Using AI Features

import { Tabs, TabItem } from ‘@astrojs/starlight/components’;

AI features overview

Klarify’s AI features are designed to make creating and managing business processes faster and more effective. These tools assist with content creation, suggest improvements, and help you discover relevant information, while always keeping you in control of the final decisions.

Think of AI assistance as having an experienced colleague who can quickly suggest ideas, identify patterns, and help with routine tasks - but you always make the final decisions about your content.

What AI can help with in Klarify

Content creation assistance:

• Suggest task descriptions and process steps
• Recommend element names and labels in process models
• Generate documentation templates and starting points
• Propose organizational structures for complex processes

Process improvement insights:

• Identify potential bottlenecks or inefficiencies
• Suggest standard industry practices for common workflows
• Recommend process model improvements based on best practices
• Alert you to inconsistencies or missing elements

Discovery and navigation:

• Enhanced search results with intelligent ranking
• Content recommendations based on your role and activities
• Related process suggestions when viewing documentation
• Pattern recognition across your organization’s processes

Enabling AI assistance

AI features are enabled by default, but you can control their visibility:

1. Navigate to Settings → AI Preferences

2. Choose your assistance level:

   • Full assistance - All AI features enabled
   • Suggestions only - AI recommendations without automatic actions
   • Minimal - Only basic content enhancement
   • Disabled - No AI features active

3. Customize specific features:

   • Content generation suggestions
   • Process modeling assistance
   • Search enhancement
   • Recommendation systems

Tip

Start with “Suggestions only” to experience AI assistance without automatic changes. You can increase the assistance level as you become comfortable with how the features work.

Content generation assistance

AI-suggested task descriptions

When creating global tasks or process elements, AI can suggest comprehensive descriptions:

How it works:

1. Enter a basic task title like “Review customer application”
2. AI analyzes the title and suggests detailed descriptions
3. Review suggestions and choose elements that fit your process
4. Customize the content to match your organization’s specific requirements

Example AI suggestions for “Review customer application”:

## Review Customer Application

**Purpose:** Evaluate customer application for completeness and accuracy

**Prerequisites:**
- Access to customer management system
- Current application evaluation guidelines
- Authority to approve/reject applications

**Steps:**
1. Open customer application in the system
2. Verify all required fields are completed
3. Check supporting documentation is attached
4. Validate customer information against requirements
5. Document decision and reasoning
6. Update application status in system

**Quality checks:**
- All mandatory information provided
- Supporting documents are current and valid
- Decision reasoning is clearly documented

Adapt suggestions to your organization:

• Add specific system names (Salesforce, SAP, etc.)
• Include your approval criteria and thresholds
• Reference your organization’s policies and procedures
• Add role-specific responsibilities and authority levels

Enhance with local knowledge:

• Include contact information for escalations
• Add links to relevant forms and templates
• Reference training materials specific to your industry
• Include compliance or regulatory requirements

Process documentation suggestions

Template generation: AI can suggest document structures based on process type:

• Standard operating procedures
• Approval workflows
• Customer service processes
• Technical procedures

Content enhancement:

• Suggest additional steps you might have missed
• Recommend quality checkpoints and validation
• Propose error handling and escalation procedures
• Offer formatting improvements for clarity

Template recommendations

Industry-specific templates: AI recognizes common business patterns and suggests appropriate templates:

• Customer onboarding - Templates for different industry types
• Invoice processing - Variations for different approval levels
• Incident response - Templates for various severity levels
• Change management - Structured approaches for organizational changes

Process modeling help

Element naming suggestions

Consistent naming conventions: When you add elements to process models, AI suggests names that:

• Follow BPMN best practices
• Match your organization’s existing naming patterns
• Use clear, action-oriented language
• Maintain consistency across related processes

Smart name suggestions:

• Tasks: “Review customer data” instead of “Customer review”
• Events: “Payment received” instead of “Payment”
• Gateways: “Is application complete?” instead of “Application check”

Process flow recommendations

Workflow optimization: AI analyzes your process models and suggests improvements:

Common suggestions include:

• Missing error handling - Paths for when things go wrong
• Parallel opportunities - Tasks that could happen simultaneously
• Decision clarity - More specific gateway conditions
• Completeness checks - Missing start or end events

Best practice alerts:

• Unconnected elements - Process elements not properly linked
• Complex gateways - Overly complicated decision logic
• Missing documentation - Elements without adequate descriptions
• Inconsistent patterns - Deviations from established conventions

Best practice alerts

Real-time feedback while modeling:

• Yellow indicators for minor suggestions
• Orange warnings for potential issues
• Red alerts for serious problems or errors
• Green confirmations when following best practices

Example alerts:

• “Consider adding an error path from this task”
• “This process might benefit from parallel execution”
• “Gateway condition could be more specific”
• “Missing approval step for financial threshold”

Caution

AI suggestions are based on general best practices and pattern recognition. Always evaluate recommendations against your organization’s specific requirements and constraints.

