Do What Feels Good

What is the first software you reach for on Monday morning? Your email? Your calendar? A project management platform, like Asana or Trello? Or maybe a dashboard in Hubspot/Salesforce?

I've noticed in the past three months that for me, my Monday software first touch is increasingly Claude. And not only is it becoming the first tool I load up in the morning while eating yogurt, but I've even been shifting which screen its on, from my secondary monitor to my primary monitor. Sidecar to main ride. Claude is quite literally becoming the center of my workflow.

It's been a slow creep, but as I look back, it's a perfectly reasonable transition. I didn't plan on moving my entire workflow, but here I am, chatting with AI as my personal assistant/project manager/CFO/(whatever else I need) while finishing my breakfast.

This incremental change is very likely happening on many of your own Mondays. Plenty of organizations are trying to artificially manufacture this kind of Monday by forcing AI down their people's throats. I bet the AI labs would love if their tool was everyone's first touch, last touch, and all the touches in between.

"TOO MUCH TOUCHING" you might be yelling right now. "DON'T TRUST THE TOUCH. YOUR BRAIN WILL ROT." Maybe. Maybe it's just fear of change. You don't have to submit to non-consensual workflow touching, but just because it's a different kind of Monday morning doesn't mean it's a sin.

If this is sinful, then send me to hell.

I like this kind of touch, and for this TL;DR, I'm going to show how I found myself in Claude's embrace every Monday morning.

Phase 1: First Kiss

I remember when I was scared to connect my email and calendar to Claude. It was just the first few weeks of my Max subscription, trying out prompts in both Claude and Gemini, and starting to slowly connect my business infrastructure to Claude. All I could think to myself when authenticating Google with Claude was, "I hope I don't regret this."

I approached this new connection cautiously, in the spirit of experimentation, with the goal of seeing how it would work. I asked simple questions to test if it could give me accurate answers. I asked it to answer trivia about my worklife. I asked it for reflections about me, but never to write an email to someone else. I had constructed a sandbox where if the AI was wrong, I would instantly know and it wouldn't affect a single customer.

Claude was right 95% of the time. When it got something wrong, I was able to see that my own prompt or assumptions were what made it to fail.

After a few experimental dates prompts, I began to trust the machine.

Phase 2: The Boudoir

In the start of June, I created "The Learn to Scale Strategy Room" Project in Claude. I wanted a place where I could align myself to my big goals, as well as make sure I was prioritizing the right tasks. I equipped this Project with my annual and quarterly goals, identified my most-likely failure modes, and instructed it to scan my email, calendar, Hubspot, and Asana whenever I started a fresh chat in the Project.

This second phase integrated Claude into my weekly habits. Over the past 7 years, I already had established weekly rituals. I've trained myself to plan my week on Monday morning and do a weekly retro on Friday afternoon, so I was pretty comfortable trying out a digital workspace in place of my paper weekly planner.

The gap my pen-and-paper planning always had, though, was that I had to continually remember what Past Dan had decided at the start of the year/quarter. I had to have discipline to double check my Big Why when deciding which little tasks to prioritize. I would sometimes ignore my strategic goals when something super fun and interesting was top of mind, and then weeks later, I'd bemoan myself, "I should have known better than to spend 20 hours repurposing an AI Bootcamp for a birthday giveaway."

I found that by introducing AI into my weekly prioritization habit, I was able to marry the bits I liked about my pen-and-paper planning and obtain the strategic alignment I needed. I would jot down everything that I thought I needed to do for the week on a sheet of paper, take a photo of it, then share it with Claude in my Strategy Room to run it down. Often it caught a few things I missed, offered ways to start tasks I hadn't realized that Claude could do for me, told me to ignore the shiny objects and to keep my eyes on the prize, and overall made me feel even more confident about how I was prioritizing my week.

If I didn't have weekly habits, this workspace would have been used twice and forgotten. Habits are harder to build than learning a new tool. Having the tool augment and improve my weekly habit with almost no additional work was a huge benefit.

I'd say Claude and I were going steady by now.

Phase 3: My No. 1 Side Piece

Four months. That's all it took to transition from a designated Strategy Room Project to being the first tool I opened on a Monday morning.

Hilariously, it's because I was so much more productive on Mondays that my habits needed to change. I found that by outsourcing some tasks to Claude (like drafting reconnect/follow up emails or updating my accounting), the improved efficiency I got on core projects (like creating Powerpoint decks and updating lesson plans), and listening to Claude's advice to delegate the right tasks to my team (thank you Ify and Kiki!!!), I was demolishing my weekly to-do list by mid-afternoon on Monday.

Read that again: With AI, I got most of my weekly work done by EOD Monday.

I didn't get suddenly smarter and the work didn't get easier. It was AI adding efficiency gains on every task, AI helping me delegate low-level tasks, and AI providing strategic feedback and support on high-value tasks.

By Tuesday morning, I was so productive that I needed to redo my weekly goals. Since my planning habit was "brainstorm tasks, review with Claude to prioritize, then go," I needed to do that a couple times a week. Then every day. Then first thing every day.

With all that improved efficiency, I was able to take on more client work. By completing more work in the same amount of time, I hit my revenue targets with fewer work hours and I actually felt like I could have leisure time. I started getting massages once a month and now I have Althea to look forward to.

