Skip to main content

Administration

Manage users, security, and platform settings for AI for Process.

Workspaces

Overview

Workspaces organize workflows, models, and users:
Workspace: Finance Automation
├── Workflows
│   ├── Invoice Processor
│   ├── Expense Approver
│   └── Report Generator
├── Models
│   ├── Invoice Classifier (fine-tuned)
│   └── Expense Categorizer
├── Prompts
│   └── Extraction Templates
└── Users
    ├── Developers (5)
    ├── Reviewers (10)
    └── Admins (2)

Create Workspace

  1. Navigate to AdministrationWorkspaces
  2. Click Create Workspace
  3. Configure:
    • Name and description
    • Default settings
    • Initial members

User Management

Roles

RolePermissions
AdminFull workspace access
DeveloperCreate/edit workflows, models, prompts
ReviewerProcess Inbox tasks
AnalystView analytics and reports
ViewerRead-only access

User Enrollment

Add users:
  1. Go to User ManagementUsers
  2. Click Add User
  3. Enter details:
    • Email
    • Name
    • Role
  4. Send invitation

Teams

Group users for assignment:
Team: Invoice Reviewers
Members:
  - alice@company.com
  - bob@company.com
  - carol@company.com
Inbox access:
  - Invoice Processing
  - Expense Approval

Integrations

External Services

Connect to external services:
ServicePurpose
Hugging FaceOpen-source models
S3 / Cloud StorageModel and data storage
Weights & BiasesExperiment tracking
Custom APIsExternal integrations

Configure Integration

Integration: Hugging Face
API Key: hf_...
Settings:
  auto_sync: true
  sync_interval: daily
  import_models:
    - pattern: "company-org/*"

Custom Scripts

Overview

Add custom Python scripts for workflows:
# custom_scripts/data_processor.py

def process_invoice(data):
    """Process invoice data with custom logic."""
    result = {
        'invoice_number': data['raw_number'].strip(),
        'amount': float(data['amount'].replace('$', '')),
        'is_valid': validate_invoice(data)
    }
    return result

def validate_invoice(data):
    """Custom validation logic."""
    return data['amount'] is not None

Script Management

ActionDescription
UploadAdd new scripts
EditModify existing scripts
VersionTrack script versions
TestTest scripts in sandbox

OCR Model Management

OCR Providers

Configure OCR for document processing:
ProviderFeatures
Built-inGeneral document OCR
Google VisionHigh-accuracy OCR
AWS TextractForm and table extraction
CustomCustom OCR models

Configuration

OCR Settings:
  default_provider: google_vision
  fallback: built_in
  language_hints:
    - en
    - es
  preprocessing:
    deskew: true
    denoise: true

Monitoring & Analytics

Workflow Metrics

MetricDescription
ExecutionsTotal runs
Success rate% completed
Avg durationExecution time
Queue depthPending tasks

Model Metrics

MetricDescription
RequestsTotal API calls
LatencyResponse time
Token usageInput/output tokens
CostTotal spend

Dashboards

Pre-built dashboards:
  • Workflow performance
  • Model usage
  • Inbox metrics
  • Cost analysis

Security

Authentication

MethodDescription
SSOSAML, OIDC integration
API KeysService authentication
OAuthUser authentication

SSO Configuration

SSO:
  provider: okta
  protocol: saml
  entity_id: https://kore.ai/saml
  sso_url: https://company.okta.com/sso
  certificate: |
    -----BEGIN CERTIFICATE-----
    ...
    -----END CERTIFICATE-----

Authorization

Role-based access control:
Authorization:
  workflow_access:
    - role: developer
      actions: [create, read, update, deploy]
    - role: reviewer
      actions: [read, execute]
  model_access:
    - role: developer
      actions: [create, read, update, deploy]
    - role: analyst
      actions: [read]

API Scopes

ScopeAccess
workflows:readView workflows
workflows:writeCreate/modify workflows
workflows:executeRun workflows
models:readView models
models:writeDeploy models

Billing & Usage

Usage Tracking

ResourceUnit
Workflow executionsPer run
AI model tokensPer 1K tokens
StoragePer GB
ComputePer hour

Cost Management

Cost Management:
  budget:
    monthly: $5000
    alert_threshold: 80%
  quotas:
    workflow_executions: 10000/month
    model_tokens: 10M/month
  notifications:
    - email: finance@company.com
    - slack: #ai-costs

Billing Dashboard

View:
  • Current usage
  • Cost breakdown by resource
  • Historical trends
  • Projected costs

Data Management

Data Retention

Configure retention policies:
Retention:
  workflow_logs:
    duration: 90 days
    archive: true
  inbox_tasks:
    duration: 1 year
  model_artifacts:
    duration: indefinite

Data Export

Export data for compliance:
  1. Go to Data ManagementExport
  2. Select data types
  3. Choose format (JSON, CSV)
  4. Schedule or download

Data Deletion

Request data deletion:
  1. Submit deletion request
  2. Specify data scope
  3. Confirm compliance requirements
  4. Execute deletion

Audit Logs

Logged Events

CategoryEvents
AuthenticationLogin, logout, SSO
WorkflowsCreate, deploy, execute
ModelsDeploy, update, delete
UsersAdd, remove, role change
DataExport, delete

Log Access

Audit Logs:
  retention: 7 years
  export:
    format: json
    destination: s3://audit-logs/
    schedule: daily