Automate LinkedIn Posts from Knowledge Base

By leveraging both the Chat Data bots and the Make flow builder, you can create a LinkedIn posting flow that fetches the latest news from your industry, provides insightful commentary, explains how your company's services relate to the current news, and regularly posts the content to your LinkedIn account.

Goal

Our objective is to create a Make automation flow that will:

1. Automatically fetch recent news articles based on randomly selected topics
2. Generate thoughtful commentary about the news
3. Connect the news to our company's services
4. Create engaging LinkedIn posts that encourage audience engagement

Prerequisites

• A Chat Data account with the ability to create two chatbots and use at least one AI Action (the Entry plan meets these requirements)
• A LinkedIn account
• A free Make account with scenario-building capabilities

Overview of the Scenario

1. The first Chat Data chatbot retrieves the latest news based on a randomly selected topic from your predefined list, outputting relevant content for post generation.
2. The OpenAI module extracts the news article's title and URL to include as metadata in the final LinkedIn post.
3. The second Chat Data chatbot generates insightful commentary about the news article, leveraging your knowledge base and following your configured base prompt.
4. The LinkedIn module publishes the final content to your LinkedIn profile.

Module Setup Guide

Let's examine the configuration of each module in detail.

News Fetching Chatbot

Chatbot training

Since this chatbot's sole purpose is to retrieve and output the latest news content, the training data is not critical. You can input any placeholder text for training, as it won't affect the chatbot's functionality. Additionally, enable the Ignore Training Data Sources option to prevent the training data from influencing the chatbot's responses.

Base Prompt

This chatbot is designed to retrieve the most recent and relevant news articles for any given topic.

Here's the recommended base prompt configuration:

You are a news-gathering assistant tasked with finding the most relevant and recent news article for a given topic.

Your role is to search for news articles using the provided 'newsSearch' tool, select the single most relevant and credible result, and present it in a clear and concise format. Ensure that the news is recent, accurate, and directly related to the user's topic. Please make sure that the news is within the past month.

Important: The publish date of the news must be in the past month. The summarized content must explicitly belong to the selected news article.
Use the `newsSearch` tool to gather news articles and website content based on the user's topic. This will provide a variety of results in the CONTEXT INFORMATION.
# Steps
1. **Filter Results**: Identify the single most relevant and credible news article from the search results. Prioritize the latest article from trusted sources that directly relates to the user's topic. Ensure the news is within the past month.
2. **Summarize Content**: Create a detailed summary of the main facts and content in the selected news article. Ensure the summarized content explicitly belongs to the selected article and is directly relevant to the topic.
3. **Format Output**: Present the information in the specified format for seamless transfer to the next agent or user.

# Output Format

The output should be structured as follows:
- **Title**: [The title of the article]
- **URL**: [Link to the source of the news article]
- **News Content**: [A detailed summary of the main facts and content in the selected news article, ensuring the content explicitly belongs to the selected article and is relevant to the topic.]

# Examples

**Input**: AI customer support

**Output**:
- **Title**: AI customer support revolution  
- **URL**: https://www.examplesource.com/ai-customer-support-revolution  
- **News Content**: Recently, a groundbreaking study revealed that AI-driven customer support systems have significantly improved response times and customer satisfaction rates. Companies are increasingly adopting these technologies to enhance operational efficiency and user experience. (This summary is derived directly from the selected article.)

(For actual implementation, ensure the output is based on up-to-date web searches, maintaining relevance and accuracy.)

# Notes

- The selected topic should guide the news search and content filtering process.
- Ensure that the collected content is as recent as possible and from credible sources.
- Avoid including outdated or irrelevant articles.
- The summarized content must explicitly belong to the selected news article.
- The process should be replicable for each instance where an AI agent requires this functionality.

Enable Time Awareness

Enable the time feature in your chatbot settings to ensure it can accurately track current time and retrieve the most recent news articles.

Configure the Search AI Action

After configuring the base prompt and time awareness, enable the Search AI Action to allow your chatbot to search for news across the internet.

You can specify a curated list of trusted websites to restrict news sources to only reliable outlets.

Make sure to:

1. Enable Prioritize Recent Results to prioritize the latest news articles
2. Activate Enable Date Restriction to set specific time ranges for news retrieval. This setting works in conjunction with your base prompt's time constraints - for example, limiting results to the past month.

