Ben's Bites
← Blog

How to use AI to automate your life and work

Max Brodeur-Urbas on how Gumloop’s modular AI workflows are transforming productivity across professions.

Event Recap free
Topic: automationTopic: productivity Role: ai-machine-learning-engineer Category: productivity-workflow-optimization

Published 2024-12-20

Max Brodeur-Urbas didn’t set out to create just another AI tool. He wanted to build something that truly changes things for anyone dealing with automation.

And after chatting to him in a recent hour-long live workshop, it’s clear he’s done that with Gumloop.

Max isn’t afraid to challenge the status quo. During our chat, he revealed how his team took on the task of simplifying what many see as the future of work—AI-driven workflows. But there’s a twist: Gumloop doesn’t just throw AI at every problem. Instead, it focuses on using AI smartly, making sure you get the most out of all your tools.

In this workshop, Max walked us through:

  • how Max’s experience with AutoGPT and the challenges users faced led to the creation of Gumloop
  • live demonstrations of workflows that:
    • turn YouTube videos into SEO-optimised blog posts using well-structured AI prompts
    • automate the process of researching and qualifying leads after sign-up, and
    • research the people Max is meeting with that day and send him a concise report every morning.

The workshop is a great watch for anyone looking to simplify their workflow, automate complex tasks, or make better use of AI in their daily operations.

Paying subscribers get access to Max’s full recording and all future events.

Here’s a quick summary of the session with Max:

The Gumloop Philosophy: Less AI, more control

A refreshing and unconventional idea Max shared was Gumloop's philosophy: less AI, more control. It’s a perspective that flies in the face of the current trend of automating everything with AI.

Max shared how this approach was born out of his early experiences with AutoGPT, where users were often disappointed by the over-reliance on AI to handle tasks that could have been managed more efficiently with a simple Python script or a few API calls.

“A counterintuitive thing we do at Gumloop is we encourage people to use as little AI as possible,” Max explained. “This drives the cost down, increases reliability, and you can normally get 90% of the work done with just data manipulation, API calls, and targeted AI interventions.”

By focusing on reducing the reliance on AI, Gumloop offers users more control over their workflows. It also cuts down on costs (something any business can appreciate).

Max’s philosophy is all about finding that sweet spot where AI can provide the most value without becoming a crutch.

Making automated content creation easy

Max walked us through using Gumloop to turn a YouTube video into a polished, SEO-friendly blog post. This is the kind of workflow that saves hours for content creators and improves quality, without overcomplicating the process. It was neat to watch.

The workflow goes like this:

  1. Gumloop automatically pulls a transcript from a YouTube video.
  2. Then, instead of asking AI to write the entire blog post in one go - a method prone to errors -Gumloop breaks the task into manageable sections:
    • Intro - focused on setting the right tone and context
    • Main content - based on key points from the video
    • Conclusion - to tie everything together
  3. There’s a tone-checking step in the flow too, to make sure the creator’s tone is maintained throughout the post.
  4. Once the content is ready, Gumloop formats it into HTML and automatically publishes it to the creator’s CMS (in this case, Ghost).

The entire process, from video to blog post, takes minutes, freeing up creators to focus on what they do best: creating.

Qualifying and researching leads in real-time

Many of us know how much of a time-suck and repetitive task lead qualification can be. Max is trying to tackle this head-on with Gumloop’s ability to automate it all.

He shared how this workflow transformed his own sales process during Gumloop’s time in Y Combinator, making it easier to identify and connect with potential customers.

A quick summary of the workflow he demonstrated:

  1. As soon as someone signs up on Gumloop (a new lead) this workflow kicks in. It starts by grabbing the email address of the new user and immediately begins researching their company, including:
    • Company domain
    • Website scraping for key details like industry, revenue, and location
    • Data enrichment to pull in details about the company’s size, web traffic, etc
  2. Within seconds, a full report is sent to a dedicated Slack channel - the team can quickly identify high-potential leads.
  3. The workflow drafts a personalised email using GPT-4 that’s tailored based on the research. The user can send it after reviewing it.

No denying, this was a perfect example of how AI, when used correctly, can take over repetitive tasks, leaving you with more time to focus on the strategic parts of your business.

Building modular workflows with AI nodes

To wrap things up, Max walked us through Gumloop's modular workflow design. In Max’s words: "This modularity is really where the power of Gumloop comes in. You can test individual actions and then piece them together in the larger flow to get much more intricate automation, all without getting overwhelmed by complexity.”

The main takeaways were:

  • AI nodes as building blocks: Gumloop’s interface uses ‘nodes’ - essentially Lego-like building blocks for your workflows that can be connected in various ways. Each node performs a specific function, whether it’s extracting data from a document, analysing content, or integrating with other services like Google Drive or Slack.
  • Subflows: You can group several nodes together into a subflow, which acts as a mini-workflow within the larger process. This allows you to test individual components separately, ensuring they work correctly before integrating them into the full workflow.
  • Looping: You can loop subflows to process large datasets or perform repetitive tasks at scale.
  • Flexibility and customisation: Users can go deep into customisation in Gumloop. You can tweak individual nodes, modify AI prompts, and even integrate custom API calls or Python scripts.

If you’re serious about harnessing AI for automation, this is one session you won’t want to miss.

Paying subscribers can catch the full recording of Max’s workshop here.

Got an interesting use case or questions about how to get started with Gumloop? I’d love to hear from you. Reach out, and let’s keep the conversation going!

Enjoyed this post? You might like these: