My Founder Story: How I Use AI to Reduce Anxiety

Foreword by Yuval, CEO of Glitter AI

A couple of years ago, I published a post on LinkedIn where I talked about the lessons I learned about self-worth after selling my first startup, Simpo.

The tl;dr of that post is that I felt extremely anxious while running Simpo because in my head Simpo = me, and its potential failure would have meant *I* was a failure.

Fast forward 3 years and hundreds of therapy sessions later, and I launched my new startup, Glitter AI.

In resolving many of my self-esteem / self-worth issues, I hoped that this time, I wouldn't feel so much anxiety running a company.

I was wrong.

However, this time, the source of anxiety was different -- it was rooted in a feeling of overwhelm: being buried with too much to do, too much context-switching, and too much pressure trying to stay on top of everything. In short: too much.

So I figured: I'm going to *AI* the shit out of it.

In this post, I detail the exact system I built to manage it all, and regain a sense of freedom.

If you're a founder as well, or just someone who wears many hats and is trying to stay afloat, I invite you to read further and learn how you, too, can take back control.

Yuval
Founder & CEO, Glitter AI

p.s. after you're done here, I invite you to read my post about how to reduce anxiety as a founder by delegating.

Background

My name is Yuval, and I run a small new startup called Glitter AI. For reasons reserved for another post, I chose not to fundraise for my startup this time around, so it's still a REALLY small operation.

In fact, as of writing, it's just me full-time, and a couple of freelancers. As a result, I end up doing A LOT of work myself that in my previous startup, I hired people to do.

Today I'd like to take you behind-the-scenes to show you how I keep up with everything that needs to get done, and keep my anxiety in check.

The Overwhelm After Winning Product Hunt

After I launched on Product Hunt and Glitter AI won "Product of the day", I ended up feeling like I was drowning.

Winning Product Hunt meant that I went from practically zero traffic and zero messages on Intercom to hundreds overnight. Requests were coming in left and right: bugs, feature suggestions, questions, and a ton of general feedback.

All as free-form text, with nothing categorized.

I remember sitting there, staring at the stream of tickets, thinking, "How the f*** am I going to keep up with all this?" If I didn’t figure out a way to manage it all, I knew I’d either miss important details or just burn out completely.

Neither of those options seemed particularly appealing, so I got to work.

The Plan: Create My Dream Assistant

My plan was to create my dream assistant that would make sense of all the noise and help me feel back in control.

I figured that with what large language models (LLMs) can do nowadays, I can at the very least roll something up that would summarize and categorize all the noise for me.

I set out to do that with a combination of Make.com, OpenAI, and Notion.

If you're unfamiliar, Make.com (a competitor of Zapier) is a tool that allows you to glue together a bunch of APIs (e.g. OpenAI, Notion, Intercom, and many more...) to automate stuff.

It’s a powerful way to set up pipelines that normally would have to be coded.

How I did it: Step-by-Step

Before I go into the steps, I do want to point out that using AI is NOT enough. Delegation was a key part of learning how to reduce anxiety as a founder. I’ve shared my approach to delegating effectively in my post How to Delegate as a Founder. I invite you to read it.

Here's exactly how I set things my pipeline up, and how you can, too:

1. Create a Make.com Scenario:

A "Scenario", in Make.com lingo, is a series of steps that take place one after the other. My scenario starts when a webhook is triggered (webhook = a link that causes something to happen when someone navigates to it).

The final scenario looks like this, for reference. I explain it in detail in the next steps.

2. Set up an Intercom Trigger:

The majority of customer interactions at Glitter AI come through Intercom (a sort of chat widget customers can send me messages through). It’s where I get bug reports, feature requests, and a lot of feedback.

Intercom has a cool feature called Custom Actions that can trigger a webhook.

When I'm ready to categorize a conversation after interacting with a customer, I hit CMD + Shift + J in Intercom and select the Custom Action I created. It triggers the webhook in Make.com, which then starts the scenario I set up in step #1.

3. Categorize with OpenAI:

In Make.com, I take the conversation ID sent by Intercom in the webhook, and use it grab the full content of the conversation using the Intercom API.

Then, I pass the full text of the conversation into OpenAI API (ChatGPT) to parse and categorize the content.

I set up a specific prompt that tells ChatGPT to act like an expert product manager.

What the prompt does: summarize each conversation into clear, actionable points. Whether it’s a bug report, a feature request, or just general feedback. Here's the prompt I use:

You are an expert product manager extracting information from a conversation between admin and user.

Given the following conversation, write a summary of every bug OR feature request OR use case OR feedback provided by the user.

Some conversations may have multiple things (for example, bugs AND feature requests), but this may not always be the case.

Use your best judgment as an expert product manager to create the correct summary for this SaaS company.

Examples of use case:
"I'm using this to teach my new employees how to use our CRM"
"I want to show my VAs how to use a new system"
"I struggle with Zoom calls and this saved me the stress of jumping on one"

Examples of bugs:
"I can't log in"
"I keep getting this error"
"I get this error message when I install the desktop app"

Examples of feedback:
"I love how sleek the UX is"

Examples of feature suggestions (sometimes masked as questions):
"Does this support converting an existing video to a guide?"
"Does this integrate with Notion?"
"I wish I could remove the button from a screenshot"

4. Send to Notion

After ChatGPT does its thing, Make.com then sends the categorized information into Notion. In Notion, I’ve set up a table ("database" in Notion lingo) called “Second Brain.” Make.com creates a new entry in my Notion "Second Brain" - whether it's a feature request, bug, user feedback, etc. Here's what it looks like:

This both gives me a birds-eye view of everything customers are asking for, and sets me up to find patterns in all the requests. More on this in the next point.