Configure the Blog Page Agent to customize how it generates blog posts based on your organization's requirements, content, and brand guidelines.
Navigate to the Blog Page Agent
Go to Content Hub.
Under Agents, select Blog page.
Select the Agents tab.
Open the Library tab to view and configure the agent workflow.
Hover over the Blog page agent and select Use agent.
The Blog Page Agent is available by default and should be further configured to suit your requirements. The workflow consists of multiple nodes, each responsible for a specific step in the blog-generation process.
NOTE:
Some nodes include a Context field that you should configure based on your requirements. The following sections explain how to configure each node in the workflow.
The Blog Page Agent contains multiple nodes that control how the agent researches, structures, writes, and sends the generated blog to the Content hub.
Configure the Blog page agent
Click the Blog page agent.
Select the Brand identity you want the agent to use.
Review the Goals and Outcomes. These fields are populated by default, but you can update them to align with your requirements.
Click Save.
NOTE:
The selected brand identity determines the locations associated with the agent.
Configure the trigger
The On demand Blog generation trigger starts the agent when a blog topic is provided.
The trigger is preconfigured, so there are no settings you need to change here. You can review the trigger description, then continue to the next node
Configure the Parse images task
The Parse images task reads and processes images attached to the blog topic. You can configure the task to provide additional context and instructions for how the agent should interpret and describe images.
Click Parse images to open the Action panel.
Under the Context tab, click Add to provide additional information like your brand voice, industry knowledge, and agent-specific context that can help the agent generate more accurate responses.
You can add:
Fields — Custom business fields and other available information.
Knowledge — You can upload or select PDF, DOC, or DOCX files up to 10 MB each from the Knowledge base.
Brand — Select relevant brand information, such as Brand Profile, Target customer, Style and Voice, Media, Guardrails, and click Save.
Review the Fields. The default inputs include Images and PDF images, which provide the images that the task processes.
Review the User and System prompt, which defines how the task interprets and describes the images.
Review the Prompt Strength indicator at the bottom to evaluate the quality of your prompt. Click View Suggestions to see recommendations for improving it. Apply relevant suggestions to your prompt as needed.
Under User prompt, add specific instructions for how the agent should handle the image content.
Click Save to save your changes.
Tip: Use Context to provide information the agent should use while processing the task, and use the System prompt and User prompt to define how the task should be processed and describe the input.
Gather supporting content
Use Gather supporting content to provide the agent with additional information and references that can help it generate accurate, relevant blog content.
Open Gather supporting content.
Review and update the Context in the action panel:
Add relevant files from Knowledge to provide additional reference material.
Select relevant Brand elements, such as Brand Profile, Target customer, Style and Voice, and Guardrails.
Add other available context as needed.
3. Review the User prompt and System prompt to understand or refine the instructions.
4. Click Save to save the updated configuration.
Tip: Add context that is directly relevant to the blog topic, such as supporting documents and brand information. This gives the agent the information it needs to produce more accurate and brand-aligned content.
Research and strategy
The Research and strategy node researches the blog topic and locks the keyword strategy that will guide the blog generation process. It uses the provided blog details and other inputs to determine the appropriate content strategy before the blog is written.
1. In the Prompts tab, review the User prompt and System prompt. The prompts are configured with Strong prompt strength and define how the task researches the topic and determines the keyword strategy. You can customize the prompts if you want to change how the task processes the provided information. Click Save to save your changes.
2. In the Fields tab, review the information used by the agent. Click Add to include additional input or output fields if needed.
3. Under Context, review the information provided for the task. By default, Brand Profile is available. Click Add to include additional context if needed.
Web page reader
The Web page reader reads the URLs provided by the Blog Page Agent and extracts their content for analysis. This helps the agent understand the source material and use relevant information when generating blog content.
Extract high-performing phrases
The Extract high-performing phrases action analyzes the provided citation URLs and extracts high-performing phrases and language patterns relevant to your blog topic. You can review and customize the instructions under the Basic, Prompts, Fields, and Context tabs to control how the action processes the input and generates its output.
