Introduction
Graphiant AI-Ready Automation as a Service is the single place to handle all of your Graphiant network automation in the Graphiant Portal. You pick a task from a catalog, provide one YAML file, and Graphiant validates it and applies the changes across your network. You don't need to install or run any automation tools yourself.
It is built on graphiant-playbooks, the open-source automation framework for Graphiant. Anything you can automate with graphiant-playbooks, you can now run as a managed, template-based service.
It is also ready for agentic workflows from the start. AI copilots and orchestrators use the same catalog, validation and audit trail as your human operators, so you can bring AI into network operations without giving up control.
One stop for all your automation needs
Automation as a Service puts everything in one place, so you no longer have to piece it together from scripts, tools and manual steps:
- One catalog for configuration tasks such as VPNs, BGP and interfaces.
- One way to submit changes: a single YAML file and a playbook choice.
- One validation step that checks every change before it reaches the network.
- One pipeline view for stages, logs and status.
- One audit trail for every run, whether a person or an AI agent started it.
Refer Graphiant Automation Workflows for usage documentation.
Why Automation as a Service
Network changes often cover many devices and services, and doing each step by hand is slow and error-prone. Running automation yourself means maintaining toolchains, credentials and pipelines. Automation as a Service removes that overhead. You describe the outcome you want, and Graphiant handles running it.
Key capabilities
Abstracted operations
You don't configure each device one at a time. You select a playbook and submit a single YAML file that describes the change. Graphiant then applies it to every affected device and service in one coordinated run.
- One submission can cover many devices and services.
- You choose a playbook for the task, such as VPNs, BGP or interfaces.
- The YAML file describes what you want, not the steps to get there.
Shift-left safety
Every submission is validated before any change reaches your network. Graphiant checks your configuration first, so errors are caught and reported early instead of showing up as problems in production.
- Validation is mandatory and runs on every submission.
- Configuration errors are reported before any device is changed.
- Only configurations that pass validation go on to execution.
Developer-grade experience
Each run is shown as a CI-style pipeline, like the ones in GitHub or GitLab. You can follow a run from start to finish and see exactly what happened at each stage.
- Stages: see where the run is, from validation through execution.
- Logs: read detailed output for every stage.
- Status: see at a glance whether each stage passed, failed or is still running.
Infrastructure as a product
Complex network tasks are offered as simple catalog choices. Work that used to need deep device-level knowledge becomes a repeatable, standard service, and every run is recorded.
- The catalog covers common tasks such as VPNs, BGP and interface configuration.
- Templates keep configurations consistent across teams and sites.
- Every run is auditable: you can see who submitted what, when, and with what result.
Ready for agentic workflows
Automation as a Service is built for AI agents as well as people. A Model Context Protocol (MCP) server lets AI copilots and orchestrators work with the service directly. An agent can find the right playbook, validate a configuration and start a run, all through the same controlled interface people use.
- List: see which playbooks are available and what inputs each one needs.
- Validate: check a proposed configuration before anything is applied.
- Trigger: start a playbook run that shows up in the same pipeline view.
Playbooks are defined by clear YAML schemas and always validated before they run, so agents can use them safely and predictably with no custom integration work.
Unified guardrails
AI agents and human operators follow the same rules. There is no separate, less controlled path for automation.
- Same governance: the same permissions and policies apply whether a person or an agent makes the request.
- Same schemas: every submission is checked against the same YAML schemas and validation.
- Same audit logs: every run is recorded in one audit trail, whoever or whatever started it.
How it works
1. Choose a playbook from the catalog in the Graphiant Portal, or let an AI agent find one through the MCP server.
2. Submit your YAML file describing the change.
3. Validation checks your configuration before anything is applied.
4. Execution applies the change to all affected devices and services.
5. Review the pipeline's stages, logs and status, and use the audit trail later.
Who it's for
- Network engineers who want reliable, repeatable changes without scripting each device.
- DevOps and AI/automation teams who want to connect copilots and orchestrators to the network safely.
- Operations and compliance teams who need one audit trail of every change, by people or by agents.
Next steps
- Browse the playbook catalog in the Graphiant Portal.
- Review the YAML reference for the playbook you plan to use.
- Explore graphiant-playbooks on GitHub to see how the service works under the hood.