Known Limitations
Understand what the Relio can and cannot do. We're continuously working to improve these areas.
Version Note: This documentation reflects the current MVP release. Many limitations are actively being addressed in upcoming versions.
Tool Coverage & Integration Limits
GitHub
- Currently supports public repositories and private repos with proper permissions
- Does not yet integrate with GitHub Actions workflow metrics
- Limited support for monorepo structures with complex cross-repo dependencies
- Cannot access branch protection rules or review requirements
Jira
- Works best with Jira Cloud; Jira Server support is limited
- Custom fields are partially supported; complex field types may not sync correctly
- Does not track Jira Service Management (JSM) requests
- Limited integration with Jira Portfolio/work-at-scale features
Slack
- Cannot access message history from before integration was enabled
- Limited support for thread-based conversations
- Emoji reactions are not analyzed
- Does not integrate with Slack workflows or automations
Other Tools
- Currently does not support Asana, Monday.com, or other project management tools
- No native integration with Git platforms other than GitHub (GitLab, Bitbucket support planned)
- Email and calendar integration not yet available
AI Analysis Limitations
Context Window
The agent can analyze recent activity (typically last 2-4 weeks) with high accuracy. Historical trend analysis beyond this window is more limited.
Natural Language Understanding
The agent performs best with clear, explicit communication. Implicit or heavily sarcastic information in pull requests and comments may not be interpreted correctly.
Code Analysis
- Cannot perform deep code review or security scanning
- Does not understand team-specific coding patterns or conventions
- Risk assessment is based on PR metadata, not actual code diff analysis
Dependency Resolution
While the agent identifies task dependencies within your integrated tools, it cannot infer unstated or implicit dependencies that exist in your team's workflow.
Data & Privacy Considerations
Data Retention
Historical data is retained for up to 1 year. Older data may be archived or removed.
Sensitive Information
While the agent respects permissions from your connected tools, it may include sensitive context (like security vulnerabilities or internal discussions) in summaries. Always review before sharing reports externally.
Real-time Processing
Not all data is processed in real-time. Some analysis runs on a delayed schedule, so very recent changes may not be reflected in summaries immediately.
Team & Scale Limitations
Team Size
The agent performs best with teams of 5-100 people. Very small teams (<5) may not see significant patterns, while very large organizations (>500) may experience slower analysis.
Project Complexity
Works best with single-track projects. Highly complex matrix organizations or cross-functional structures may require manual configuration of project boundaries.
Multi-timezone Teams
- Standup and report times are currently global; per-timezone scheduling not yet available
- Agent assumes work follows a standard 9-5 weekday pattern
- May not detect patterns in async-first or heavily distributed teams
Automation & Workflow Limitations
Actions the Agent Takes
Currently, the agent is read-only. It analyzes and reports but does not:
- Create or close issues
- Assign work to team members
- Comment on pull requests or issues
- Change labels or status values
Custom Workflows
Cannot accommodate non-standard workflows. Works best with Agile/Scrum or Kanban methodologies.
Current Known Issues
Occasional Summary Inaccuracy
In rare cases, the AI may misinterpret complex scenarios or produce summaries that don't fully capture nuance. We recommend always reviewing agent summaries before sharing with stakeholders.
Integration Sync Delays
Occasionally, data from one integrated tool may be delayed in syncing to others (typically 5-15 minutes). This is temporary and resolves automatically.
Duplicate Detection
The agent may occasionally flag the same issue multiple times if it's referenced across different tools. We're working to improve deduplication logic.
Report Generation Performance
Reports for very large projects (500+ active issues) may take longer to generate during peak hours.
Workarounds & Recommendations
To Work Around Limitations
- For complex dependencies: Document critical dependencies explicitly in issue descriptions or Jira link types for the agent to detect them.
- For multi-timezone teams: Use Slack threads to group discussions by timezone and configure separate channels by region.
- For custom workflows: Map your workflow to standard Agile/Kanban status values in your tool configurations.
- For read-only limitations: Have the agent send recommendations to a specific Slack channel monitored by your project manager for action.
Best Practices
- Keep PR descriptions and issue details clear and explicit
- Use consistent labeling and status naming across your tools
- Regularly review and provide feedback on agent summaries
- Monitor agent output during the first week to calibrate expectations
What's Coming
We're actively working on addressing these limitations. Planned improvements include:
- Support for GitLab and Bitbucket
- Asana and Monday.com integrations
- Write-enabled actions (creating issues, updating status)
- Per-timezone standup scheduling
- Advanced code analysis and security scanning
- Custom workflow support
- Extended historical data analysis
Need Help?
If you're running into a limitation that blocks your workflow, we want to hear about it.
→ Contact Support or Send Feedback