Every founder I know has the same relationship with their CRM: guilt.
You know you should be logging calls, updating fields, writing follow-ups. But you're running a company. It doesn't happen.
Lightfield is built for exactly this problem. It's an AI-native CRM that updates itself after every meeting — no manual data entry, ever. It reads your emails, records your calls, and builds a complete picture of every customer relationship automatically.
Some of the things I wish I had when we were in the trenches doing early-stage sales:
Automated meeting prep pulled from your full account history
Post-call follow-up tasks drafted for you after every conversation
An AI agent that can answer "which customers asked for this feature and why?"
ChatGPT-like search across all your deals, calls, and emails
Over 1,900 startups are already using it. Setup takes minutes — connect your email and calendar, and it starts building itself.
Use code BIGDESKENERGY for 3 months off 🫡.

Like many of you, when ChatGPT launched in late 2022 it fundamentally changed how I work.
Since then I’ve adopted plenty of AI tools for my own day-to-day: Claude, Perplexity, Gemini, Granola, etc. And as a team we leverage some pretty powerful AI Slack integrations (i.e. Linear and Unblocked), coding tools like Cursor, and tons more for various different purposes.
We’ve successfully been able to find leverage in doing everything from note-taking to research to creating mockups to shipping code to extracting important information from legal documents.
But despite all of that, I’d still consider myself a pretty novice user of AI. I'd use AI when it was convenient, and it absolutely saved me a ton of time with specific workflows – but I always knew that I was just barely scratching the surface.
Then something changed.
In December, Claude Code had its big moment… likely due to the release of Opus 4.5 in November and people having some free time over the holidays. I couldn’t escape my feed where seemingly everyone was raving about building their own personalized apps.
In mid-January, Anthropic released Claude Cowork, which allowed an agent to take over your desktop and start performing everyday tasks – organizing your files, updating spreadsheets, and responding to emails. This felt like a crossing of the Rubicon. AI went from what felt like being primarily for engineers to being able to tackle everyday tasks for anyone.
Then in late January, OpenClaw (formerly known as Moltbot, formerly known as Clawdbot) went mega-viral on X. Every single post was about someone buying a Mac Mini or satire about someone buying a Mac Mini.
OpenClaw is an open-source autonomous agent that runs locally on your machine and executes tasks async in its own environment. You can connect it to WhatsApp, Telegram, iMessage – basically text your personal AI agent and it goes off and handles things on its own.
I spent hours watching tutorials and reaction videos where people would showcase how they were leveraging all of these new tools. It was fascinating to witness what felt like a step-function change in what’s possible with AI today.
Here's the problem though: by far my least favorite part of being a founder is the lack of time I have to explore anything outside of my core work.
During the week I wake up around 5:30am, hit the gym, and am at my desk by around 8am. That's pretty much where I remain (meals and all) until around 9pm, when I finally retreat to bed to read for a bit.
My weekends consist of wrapping up work I couldn't complete during the week, writing this very newsletter, and preparing for the week ahead. It's also quite frankly the only opportunity I have to get out, see people, touch grass.
So despite my fascination with all of these developments, my love of technology, and my desire to 100x my output… I’ve mostly been on the sidelines. I’d love nothing more than to irrationally purchase a Mac Mini and spend 12 hours building a personal agent. But it all felt way too experimental at a time where beehiiv demands so much from me.
It’s hard to justify tinkering when the day-to-day demands of the company are so overwhelming and tangible.
As I documented in my last newsletter, last Saturday I had a 12-hour flight home from Europe where I consumed some small blue pills, didn't blink for the duration of the flight, and annihilated my to-do list.
That bought me the first bit of freedom I've had in a while on Sunday. After a hot yoga sesh, I finished my usual Sunday work then made the mistake of downloading the Claude desktop app a few hours before bed.
I slept for 3 hours.
It wasn’t without trying though. I just couldn’t stop thinking about all of the repetitive tasks that I could begin to automate, or the performance gains we could get across the company by building some of these new workflows. I legit just tossed and turned thinking about what to build.
The first thing I built was a weekly reporting tool.
Over the past several years, our data team has built hundreds of invaluable dashboards that allow us to deep dive into specific metrics and trends. The issue is that it now requires someone loading and looking through 100+ dashboards on a regular basis to extract those insights. Nobody has the time for that.
So I built a Claude skill that analyzes all of our most important dashboards, cross-references trends against our database, pulls in additional context from Slack conversations, and then surfaces everything that is remotely relevant or concerning.
The second thing I built was a little petty.
To start each week, everyone submits their top priority in a Google Form. It’s meant to be a forcing function to show up to work with intentionality, and it also provides a level of accountability (we review them at the end of the week). Participation usually hovers between 70-80%... which drives me absolutely nuts.
I've always wanted to message everyone who didn't submit theirs, but that was never feasible… until now. I simply asked Cowork to cross-reference our employee list with the submissions and message everyone who didn't submit. That agent ran while I was in the middle of another meeting.
Both examples are pretty simple. But one thing I know for sure: the capabilities and accessibility of AI today are 100x what they were just a year ago.
I genuinely want to cancel all of my meetings for the next month and rebuild all of our processes to be AI-first. You can probably sense my excitement – and again, I’m just a week into diving deeper here (translation: I’m still such a n00b).
In one of the most recent episodes of My First Million, Sam Parr does this thing where he regularly expresses how overwhelming the pace of change is with AI. Something I totally relate to, and I assume most others do too.
But Shaan's response really resonated with me:
I think if you put on yourself that you need to become a top 1%, 5%, or even 10% kind of AI builder/user, that's the wrong frame. All you need to do is say — 'All right, I'm already top 10 percent in business, in content, right? You've already done all that in your life.' And now if you just get to like 50th percentile in AI, which is not that hard to do, it multiplies against all your other skills.
(By the way — I simply described the quote I was looking for to Claude and it dug that out of the YouTube transcript for me in a few seconds.)
I’m a founder of a startup that demands 80+ hours a week from me and I truly won’t ever have the luxury to be in the top 1% of AI users. But that’s alright – there’s an 80/20 rule at play here where I can spend an hour here or there pushing the limits of the latest tools and rebuilding my own workflows with AI.
But the real leverage won’t come from me alone. I’m not looking to become an indie hacker and add “$7K MRR” to my bio – I’m looking to build a multi-billion dollar business.
That means that I need my team of 120 employees to be just as excited about AI as I am.
We’re an unprofitable startup that is burning ~$700K each month. The only way for us to escape the venture flywheel of trading control for capital is to radically transform how we’re growing and operating.
We need to exponentially scale growth and revenue.
We need to increase output while reducing costs (or prevent future costs).
The good news: the exponentially scaling revenue part is already happening. We’re now adding well over $1M ARR per month in 2026 and should make a run at adding $2M ARR in March.