Not only was I able to do more, I was able to do better work. I was privately embarassed when Claude found zounds of typos and inconsistencies in my past work, but now I felt so much more confident when delivering new AI-reviewed work. I was even able to add new value to existing engagements by creating interactive data dashboards. transforming traditional data exports into interactive analyses. Clients LOVE these. It took one prompt.

Claude, you might as well move in since your toothbrush and clothes are already here.

Interlude: Not My Kind of Kink

Claude tried, but I found a few of its suggestions left something to be desired. The big ones were efficiency hacks that didn't click with me:

• Automated Briefings: Having Claude summarize my week as a briefing was just a crappier version of what I already knew was going on. I tried dressing it up by pushing the briefing through the Tiddlywinks persona and generating an ElevenLabs voice narration so I could listen on CRO Roman Noodle's morning walk. It was meh.
• Inbox Triage: For a short while, I tried having Claude organize my inbox with labels and folders. I hated it: my email management habits were far superior than a perfectly logically organized inbox. I spent more time trying to find things and making sure the labeling nomenclature was internally consistent than just writing the damn email and moving on in my life.
• Company Brain: A few times I've tried to equip Claude with a brand book, a single source of context, or a consolidated set of SOPs. However, one of my failure modes is to overbuild and when Claude pointed out my MULTIPLE failed attempts at this knowledge consolidation (and then the upkeep, my real pitfall), I realized that I was just rearranging internal furniture to avoid doing sales outreach. You caught me, Claude.

I won't yuck someone's yum, but I will concede that by trying what it suggested, I was able to figure out what I didn't like and, more importantly, why I liked other things. I didn't like when Claude organized my life, but I liked when Claude built off my own work. I didn't like setting up universal standards for myself and having Claude enforce them, but I did like how Claude pointed out my uncharacteristic or self-defeating behaviors.

You won't know if you don't like getting spanked by an AI until you try it.

We All Have an AI Kink To Discover

If I was helping a new client adopt AI, I wouldn't recommend starting with Monday morning task triage. That's like sending someone to a munch without teaching them the basics of BDSM consent.

Instead, curating spaces for people to discover their own relationship with AI is a far less painful and far more effective way to drive adoption. When you're not in an "AI WILL ERODE MY CRITICAL THINKING" headspace, you can admit that having AI doublecheck your work and then find three real things you got dead wrong isn't the worst thing ever. Try it a few times and you might even like it.

Even still, when you're at the organizational policy level, there's a difference between giving permission and enthusiastic permission. Simply writing a policy won't get people to change. Even explicitly identifying approved AI-assisted workflows or offering a fundamental AI 101 won't get people to experiment. Removing roadblocks doesn't drive change: I believe that ineffective enablement programs fall down because they lack an essential enthusiasm energy.

I believe that long-term change happens by gently guiding people through a workflow that's relevant to their day job, helping AI newbies connect that AI workflow to their existing habits and workflows, and then giving room to play. In instructional design parlance, "I do, We do, You do."

Alongside this training approach, consent needs to happen at each phase. An organization that wants to see successful change needs to right-size the AI experience, be supportive and encouraging, offer options, and celebrating progress without shaming others.

This makes so much sense to me. Just like a human relationship, Phase 1 was negotiated limits, creating a sandbox where mistakes couldn’t reach a customer. Phases 2 and 3 was built off of that exploration, continually iterating or deciding- with experience- what wasn't for me.

Monday morning prioritization was a major arena for me to gradually expose to AI, but for you it might be tracking flight prices, creating a text-based video game, or creating Spotify playlists. I encourage you to try, and that it's ok to have the safeword "I didn't like that."

This is my open secret behind why I continue to host Show Me Something Cool With AI. I want to be the person to invite you to try something weird, share if you liked it or not, and to normalize the process. I get a thrill when someone steps a toe out of their comfort zone and realizes, "Yeah, I'm into that. What else could I try?"

Please, step into my lair.

I've collected AI prompts to try in an ever-growing list of Cool Things that people have demoed in past SMSCWAI meetups. It's designed so you can simply click a prompt and paste it into your AI tool of choice to get started right away. Everything there has been done by someone else and remember: that someone else had a first time, too. Perhaps they were afraid that it might be naughty, but if they had the courage to share at SMSCWAI, then at least one real human thought it was cool.

Go pour yourself a glass of wine, put on some soft clothes, turn on whatever music gets you in the mood, and try prompting your AI with something new...and maybe a little kinky.

A TL;DR from Roman Noodles

You should at least try la Pasta dei cornuti, no matter how long it took to cook. Perhaps you'd be into sharing...?

-Roman Noodles

This week I'm on holiday in Italy! I was putting the finishing touches on this week's TL;DR while on a rooftop bar in Rome, watching the sunset:

I'm going to go back to my vacation, but you'll notice that the typical update from CRO Roman Noodles is a little different. Get it? Roman Noodles? I'm hilarious.

Allora,

Dan from Learn to Scale​

​Opt-out from the newsletter | Unsubscribe from all emails | Update your Preferences | www.learntoscale.us, Boston, MA 02119 |

​PS. PEAS​