Inistate

@Inistate show me the schema for the 'support-tickets' module, including all activities and state flows.

The support-tickets module has 14 fields (ticket_id, subject, priority, assigned_to, customer_email, description, status, created_at, updated_at, resolution_notes, category, severity, sla_deadline, tags) and 5 states (new, assigned, in_progress, waiting_customer, resolved). Available activities include 'create_ticket' (requires subject, priority, customer_email), 'assign_ticket' (requires assigned_to), 'escalate' (requires severity, notes), and 'resolve' (requires resolution_notes). State flows allow transitions from new→assigned, assigned→in_progress, in_progress→waiting_customer or resolved, and waiting_customer→in_progress.

This example demonstrates schema introspection before writing data. Inistate modules are custom-defined per workspace, so field names and state flows vary. Always fetch the schema first when working with an unfamiliar module to avoid validation errors on create or update operations.

@Inistate create a new entry in 'project-tasks' with title 'Migrate legacy API endpoints', priority 'high', assigned to user ID 'u_4729', and due date March 15.

Entry created successfully with ID task_8821. Title: 'Migrate legacy API endpoints', priority: high, assigned_to: u_4729, due_date: 2025-03-15, state: backlog. Confidence score: 0.92 (above threshold). The record was not flagged and is immediately visible to the team. Created at 2025-01-15T14:32:18Z by AI assistant via MCP.

This example shows single-record creation with Inistate's confidence scoring. If the AI's field mappings had fallen below the module's confidence threshold (typically 0.85), the entry would be created but flagged for human review. Requires write permissions on the target module and valid user IDs for assignment fields.

@Inistate pull the last 50 history events for entry 'ticket_3394' and summarize what caused the SLA breach.

Ticket_3394 history (page 0, 50 events): Created Jan 10 at 09:14 (state: new), assigned to agent u_2201 Jan 10 at 11:03, transitioned to in_progress Jan 11 at 08:45, moved to waiting_customer Jan 11 at 16:22 with comment 'Requested log files', customer replied Jan 14 at 10:05 (3-day gap), returned to in_progress Jan 14 at 14:30, resolved Jan 15 at 09:12. Summary: The SLA breach occurred during the 3-day waiting_customer period (Jan 11-14) when the customer delayed providing logs. The ticket spent 28 hours in-progress across two intervals, but the waiting period pushed resolution past the 72-hour SLA.

This example pairs Inistate's audit trail with AI reasoning to diagnose workflow delays. History pagination is zero-based and capped at 50 events per page; longer trails require multiple calls. The AI traceability context in each event shows which assistant actions were automated versus human-initiated, useful for compliance audits.

When Inistate wins for regulated-team record-keeping

A 6-person medical-device startup needs every design-control change logged with AI provenance for FDA submissions. Inistate's Get Entry History tool returns 50-item audit pages that include field changes, comments, and AI traceability context in one call—exactly what auditors ask for. The MCP's confidence thresholds let the team flag low-confidence AI edits before they land in the official record, so human review happens inline instead of during post-hoc cleanup. If your compliance regime requires immutable logs and you're under 500 records per module, this MCP handles the entire chain of custody without custom scripting. Teams that don't face regulatory scrutiny won't need the audit overhead; stick with a simpler state-machine MCP.

When bulk state changes beat one-at-a-time workflows

A 3-person QA team closes 40 test cases every Friday after regression runs. Inistate's Bulk Change Entry States tool processes 1-100 transitions in one request and returns fail-soft per-item results, so one malformed case ID doesn't block the other 39. The MCP surfaces low-confidence flags inline, letting the team spot ambiguous transitions before they commit. If your workflow involves fewer than 10 state changes per session or you need atomic all-or-nothing commits, the bulk tools add complexity you don't need—use the single Change Entry State tool and loop in your prompt. For teams that batch-process dozens of records weekly, the bulk endpoint cuts latency and keeps the conversation moving.

When Get Activity Form saves round-trips during onboarding

A 5-person customer-success team onboards 15 new accounts a week, each requiring a multi-field intake form with conditional defaults. Inistate's Get Activity Form tool returns required fields, current values, options, defaults, and confidence thresholds in one call before the agent submits the Create Entry. That means the AI can prefill the form, validate required fields, and surface confidence warnings without a second round-trip to the user. If your intake forms are static or you onboard fewer than 5 accounts a month, the schema-fetch overhead isn't worth it—just hardcode the fields. For teams with variable or conditional forms, this MCP turns a 4-message back-and-forth into a single confirmation prompt.