As for increasing output, we need to level up our team to leverage the latest models and tools to entirely rethink how we do work:
If engineers used to complete 5 tickets per day, with AI they should be able to complete 15.
If customer support answered 50 tickets per day, with AI they should be able to answer 100.
If customer success could manage 40 customers at a time, with AI they should be able to handle 80.
This isn’t optional – this is the path forward.
You wouldn’t hire a growth marketer who refused to use a laptop for work, the same way I don’t think you can expect most roles today to perform tasks without leveraging AI. It’s a technology that should empower those who harness it to considerably expand their contributions to their work.
The best companies, more than ever, are doing more with less. And the underlying technology is only going to compound.
A few weeks ago we had our first ever “AI Show and Tell”. The idea was simple: have employees showcase what they’re building with AI, which would hopefully inspire others. I intentionally scheduled it on Friday afternoons, so anyone who struck inspiration had the full weekend ahead to tinker.
But I came to the conclusion that “hopefully inspiring others” wasn’t an urgent enough timeline for the current situation. So we just launched an “AI Club” internally and mandated that at least one person from each team participate.

Early action from the #AI-Club
Each week, participants will have to showcase at least one new workflow or task that they accomplished leveraging AI. In the next few months, the difference in output between those participants leveraging AI and those on their team who aren’t should become fairly apparent – causing a cascading effect of adoption across the company.
The future is here, and it’s time to build it (with the help of AI).
If you enjoyed this post or know someone who may find it useful, please share it with them and encourage them to subscribe: mail.bigdeskenergy.com/p/slept-3-hours


Credit: Michael Finn
Shoutout Michael for the reader submission 🫡.
It’s only fitting to showcase an old western office with a typewriter after writing 1,600 words about AI. Life is all about balance.
Think you can generate a better office? Reply with your submissions 📨.

Some of my favorite content I found on the internet this week…
Ben Thompson joins the Cheeky Pint podcast (hosted by John Collison, cofounder of Stripe) to talk about AI ads, the end of SaaS, and the future of media.
The head of Claude Code joins Lenny’s Podcast. Very relevant discussion after writing today’s newsletter.
I can’t remember if I already plugged my episode on My First Million… but if you like this newsletter, you’ll love the episode. Check it out.
Should you be afraid of the SaaSpocalypse? Solid episode of Hard Fork with Casey Newton.
Want to understand how newsletter operators are leveraging AI? Complete the free survey and be the first to get the results.

Chat with DenkBot — my AI clone. It’s trained on everything I’ve ever published and the entire beehiiv knowledge base 🧠.
Or you can book an hour session with me (real me) directly → here.

Turn on, tune in, drop out. Click on any of the tracks below to get in a groove — each selected from the full Big Desk Energy playlist.

Share this newsletter with your friends, or use it as a pickup line.
👉 Your current referral count: {{ rp_num_referrals }} 👈
Or share your personal link with others: {{rp_refer_url_no_params}}
📥 Want to advertise in Big Desk Energy? Learn More










