We’ve been talking about “AI in the CMS” for years. Most of what existed before was autocomplete dressed up as intelligence. Sitecore Agentic Studio is something different, not because it’s magic, but because it’s structured, extensible, and built with the digital experience lifecycle in mind.

This post covers the full picture: what Agentic Studio is, why agents matter in a Sitecore context, the types of agents available, what marketing teams can do with it, and how developers can start building. I’ll also share lessons from my own experience building with the platform.
 

What Is SitecoreAI and Agentic Studio?

SitecoreAI is Sitecore’s unified AI platform, a single workspace designed to sit across the digital experience lifecycle. It supports everything from content creation and personalization to campaign planning and optimization. It is not a plugin or an add-on. It is the layer that connects Sitecore’s broader product ecosystem with AI-driven intelligence.

Agentic Studio is the part of SitecoreAI where the action happens. Launched at Sitecore Symposium 2025, it is the workspace where you build, manage, and run AI agents. Think of it as the operational hub for your AI workforce.

What makes it different from a standard AI assistant or chatbot integration is the concept of agentic AI. This is AI that does not just respond to prompts, but understands goals, plans steps, makes decisions, and executes tasks autonomously. It is the difference between asking an assistant what to write and assigning a teammate to draft the content, check it against brand guidelines, and save it as a versioned artifact.

Agentic Studio organizes this work across three core concepts:

• Agents are specialized AI workers built for a specific task or category of tasks.
• Flows are orchestrated sequences that chain multiple agents together, with logic, conditions, and human checkpoints.
• Signals are contextual data and preferences that ground agent behavior in your brand, content strategy, and business rules.

You can access Agentic Studio directly from the SitecoreAI navigation under Agentic, and you can also invoke it from the AI chat icon anywhere in the platform. That means agents are always one click away, regardless of where you are in your workflow.
 

Why SitecoreAI Needs Agents and When They Help

Not every AI task needs an agent. If you are generating a headline or summarizing a brief, a simple prompt may be enough. Agents become valuable when the work is multi-step, repetitive, or requires coordination across different inputs and outputs.

In the Sitecore world, a few patterns appear again and again:

• Content at scale. A large enterprise site may need hundreds of product pages, localized variants, or campaign assets refreshed quarterly. Manually prompting AI for each one does not scale. An agent can take structured inputs, apply schemas and templates, and generate outputs in a single run.
• Research-to-brief pipelines. Before a campaign brief exists, someone has to do competitive research, analyze keywords, and review the market landscape. Those are the kinds of tasks that can be delegated to a research agent.
• Content governance and quality checks. Once content is generated, it often needs to be validated against SEO requirements, brand tone, or AEO readiness. Agents can make those checks part of the workflow.
• Cross-functional coordination. When strategists, content teams, and developers are all contributing to a campaign, Agentic Studio’s Spaces feature helps keep artifacts, context, and history visible to everyone involved.

I built a Release Notes Generator using Agentic Studio, an agent that takes completed tickets, groups them by category, applies a consistent format, and produces release notes ready for publishing. What used to be a tedious manual task at the end of every sprint now takes just a fraction of the time.
 

Types of Agents in Sitecore Agentic Studio

There are three categories of agents in Agentic Studio, and choosing the right one is your first architectural decision.

1. Built-In Agents

These are prebuilt agents that ship with SitecoreAI and are designed for common marketing workflows. You do not need a Builder license to use them. You simply run them.

• Bulk Content Generator generates content at scale from structured inputs using your templates.
• AEO/SEO Researcher analyzes keywords, semantic clusters, and AI citation readiness.
• Account Data Enrichment enriches account records with research-driven context.
• ABM Campaign Agent helps plan and draft account-based marketing campaigns.
• Brief Generator turns campaign goals into structured creative briefs.
• Translation Agent supports localization across languages.

These agents are a strong starting point and a useful way to understand what a well-structured agent looks like before building your own.

2. Standard Agents

Standard agents are custom agents built for more flexible, chat-style interactions. They are not fixed pipelines. Instead, they respond dynamically based on the instructions, tools, skills, and context you configure.

With standard agents, you define:

• The agent’s instructions
• The tools it can use
• The skills it should apply
• The output format

Standard agents are best for conversational or exploratory work where a user may want to ask follow-up questions, refine outputs, or collaborate interactively with the agent.

3. Workflow Agents

Workflow agents are where developers will spend most of their time. These are structured, repeatable, multi-step pipelines built in the Workflow Editor, Agentic Studio’s visual drag-and-drop canvas.