Find SEO keywords
The Find SEO keywords node finds related SEO keywords and their average monthly search volume based on the seed keywords provided by the workflow. These keywords can then be used as part of the blog's keyword strategy.
The Seed keywords provided to the tool are used as the starting point for finding related SEO keywords.
Cluster and map keywords
The Cluster and Map Keywords action groups related keywords by theme and matches them to search intent. It filters out irrelevant keywords and classifies the remaining keywords based on the blog's intent, objective, and funnel stage.
You can review and customize the instructions under the Basic, Prompts, Fields, and Context tabs. Because the default prompt has Strong prompt strength, no changes are required unless you want to tailor the task to your specific content strategy.
Build blog outline
Creates the blog structure with SEO and answer engine optimization (AEO) in mind, using the configured content and keyword strategy.
Review the User prompt to see how the blog topic, content depth, keywords, intent, funnel, and other configuration values are passed to the node.
Review the System prompt to understand how the node structures the blog outline based on the provided inputs. The prompt strength is Strong by default, but you can customize the prompt if required.
Review the Fields and Context used to provide the information required for outline generation. Click Add to include additional context or inputs if needed.
Recommend blog images
This node suggests relevant images for the AI-generated blog based on the blog content and configured media sources.
Click Recommend blog images to open the node.
Media sources: Select the sources the agent can use for image recommendations. You can enable or disable sources and drag them to change their priority; the agent checks the sources from top to bottom.
Write blog
Click Write blog to open the node.
In the Action panel, review the configured Fields and Context. Click Add to include additional context or input fields if needed.
Review the User prompt to understand how the blog topic, content depth, images, heading style, word count, internal links, intent, and other inputs are passed to the node.
Review the System prompt, which defines the instructions the agent follows when generating the blog.
The Prompt Strength is set to Strong by default. No changes are required, but you can customize the prompts if needed.
Click Save after making any changes.
The Write blog node uses the outline, research, keywords, recommended images, and other configured inputs from earlier steps to generate the full blog draft.
Send blog to Content hub
Click Send blog to Content hub to open the node.
Review the Blog content, Reasoning, and Blog metadata fields to confirm that the outputs from the previous steps are mapped correctly.
No additional configuration is required for this node. Review the mapped fields and click Save if you make any changes.
This node adds the generated blog to the Content hub project, using the blog content and metadata produced by the preceding workflow steps.
Test the agent
After configuring the agent, run a test to verify how the workflow processes a blog topic and generates the blog.
Click Run test.
In the Preview settings window, enter a Blog topic. You can also add optional information to provide additional context for the blog and select your preferred language: English, French, Spanish, or German. The agent will generate the blog content in the selected language.
Click Run test.
The agent runs the configured workflow and displays the outcome of each task in the Test details panel.
Review the test results
The Test details panel shows the results produced by each task in the workflow. Depending on the task, you can expand sections such as Result or View inputs to review the information processed or generated by that task.
For example, the test results can show:
The blog topic provided to the agent.
Image analysis from Parse images.
Supporting sources collected by Gather supporting content.
Research and strategy details, including intent, funnel, target word count, and content depth.
Content extracted from citation URLs by Web page reader. In the test results, you can review the task's status and expand View inputs to inspect the URLs provided to it.
High-performing phrases and language patterns identified by Extract high-performing phrases from the citation URLs. Expand Winning Phrases to review the extracted results.
SEO keywords and their search volumes.
Primary and secondary keyword clusters.
The generated blog outline.
Recommended blog images.
The final blog generated by the Write blog task.
You can expand the available sections under each task to inspect the corresponding inputs and outputs.
After expansion
Once you've reviewed the test results and confirmed that the agent is producing the expected output, you can exit the test and proceed with the agent's next step, such as activating or saving the agent as a draft.
Next steps
Once you have configured and tested the Blog Page Agent, you can use it to generate blog posts.
Read: How to Create Blog Posts with the Blog Page Agent to learn how to generate and review blog posts using the configured agent.