Sensitive information identification
Detect and classify personal and sensitive information in documents and databases
Data Governance · Compliance
Connects identification and de-identification of sensitive data, role-based access control, and quality, genealogy, and audit trail under one standard. Leverage data and AI safely, even in regulated environments.
OVERVIEW
Rather than controlling data to block it, we create standards for safe use.
By linking the location, purpose of use, accessor, and change history of sensitive data into one policy, we enable data and AI to be utilized in actual work even in regulated industries.
We consistently apply authority, de-identification, quality, and audit standards to the entire process of collecting, processing, sharing, and utilizing data.
De-identification and application of permissions
Apply masking and role-based access policies to the data layer
Quality · Genealogy · Audit
Track and document the source, change, and access history of data
Safe use of AI
Control to ensure that only permitted data is used for RAG and AI queries
Who used it, when and what data was usedStructure that can be explained and provenCreate .
CHALLENGES
Even if you have data, It is difficult to share and utilize safely.
If sensitive information protection, departmental authority, audit trails, and AI usage standards are separated, data utilization slows down and the burden of regulatory response increases.
There are no standards for data sharing across departments and roles.
Because personal information and sensitive information are mixed, it is difficult to provide only the necessary information to those who need it.
It is difficult to prove the access history required for regulatory audits.
Logs for each system must be re-collected to see who viewed and changed what data and documents and when.
I am anxious about using internal data for AI and LLM.
If the original text containing sensitive information is used as is in the search, learning, and inquiry process, there is a risk of leakage and misuse.
It is difficult to track the process of data creation and change.
It is difficult to demonstrate data integrity and quality as sources, cleansing processes, personnel, and change history are not linked.
SOLUTION
From authority to audit trail, we internalize control standards for data use.
Policies are applied consistently across the data and knowledge base layers, rather than being implemented separately for each application.
01
Access rights by group and role
The scope of inquiry, processing, and sharing is segmented according to organization, department, job, and user role.
02
Automatic detection and masking of personal information
Personal information is detected in documents and DB and deleted, replaced, and partially masked according to the policy.
03
Separation of public domain and knowledge base
Isolate datasets and knowledge bases and distinguish acceptable ranges according to department and business purpose.
04
Tamper-resistant audit logs
Access, inquiry, processing, and change history are recorded in tamper-proof format and used as audit trail.
05
Quality judgment and data lineage
Ensure integrity and traceability by linking data quality, provenance, processing steps, and change history.
06
CSV · GxP verification trace support
We support verification documents and operational evidence systems that take into account IQ, OQ, PQ and data integrity principles.
ARCHITECTURE
From source data to AI queries, an uninterrupted flow of authority and audit standards.
AkashiQ is responsible for collection, de-identification, quality, and genealogy, and SAMANDA provides safe AI queries within the permitted data range.
Governed Data-to-AI Flow
Data is provided to AI services in an auditable form after identifying sensitive information from the point of collection, applying permissions and scope of disclosure.
in-house documents
Regulations, contracts, reports, research materials
business database
Customer, quality, production, and operational data
system log
Access/processing/change/approval history
AkashiQ · GOVERNANCE LAYER
Collection, preprocessing, control
We apply protection, quality, and audit standards from the stage of storing data.
01
Privacy Detection
Automatic identification of sensitive information and regulated items
02
Automatic de-identification
Apply masking, replacement, and deletion according to policy
03
Isolation of knowledge base by department
Separate data by organization and authority group
04
Quality · Lineage · Audit Records
Accumulate by linking source, change, and access history
SAMANDA / RAG
Data-driven AI queries with permissions
Utilizing departmental data
Analysis, report, and search that fit the scope of work
Audit and regulatory trail
Check access history, genealogy, and quality results
Access permission control
Consistently apply data access rights and disclosure scope for each group and role.
Audit logs and lineage tracking
All inquiry, processing, and change history is recorded and even the original data is tracked.
DATA GOVERNANCE
AkashiQ
Responsible for governance, personal information masking, audit trail, quality determination, and data lineage.
GOVERNED AI QUERY
SAMANDA
Provides secure AI queries within the scope of data that is de-identified and subject to user permissions.
BEFORE / AFTER
In data that is stopped for security reasons, Transform your data into actionable data.
Before construction
- Restrict data sharing between departments due to sensitive information
- Different permissions and masking policies on different systems
- Manually collect logs and traces upon audit request
- Concerns about original text leakage and misuse when using AI
- Difficulty confirming data source and change process
After building
- Share safely according to the allowable range for each department and role
- Batch application of permissions and de-identification at the data layer
- Immediately utilize access and change history as audit trail
- Provide only permitted data to RAG and AI queries
- Prove integrity and traceability through quality and lineage
BUSINESS VALUE
Control for regulatory response becomes the foundation for safe data use.
01
Reduce regulatory audit response costs
Reduce audit preparation time by immediately checking access history and data lineage.
02
Secure transcription of sensitive data
Share data to the extent necessary through de-identification and role-specific permissions.
03
Secure the security premise of AI introduction
Only permitted data is provided to AI and RAG, reducing the risk of leakage and misuse.
04
Strengthening data integrity and traceability
Link sources and processing history to explain the reliability of data and the basis for changes.
05
Access control by department and role
Apply detailed data use policies that fit your organizational structure and business purposes.
APPLICATIONS
민감 데이터를 활용하면서 규제와 감사 요건까지 충족해야 하는 산업에 적용합니다.
데이터 접근과 변경을 통제하고 활용 이력을 추적해 안전한 데이터 활용과 규제 대응을 지원합니다.
제약 · 바이오 데이터 거버넌스
연구 · 생산 · 품질 데이터를 안전하게 활용하면서 데이터 무결성과 감사 추적성을 확보합니다.
- GxP · 21 CFR Part 11 compliance evidence
- 품질 · LIMS 데이터 접근권한 관리
- ALCOA+-based data lineage and integrity
- CSV 검증 문서 · 운영 기록 지원