Search and discovery

AI-enhanced search results

Intelligent result ranking: AI improves search by considering:

• Your role and department - Prioritizes relevant content
• Previous activity - Learns from your usage patterns
• Content relationships - Finds related processes you might need
• Team collaboration - Surfaces content from your colleagues

Smart query interpretation:

• Natural language understanding - “How do we handle refunds?” finds refund processes
• Synonym recognition - “Customer” matches content tagged with “Client”
• Context awareness - Search results adapt based on your current work

Content recommendations

Personalized suggestions:

• “You might also need…” - Related processes for your current task
• “Popular in your department” - Content colleagues find useful
• “Recently updated” - Fresh content relevant to your work
• “Training recommendations” - Skill-building content for your role

Related process suggestions

Discovery connections: When viewing any process, AI suggests:

• Prerequisites - Processes that typically happen before this one
• Follow-up processes - What usually comes next
• Alternative approaches - Different ways to achieve the same outcome
• Supporting documentation - Global tasks and reference materials

Cross-functional insights:

• Upstream dependencies - What other departments provide as inputs
• Downstream impacts - How your process affects other teams
• Integration points - Where your process connects to systems or external parties

Data privacy and AI

How your data is used

Privacy protection principles:

• Your content stays private - AI analysis happens within your organization’s data
• No cross-organization learning - AI doesn’t share insights between different organizations
• Aggregate analysis only - Individual user behavior isn’t tracked for AI training
• Secure processing - All AI analysis follows the same security standards as your other data

What AI analyzes:

• Content patterns - Structure and organization of your processes
• Usage patterns - What content gets accessed frequently
• Naming conventions - How your organization labels and categorizes
• Process relationships - How different workflows connect

Opting out of AI features

Granular control options:

1. Global AI disable - Turn off all AI features organization-wide
2. Individual feature control - Keep some AI assistance while disabling others
3. Personal preferences - Each user can set their own AI assistance level
4. Content-specific settings - Disable AI for sensitive or confidential processes

Steps to adjust AI settings:

1. Navigate to Settings → AI Preferences
2. Choose your preferred level of AI assistance
3. Configure specific features individually
4. Set content sensitivity levels for different process types

Compliance considerations

For regulated industries:

• AI suggestions can be logged and audited like other system actions
• Human approval required for all AI-suggested changes
• AI assistance can be disabled for compliance-critical processes
• Audit trails include information about AI involvement in content creation

Providing feedback on AI suggestions

Improving AI accuracy

Rating suggestions:

• Thumbs up/down on individual suggestions
• “This helped” or “Not useful” feedback
• Detailed comments on why suggestions did or didn’t work
• Feature requests for additional AI assistance areas

Training AI for your organization

Pattern recognition improvement:

• Use suggested names when they fit your needs
• Customize suggestions to match your preferred style
• Provide alternative phrasings when suggestions are close but not perfect
• Flag organization-specific terms that AI should learn

Feedback channels

How to provide feedback:

• In-line rating on individual AI suggestions
• Feedback forms accessible from AI preference settings
• Support tickets for detailed feedback or issues
• User research participation for improving future AI features

Tip

The more you interact with AI suggestions - accepting good ones and declining poor ones - the better the system becomes at providing relevant assistance for your organization’s specific needs.

Advanced AI usage tips

Getting the most from AI assistance

Content creation workflow:

1. Start with AI templates to get basic structure
2. Customize suggestions to match your organization’s voice and requirements
3. Use AI naming for consistency across team members
4. Leverage search enhancement to find related content more efficiently

Quality improvement process:

1. Review AI alerts during process modeling
2. Consider best practice suggestions even if you choose not to implement them
3. Use AI insights to identify process improvement opportunities
4. Share AI discoveries with team members for organization-wide benefit

Combining AI with human expertise

Effective collaboration patterns:

• AI provides suggestions - Human makes final decisions
• AI identifies patterns - Human interprets significance
• AI handles routine tasks - Human focuses on strategy and exceptions
• AI surfaces options - Human chooses based on organizational context

When to rely more heavily on AI:

• Routine content creation with well-established patterns
• Initial drafts that will be reviewed and refined
• Consistency checks across large volumes of content
• Discovery of content and connections you might miss

When to rely more on human judgment:

• Strategic process decisions that affect business outcomes
• Sensitive or confidential content requiring careful handling
• Complex organizational politics or relationship considerations
• Industry-specific requirements not covered by general AI training

Ready to explore more advanced features? Learn about organizational structure management or dive into access control and permissions to secure your AI-enhanced content.