Cisco Meraki Workflows: What It Is, Why It Matters, and Where to Start

The Problem Workflows Solves

Network engineers spend a disproportionate amount of time on tasks that are repetitive, procedural, and low-value — but still require expert knowledge to execute safely. Onboarding a new branch location. Applying a consistent security policy across 200 SSIDs. Replacing a failed switch and migrating its configuration. Generating a weekly compliance report.

These tasks aren’t intellectually demanding. They’re risky because they’re manual, slow because they require human coordination, and error-prone because they depend on people following runbooks correctly every time.

Cisco Workflows exists to eliminate this class of work entirely.

What Is Cisco Workflows?

Cisco Agentic Workflows is a cloud-hosted automation platform built directly into the Cisco Meraki Dashboard. It was announced at Cisco Live 2025 and became Generally Available in October 2025. It lives in the Automation menu in your Meraki Dashboard — no separate tool to install, no additional portal to log into.

The core idea: build multi-step automations across Cisco and third-party systems using a drag-and-drop visual designer, without writing code. Connect actions, decisions, loops, approvals, and AI steps into a single orchestrated workflow that runs on demand, on a schedule, via webhook, or through the Cisco AI Assistant in natural language.

Critically: you don’t need to be a Meraki customer to use Workflows. Any organization with an active license to any Cisco network management console can access it. And basic Workflows usage requires no additional license — it’s included.

What Makes It Different from Scripting

If you’ve been automating Meraki via Python and the Dashboard API, you might ask: why does this matter? I can already do all of this with scripts.

That’s true. But Workflows solves different problems:

Accessibility. A Python script requires someone who can write Python. A Workflow can be built, modified, and run by anyone on your team — including NOC engineers, help desk staff, and IT managers who don’t code. This is the difference between automation that lives in one engineer’s brain and automation that becomes institutional capability.

Visibility and governance. Scripts run in the dark. Workflows run in the dashboard with a full audit trail — who ran it, when, what parameters were used, what the outcome was. Every run is logged. Approvals can be built in, requiring a manager to sign off before a change executes.

Cross-domain orchestration. A Python script typically talks to one API. Workflows natively integrates with Meraki, Catalyst Center, Cisco ISE, SD-WAN Manager, Nexus Dashboard, and third-party systems — all in the same workflow. Stitching these together in code requires separate API clients, authentication management, and error handling. Workflows abstracts all of that.

AI integration. Any workflow you create is automatically exposed to the Cisco AI Assistant via MCP (Model Context Protocol). Ask the AI Assistant “set up a new guest SSID at the Chicago branch” and it identifies the matching workflow and executes it. Your validated, tested automation becomes a natural language interface.

The Four Ways to Trigger a Workflow

Workflows supports four execution modes:

Manual — Run directly from the Dashboard with a button click. Useful for tasks you want on demand but controlled: swap a device config, generate a report, push a policy change.

Scheduled — Run on a cron-like schedule. Useful for compliance reports, daily health checks, nightly config backups, weekly cleanup tasks.

Webhook — Triggered by an external event via HTTP POST. Meraki alerts, monitoring tools, ticketing systems, and custom applications can all trigger workflows. A client VPN disconnect alert triggers a remediation workflow. A Meraki MT sensor threshold breach triggers a notification workflow.

AI Assistant — Discovered and executed through natural language conversation. “Enable client isolation on all guest SSIDs in the Seattle office” becomes a one-sentence operation instead of a manual multi-step process.

The Workflows Exchange

The fastest way to get started isn’t building from scratch — it’s the Workflows Exchange. This is a built-in library of pre-built, Cisco-validated automations accessible directly from the Automation menu in your Meraki Dashboard.

Exchange workflows cover common scenarios like:

• Branch network provisioning
• Device lifecycle replacement (automatically copying configuration from a failed device to its replacement)
• Wi-Fi setup with ISE authentication
• Security policy enforcement across sites
• Compliance and inventory reporting
• Alert-driven remediation

You can run Exchange workflows as-is, or open them in the drag-and-drop editor and customize them for your environment. This is the difference between automation being a six-month project and being a same-day capability.

Key Concepts You Need to Know

Targets — The systems your workflow connects to. A Meraki organization, a Catalyst Center instance, an ISE server, or a third-party API endpoint. You define targets once and reference them across multiple workflows.

Adapters — Pre-built integrations for specific systems. Adapters define what actions are available for each target (e.g., the Meraki adapter includes actions like “Get network clients,” “Update SSID settings,” “Claim device to network”). If an adapter doesn’t exist for a system, you can use the generic HTTP Request activity to call any REST API.

Activities — Individual steps in a workflow. Activities can be API calls (using adapters), conditional logic (if/else branches), loops (iterate over a list of networks), delays, human approval steps, AI LLM actions, or variable assignments.

Variables — Data that flows through a workflow. The output of one activity becomes the input of the next. A workflow that fetches all networks in an org can iterate over that list, applying a configuration to each one in turn.

AI LLM Actions — A step that sends data and a prompt to an AI model (like Claude or GPT) and uses the response to make decisions or generate content. This is what Cisco means by “AgenticOps” — workflows that don’t just execute predetermined steps, but reason about data and choose next actions dynamically.

AgenticOps: What Cisco Is Building Toward

The term AgenticOps appears throughout Cisco’s Workflows documentation and represents their vision for the next phase of network operations: AI agents that autonomously reason, decide, and act — not just execute scripts.

The practical meaning today: Workflows + AI LLM Actions + AI Assistant together create a system where:

1. An event occurs (device goes offline, client count spikes, a compliance check fails)
2. A webhook triggers a workflow
3. The workflow gathers context (what changed, what the normal baseline is, what other affected devices look like)
4. An AI LLM Action analyzes the context and recommends or chooses a remediation path
5. The workflow executes the remediation
6. The result is logged and optionally surfaces a summary to the AI Assistant

This is significantly different from traditional automation, which executes a fixed decision tree. The AI step can handle ambiguity, synthesize multiple inputs, and make a contextual judgment call — the same thing a senior engineer would do, but in seconds.

Who Should Use Workflows

MSPs managing multiple customers. The cross-organization visibility and templated Exchange workflows make it practical to apply consistent automation across every customer environment, not just the ones large enough to justify custom scripting.

Enterprise IT teams without dedicated automation engineers. Workflows removes the Python/scripting requirement. Your NOC team can build and run automations themselves, without waiting for a dedicated automation engineer.

Organizations already using Cisco Catalyst Center or ISE. The cross-domain integration means Workflows can orchestrate changes that touch Meraki, campus switching, and identity simultaneously — something that previously required multiple separate tools or complex custom scripts.

Anyone doing repetitive network management tasks manually. If you’re doing the same sequence of dashboard clicks more than once a month, that task belongs in a workflow.

Getting Started

1. Log into your Meraki Dashboard and navigate to Automation → Workflows
2. Browse the Exchange — find a workflow that matches something your team does manually today
3. Click Install and run it in your environment
4. If it works, schedule it or connect it to a webhook trigger
5. When you’re ready to build custom workflows, start the YouTube Training Series linked in the Meraki documentation — it walks through everything from first workflow to advanced AI activities

The best first workflow is the one that replaces a task your team actually does this week. Find that task, automate it, and build from there.