New AI Tool Builds ANY Automation in Minutes (Live Demo)

Key Concepts

• Building Automation Frustration: The gap between the perceived ease of automation and the technical expertise required.
• AI-Native Automation: Automation platforms designed with AI integration as a core feature.
• MCP (Managed Code Prompts) Nodes: A way to call APIs by describing what you want in natural language, simplifying complex API interactions.
• Subflows: Reusable, modular components within a larger workflow, promoting efficiency and collaboration.
• Loop Mode: Automatically iterating a node over a list of inputs, enabling batch processing.
• Smooshing: Combining a list of items into a single text string, often separated by a delimiter.
• Gummy: Gumloop's AI assistant that helps users build and understand workflows.

Gumloop vs. Traditional Automation Tools

• Main Point: Gumloop differentiates itself from tools like Zapier, Make, and N8N by natively integrating AI and simplifying complex tasks.
• Key Differences:
  • AI-First Approach: Gumloop assumes AI will be used in workflows, providing tools for easy AI integration.
  • List Handling: Gumloop simplifies working with lists of data, enabling looping and aggregation.
  • Usability: Gumloop aims to be intuitive for both technical and non-technical users.
• Aaron's Perspective: "Gum Loop is different in a few assumptions. The first assumption is that AI is kind of native and that your workflows are just going to use AI."

Gumloop Pricing

• Free Plan: Offers a limited number of credits for experimentation. Running flows is free unless external models or providers are used.
• Starter Plan ($97/month): Provides 30,000 credits.
• Team Plan ($300/month): Offers more credits and team functionality.
• Enterprise Plan: Tailored for organizations with specific security and user management needs.
• Credit Usage: Credits are consumed when using AI models or external providers.

Live Build: Research Assistant Workflow

Overview

• Goal: To create a workflow that identifies external meeting attendees, gathers information about them, and generates a report.
• Process:
  1. Input Node: Takes an attendee's email as input.
  2. Extract Domain: Uses an AI step to extract the domain from the email address.
  3. Perplexity Search: Searches Perplexity for recent news about the company.
  4. Apollo Enrichment: Uses Apollo to enrich the contact with professional information.
  5. HubSpot Tickets: Retrieves relevant HubSpot tickets associated with the contact.
  6. Combine Text: Combines the information from Perplexity, Apollo, and HubSpot into a single text string.
  7. Subflow (Do Research): Encapsulates steps 2-6 into a reusable subflow.
  8. Google Calendar MCP Node: Retrieves emails of external attendees from Google Calendar.
  9. Loop Mode: Iterates the "Do Research" subflow over each external attendee.
  10. Join List Items (Smooshing): Combines the individual research reports into a single text string.
  11. Ask AI: Uses AI to generate a formatted report based on the combined research.
  12. Email: Sends the generated report via email.

Step-by-Step Breakdown

1. Input Node:
   • Defines the input for the workflow (attendee's email).
   • Example: [email protected]
2. Extract Domain:
   • Uses the "Extract Data" AI node to extract the domain from the email.
   • Model: GPT 4.1 nano (1 credit).
   • Prompt: "Extract the domain from the email."
   • Example: Input: [email protected], Output: gumloop.com
3. Perplexity Search:
   • Searches Perplexity for recent news about the extracted domain.
   • Search Query: "Get recent information about gumloop.com"
4. Apollo Enrichment:
   • Uses a custom-built MCP node to retrieve contact information from Apollo.
   • Input: Attendee's email.
   • Includes current position.
5. HubSpot Tickets:
   • Uses a custom-built MCP node to retrieve relevant HubSpot tickets.
   • Input: Attendee's email.
   • Returns a maximum of 10 tickets as text.
6. Combine Text:
   • Combines the outputs from Perplexity, Apollo, and HubSpot into a single text string.
   • Includes labels for each section (e.g., "Recent Company Information," "Person Details," "Recent Tickets").
7. Subflow (Do Research):
   • Encapsulates steps 2-6 into a reusable subflow.
   • Input: Attendee's email.
   • Output: Combined research report.
8. Google Calendar MCP Node:
   • Uses a custom-built MCP node to retrieve emails of external attendees from Google Calendar.
   • Prompt: "Return all emails of external attendees for meetings in the next 24 hours."
   • Includes meeting details.
   • Filters out emails that include gumloop.com.
9. Loop Mode:
   • Automatically iterates the "Do Research" subflow over each external attendee retrieved from Google Calendar.
   • Enables batch processing of multiple contacts.
10. Join List Items (Smooshing):
    • Combines the individual research reports from the "Do Research" subflow into a single text string.
    • Separator: New line (\n).
11. Ask AI:
    • Uses AI to generate a formatted report based on the combined research.
    • Model: GPT 3.5 Turbo (0.3 credits).
    • Prompt: (Copied and pasted from a pre-written prompt generated by Gummy).
    • Includes meeting title information.
12. Email:
    • Sends the generated report via email.
    • Recipient: [email protected].
    • Subject: "Your daily report for today's meetings."
    • Sends as HTML.

MCP Node Creation and Customization

• Process:
  1. Describe the desired functionality in natural language (e.g., "Return all emails of external attendees for meetings in the next 24 hours").
  2. Gumloop analyzes the prompt and suggests inputs and outputs.
  3. Confirm the suggested inputs and outputs.
  4. Gumloop generates static code that calls the MCP tools.
  5. Test the node and request changes if needed.
  6. Gumloop regenerates the code based on the requested changes.
  7. Save the node for future use.
• Benefits:
  • Simplifies complex API interactions.
  • Reduces the need for technical expertise.
  • Ensures consistent output.

Subflows and Reusability

• Benefits:
  • Promotes modularity and code reuse.
  • Simplifies complex workflows.
  • Enables collaboration and sharing of building blocks.
• Example: The "Do Research" subflow can be reused in other workflows, such as researching new contacts from an email list.

Interfaces and Forms

• Functionality: Gumloop allows users to create interfaces and forms that trigger workflows.
• Benefits:
  • Makes workflows accessible to non-technical users.
  • Simplifies data input and workflow initiation.
  • Enables the creation of custom applications.
• Example: A form can be created to research a specific contact by entering their email address.

Gumloop's Internal Automation

• Commit Summarization: Automates the process of summarizing GitHub commits for email updates.
• Internal Linking: Uses AI to identify cross-linking opportunities within Gumloop's documentation and blog posts.
• Enrichment and Notifications: Automates the process of enriching user data and sending notifications based on user activity.

Key Takeaways

• Gumloop aims to bridge the gap between the expectation and reality of building automation by providing an AI-native platform that simplifies complex tasks.
• MCP nodes enable users to interact with APIs using natural language, reducing the need for technical expertise.
• Subflows promote modularity and code reuse, enabling the creation of complex workflows from reusable building blocks.
• Gumloop's internal automation demonstrates the power of the platform for streamlining business processes and improving efficiency.
• The platform's focus on usability and accessibility makes it suitable for both technical and non-technical users.
• Aaron: "Everyone's a builder now. You can just do things now."