One runtime for records, approvals, and AI-assisted work.
bontik combines records, governed workflows, AI-assisted review, files, and approval controls in one tenant-safe runtime built for production operations.
What powers the workflow
Generated records, forms, and files
The workflow runs on platform-owned records, forms, and packet files inside one runtime, not as notes layered on top of a separate tool.
Governed workflow runtime
Approvals, lifecycle control, reruns, and recovery are built into the same runtime that owns the record.
AI inside the control model
AI can summarize, route, and assist with execution without bypassing RBAC, approvals, or auditability.
Clean handoff and room to grow
The product owns review and routing, then hands off a complete approved packet while leaving room to support nearby approval-heavy work.
One control model across fixed and AI-assisted work
Repeatable approval paths
Teams can run fixed approval-heavy workflows where the path is defined once and executed many times.
AI-assisted review and routing
Teams can also let AI review, summarize, and route work inside the same governed runtime.
Shared permissions and state
Both patterns use the same record model, permission system, and workflow-state controls.
The same audit trail
The execution history stays attached to the record whether a step is fixed, human-approved, or AI-assisted.
Built to keep review, approvals, and handoff in one system.
The platform combines records, files, workflows, approvals, and AI-assisted execution so the packet stays in one operational system from intake through decision.
- Files, records, approvals, and history stay attached to the same request.
- Human decisions and AI-assisted steps follow the same permission and audit model.
- The runtime keeps control when workflows branch, pause, or change.
What the runtime proves
Role-based authorization
Users, AI-driven steps, and tool calls operate within the same permission model.
Approvals and lifecycle control
The runtime keeps state, pauses for approvals, and supports reruns and recovery.
Traceable execution
Actions, decisions, and workflow movement remain auditable from packet intake through final handoff.
What the runtime owns from start to handoff
The same governed runtime carries the record through every stage, including the parts where AI takes over.
Create the record
The packet, files, ownership, and state live in one platform-owned record instead of a loose coordination layer.
Run the review path
The same runtime supports AI-assisted review, request-more-info loops, and branch decisions without splitting execution into side tools.
Pause and approve with control
Approvals, reruns, recovery, and authorization checks remain part of the same workflow state model.
Hand off with history intact
The record, files, summaries, and execution trail stay intact even when the workflow ends with an export, notification, or the next operational setup step.
Where the runtime sits relative to the alternatives.
Most teams already use scripts, automations, or a generic workflow tool. Here is what shifts when the platform owns the governed record itself.
| Scripts and bots | Workflow tools | bontik | |
|---|---|---|---|
| Owns the governed record | No, lives in another system | Sometimes, often a thin overlay | Yes, the record is platform-native |
| Approval and recovery | Hand-rolled, often brittle | Available, separate from the record | Built into the same runtime |
| AI inside the control model | Bolted on as a side tool | Calls AI from outside the workflow | AI runs under the same RBAC, audit, and approvals |
| Audit trail across humans + AI | Reconstructed from logs | Partial, often manual to assemble | One execution history per record |
| Handoff to the next operational step | Custom coordination each time | Depends on the connector model | Approved package or notification with the record intact |
Platform questions
What makes the platform different from loose automation?
bontik owns the record, the workflow state, the files, and the approval history. It is not just a layer of prompts or disconnected automations.
Can fixed approvals and AI-assisted review run together?
Yes. Teams can mix repeatable approval paths with AI-assisted review, routing, and tool use inside the same permission, state, and audit model.
What can the platform support beyond vendor onboarding?
Beyond vendor onboarding, the same runtime can support supplier intake, procurement approvals, and other approval-heavy workflows that need governed records and clear handoffs.
Want the technical walkthrough behind the workflow?
We can show the runtime, the control model, and the approval path working together on a real vendor-onboarding flow.