Loopmaxxing

The best consumer founders know the game. 35K+ of them read Consumer Startups every week.

Stay ahead. Get the playbook behind today's breakout startups.

Read time: 2 mins 11 seconds

I found myself loopmaxxing lately.

It has completely changed how I approach work. I spend 50% of my day creating new loops or improving the ones I have.

For those of you who aren't familiar, a loop is an AI agent that keeps working until a task is done.

There are four types of loops:

1/ Turn-based. You prompt, AI works, you check, you prompt again. Most people are here.

2/ Goal-based. You define what done looks like, and the AI iterates until it gets there, e.g. stop when the score hits 90.

3/ Time-based. The loop runs on a schedule. 

4/ Proactive. The loop watches for events and acts without you in the room.

P.S. For those who want to dive deeper, here is an awesome Anthropic workshop on loops.

Here are some GTM loops I have built for my startup, Pikes, inside Claude.

1/ SEO/GEO. This loop reads my Search Console data, finds what people in my space are searching for, and drafts two blog posts a day. It writes against a playbook file of SEO and GEO practices from the sources I trust. 

2/ Outbound. Every morning this loop scans social accounts and industry publications for funding news, product launches, and retail expansions, then finds brands worth reaching out to. It drafts the email and finds the contact info, so all I do is review and send. I have found this approach to have a higher conversion rate than a spray and pray volume play.

3/ Follow-up. I used to spend 1-2 hours a day drafting follow-up emails and updating my CRM. Now a loop watches my calendar and fires eight minutes after each call ends, pulls my Granola recording, and drafts the follow-up in my tone of voice. A second sub loop keeps the CRM current.

4/ Upsell. I never had time to look for upsell signals in my existing customer base. Now this loop assesses trial signups and paid accounts every afternoon and pings me on Slack with who I should reach out to.

Loop orchestration will become a full-time job, or at least a big part of the job, for most knowledge workers.

Over the past month of building loops, I realized 80% of my work can be replaced by some version of a loop, as long as:

The other 20% is people-centric such as getting on a call with a customer or managing my team.

The other advantage is that everything a loop does can be hyper-personalized and data-driven. Most of my decisions used to be instinct calls since I never had the bandwidth to look at every relevant datapoint. A loop does. It reads everything before it acts, which increases the quality of an action.

This new paradigm also brought unique challenges.

I now spend most of my time giving feedback to existing loops, or project managing permissions, tool access, and context windows. 

What is the right user experience for this new paradigm? How do we get to a point where the loop learns from my feedback on its own? And in that world, what does it actually mean to be the human in the loop?

I don’t have the answer to most of those questions but what a time to be a builder!!

—

What are some loops you've built that are helpful? Would love to hear from y'all!

Until next time,

Leo

Follow me on X and LinkedIn

📥️ Want to advertise in Consumer Startups? Learn more.