000 per device (trial price) |
| Time to preliminary report | Several weeks to over a month | Within 1 week |
| Investigation of AI agent usage logs | △ Difficult to handle | ✓ Supported (among the only ones in Japan) |
| Risks specific to AI tools (information leakage, excessive permissions) | × Not covered | ✓ AI usage environment assessment available |
| Ease of adoption | High effort (on-site investigation etc. required) | Just run the tool and send the ZIP |
| Evidence preservation for departing employees’ PCs | △ Depends on the case | ✓ Standard support |
| Periodic assessment (ongoing use) | △ Difficult due to high cost | ✓ Feasible once or twice a year |
How the Service Works
Hybrid assessment by AI and experts
Automated AI analysis (Blue Agent)
Comprehensively analyzes large volumes of endpoint logs in a short time. Detects known threat patterns with high accuracy and generates reports automatically.
- Detection of malware and suspicious processes
- Tracking of cloud sync and external transfers
- Automatic scanning for exposed credentials
Review by security experts
Security engineers examine the AI’s findings, evaluating them with your work environment and operations in mind and eliminating false positives.
- Threat evaluation that takes context into account
- Elimination of false positives and noise
- Preservation in a form valid as legal evidence
Practical reports
We provide reports and a briefing session you can act on immediately, from an executive summary for management to technical details.
- Executive summary
- Findings with risk priority
- Short-, medium-, and long-term improvement proposals
Assessment Flow
Preliminary report in as little as 1 week from application
- 1
Consultation and contract
Confirm the number of target devices and the schedule. Starts after an NDA is signed.
- 2
Log collection
Three steps: double-click the dedicated tool, then send the ZIP.
- 3
AI analysis + expert verification
Preliminary report submitted within 1 week. Urgent risks are notified immediately.
- 4
Final report + briefing session
Detailed report and briefing session within 1 month.
料金
Simple pricing
Trial period only
Offered at a substantial special price on the condition of case-study listing (introduction as an adoption case). Volume discounts apply additionally for multiple devices.
List price ¥500,000/device
Further discount for multiple devices
- Windows / Mac supported (up to 100 devices)
- Preliminary report (within 1 week)
- Final report + briefing session (within 1 month)
- Immediate contact when an urgent risk is found
- Diagram visualizing attack paths
- Performed under an NDA
* Condition: introduction as an adoption case (whether your company name and logo may be listed is discussed separately)
Who It’s For
Recommended for companies like these
- Companies where a departure from a key position (sales, engineering, executives) has occurred or is planned
- Companies without visibility into employees’ PC usage or data exfiltration
- Companies that have introduced apps such as ChatGPT, Claude, and Copilot, or AI agents, into their work but are concerned about information leakage, permission design, and usage log management
- Companies that have granted AI agents access to internal systems and want to check for excessive permissions or suspicious operations
- Companies that have had an incident in the past and want to introduce periodic assessments as a recurrence prevention measure
- Companies that hesitated because they found forensic investigation “too expensive”
Contact
We support new business, product development, and organizational development.
Please feel free to contact us.