A workflow agent runs a defined sequence of actions, passing data through variables from one step to the next. The five action categories available are:

• Context Parameters to extract, validate, and normalize structured data from user input
• Content Generation to run AI prompts and create outputs
• Research to pull from the web, analyze data, and enrich with external sources
• Flow Control to add conditions, loops, and branching logic
• Variables to store and pass state across the workflow

You can also chain workflow agents together, using the output of one as the input of the next. That is how more sophisticated multi-agent pipelines are built inside a coordinated flow.
 

How Marketing Teams Can Use Agentic Studio

You do not need to be a developer to get value from Agentic Studio. Marketers can run agents, configure inputs, review outputs, and collaborate without ever touching the Workflow Editor.

A typical marketing workflow might look like this:

• Campaign kickoff: A marketer opens the ABM Campaign Agent, enters the target account, product focus, and campaign goal, and receives a structured brief saved as a versioned artifact in a shared Space.
• Content production: The Bulk Content Generator uses that brief and a list of content needs, such as landing page copy, social posts, and email variations, to create structured drafts aligned to brand requirements.
• SEO and AEO optimization: Before publication, the AEO/SEO Researcher identifies keyword gaps, semantic opportunities, and answer engine optimization issues that can inform revisions.
• Human review at every step: This is one of the platform’s most important principles. Agents do not publish content without human review and approval. Teams remain in control of strategy and final decisions.

The result is that lean marketing teams can operate at a scale that previously required far more headcount. Agents do not replace people. They reduce repetitive work so teams can focus on higher-value thinking and decision-making.

How Developers Can Extend and Build in Agentic Studio
 

The Workflow Editor

The Workflow Editor is a no-code visual canvas, but it is built with developer-grade precision. Teams can:

• Chain actions with explicit variable passing
• Enforce structured outputs using JSON schemas
• Apply HTML templates for formatted rendering
• Set model parameters such as temperature
• Invoke external APIs within workflow steps
• Call Agent API tools for deeper Sitecore integrations

One lesson I learned quickly is that model behavior can vary significantly depending on which model is selected. In my Release Notes Generator work, one model added conversational text even when the schema required structured JSON output. Switching to a more controlled model and tightening the system prompt made the output much more reliable. The takeaway is simple: treat your agent prompt like production code. Version it, test it, and refine it.
 

MCP Integration

Agentic Studio supports MCP, or Model Context Protocol, servers. That means agents can connect to external tools and data sources. Sitecore’s Marketer MCP enables agents to take action directly inside SitecoreAI workflows, helping bridge Agentic Studio with the broader Sitecore ecosystem.

This opens the door to scenarios such as:

• Agents that pull from a CRM to enrich content decisions
• Agents that push content directly into Content Hub
• Agents that use CDP signals to tailor outputs
   

Agent API

The Agent API allows external systems to trigger and interact with agents programmatically. This is how Agentic Studio can connect into CI/CD pipelines, internal tools, and third-party orchestration platforms. It acts as the headless interface for your agent workflows.
 

Builder License

To create and modify agents in Agentic Studio, users need a Builder license. Viewers can run agents and review outputs. Builders can design workflows, create custom agents, manage schemas and templates, and configure shared settings.
 

Creating Your First Agent

A simple five-step pattern works well for most workflow agents:

1. Define your inputs. Start with the fields users will complete before the agent runs, such as a prompt, context, or file upload.
2. Extract and validate. Use a Context Parameters action to normalize user input into structured data.
3. Run the core AI action. Add a Content Generation or Research step with explicit instructions about format, tone, and structure.
4. Add flow control if needed. Use conditions, loops, or branching logic to support more advanced workflows.
5. Save and format the output. Store the result in a variable and optionally render it through an HTML template or save it as a versioned artifact.

A few practical tips help:

• Start small and test each action before adding the next
• Use descriptive action names to make debugging easier
• Name variables consistently so data flow stays clear
• Use lower temperature settings for more factual, structured output
   

Where This Is Heading

Agentic Studio is still relatively new, but it is evolving quickly. In a short period, the platform has expanded standard agent creation, added more workflow actions, improved collaboration through Spaces, introduced better visibility controls, and refined its built-in agents.

What stands out most is the composability. Marketers can build and run agents in a governed SaaS environment, while developers can extend those workflows through APIs and MCP integrations. That is a very different approach from simply bolting AI onto a CMS as an afterthought.

The human-in-the-loop philosophy is also worth calling out. Agentic Studio is designed to accelerate human work, not replace human judgment. That is the right model for a platform where content ultimately reaches live production environments and real audiences.

If you are a Sitecore developer and have not explored Agentic Studio yet, the best place to start is with the Workflow Editor. Build something small first, such as a content formatter, brief builder, or release notes generator. The learning curve is manageable, and the upside in terms of speed, consistency, and scale is real.