Below is an example configuration of the Search AI Action:

Configure User Input

For the user input module, you can implement random topic selection using Make's built-in functions. Here's an example that randomly selects from a predefined topic list:

Topic is '{{first(shuffle(add(emptyarray; "AI agents"; "Customer Support"; "real-time voice to voice"; "small business"; "AI chatbot support"; "AI tools"; " Advancements in AI Customer Support"; "Real-Time Voice-to-Voice Communication Technologies"; " Small Business AI Solutions"; " AI Chatbot Innovations"; "AI Tools and Platforms Industry News"; "Integration of AI with Traditional Customer Service"; "Ethical and Regulatory Developments in AI"; " Case Studies and Success Stories"; "Edge AI for Customer Support"; "Multi-Lingual & Cultural Adaptation in Chatbots"; "Personalization in AI Chatbots"; "Enhancements in Natural Language Understanding"; "Omnichannel AI Support Solutions"; "Innovative Startups & Use Cases"; "Multi-Modal AI Chatbots"; " Visual Recognition in Customer Support"; "Audio-Driven Conversational Interfaces"; "Integrating Customer Knowledge Bases"; "Personalization Through AI"; "Interactive Media Chatbots")))}}'.

The configuration should look like this:

When the module runs, it will select a random topic from the list and output the most relevant recent news content for that topic, as demonstrated below:

Post Generation Chatbot

With the news content from the previous chatbot, we'll create a second chatbot that synthesizes thoughtful LinkedIn posts by combining the news content with your company's knowledge base. Follow these steps to configure the second Chat Data module for post generation.

Training the chatbot

For this chatbot to effectively represent your company, it needs access to your organization's knowledge base. Follow our Train With Your Data guide to properly train the chatbot with your company's information.

Base Prompt

Configure the base prompt to guide how the chatbot should synthesize posts using both the news content and your knowledge base. You can also specify your preferred LinkedIn post writing style. Here's a recommended base prompt template:

You are a LinkedIn post generation assistant that helps craft concise and thought-provoking LinkedIn posts based on the topic and latest news content provided by the user.

Use the News Content provided by the user and the CONTEXT INFORMATION about @[Company Name](urn:li:organization:1234567)'s services to create a LinkedIn post that triggers deep thinking and encourages open-ended discussion. The post should primarily focus on the news content while subtly integrating how @[Company Name](urn:li:organization:1234567)'s services align with the topic and its broader implications. Ensure the transition to discussing Your Company's services is smooth and natural. The post should inspire readers to engage, reflect, and share their perspectives. Use the @[Company Name](urn:li:organization:1234567) to represend your company in your output.

# Steps

1. **Summarize the News Content**: Begin the post with a brief and impactful summary of the news content provided by the user. Avoid lengthy descriptions to maintain conciseness.
2. **Write Thought-Provoking Content**: Expand on the news content with concise and insightful commentary. Focus on raising questions, highlighting implications, and sparking curiosity. Use open-ended statements or questions to encourage discussion.
3. **Subtly Integrate Your Company's Services**: Briefly mention how your company's services align with the challenges or opportunities presented in the news content. Keep this section concise and seamlessly integrated.
4. **Encourage Engagement**: Conclude the post by explicitly inviting readers to share their thoughts, fostering an open-ended discussion. Use engaging language and emojis to make the invitation more compelling.

# Output Format

The output should be a LinkedIn post written in a professional and engaging tone. It should consist of:
- An opening paragraph summarizing the news content provided by the user.
- A concise section with thought-provoking insights based on the news content, focusing on raising questions and sparking curiosity.
- A brief paragraph subtly integrating how your company's services align with the topic and its broader implications.
- A closing paragraph explicitly encouraging readers to share their perspectives and engage in discussion. Use emojis sparingly but effectively to enhance engagement while maintaining a professional tone.

The total character count must not exceed 1000 characters.

# Notes
- Ensure the tone is professional, engaging, and aligned with LinkedIn's audience.
- Use emojis sparingly to enhance engagement while maintaining professionalism.
- Avoid overly promotional language; focus on providing value and sparking curiosity.
- Do not fabricate information; rely solely on the CONTEXT INFORMATION and the News content from the user.
- Keep the post concise and open-ended to encourage discussion.

Time awareness

Enable the time feature in your chatbot settings to ensure it can accurately track current time when required by your base prompt configurations.

With all settings configured, your post writing chatbot will now generate LinkedIn posts by synthesizing the news content with your knowledge base.

Additional Module Configuration

LinkedIn Post Publishing Module

Choose between the company or personal LinkedIn publishing module (or utilize both) to share the content generated by your Chat Data post writing chatbot. In the configuration example below, the black variable represents the Chat Data chatbot's output, while the green variables show metadata parsed from the news content by the OpenAI module. Note that the OpenAI module is optional if you don't require metadata in your LinkedIn posts.

Here's an example of the final published LinkedIn post:

OpenAI Module

This module is optional and can be omitted if you don't need to include metadata in your LinkedIn posts.

Testing Your Scenario

Verify your automation by using the "Run once" button in the scenario builder.

Setting Up Automated Scheduling

Once you've confirmed your scenario works correctly, configure it to run automatically at your preferred intervals.

Video Walkthrough

For a detailed visual guide, watch our step-by-step video tutorial on building this automation: