Now that we’re into the Sep–Dec great lock-in season of the year, I thought it would be fun to create a super productive morning routine… with AI!

I’ll be honest: my morning routine over the last few years has been genuinely terrible. 😱 It usually looks something like: snooze the alarm 3 times, grab my phone and doom-scroll for an hour, then wake up later than I wanted and dive straight into work.

So this year, I’m committing to actually owning my mornings - not by joining the mythical 5am club (I’ve accepted that’s just not me), but by building a routine that sets me up to win the day regardless of what time I wake up.

And of course, I thought: why not use AI to make this even more effective?

So I’ve researched scientifically-proven habits that actually work, then figured out exactly where Claude/ChatGPT can slot in to make each one easier, more personalised, and genuinely stick.

Habit #1: We’re not touching our phone for the first hour of the day

When you reach for your phone the second you wake up, you flood your brain with everyone else’s demands, other people’s highlight reels, and low-level anxiety before you’ve even had a chance to think your own thoughts (I do this to myself daily!!).

The research is clear: those first waking minutes set your mental state for the entire day.

The Claude/ChatGPT twist: Instead of doom-scrolling, replace that input with something that actually serves you and sets you up for a successful day. This is where my Morning Briefing Magazine comes in - I have Claude build me a personalised magazine every morning that pulls the news I actually care about, my calendar, and my to-do list into one beautiful personal “publication”. So I still get “caught up,” but on my terms - not the algorithm’s. (If you missed that issue, go read it - it’s one of my most popular ever.)

Habit #2: We’re getting sunlight (or any daylight) in our eyes

One of the simplest, most powerful things you can do is to get natural light into your eyes early kickstarts your circadian rhythm, regulates your cortisol, boosts your mood, and helps you actually sleep better that night.

No sun where you are? (I relate as I’m in the UK) A bright light or sunlamp does the trick too.

The Claude/ChatGPT twist: Take your morning light outside and pair it with a walking brain-dump. Open Claude or ChatGPT on voice mode and just talk - your worries, your ideas, whatever’s on your mind. Then ask: “Sort through everything I just said and give me the one thing I should focus on today.” You get sunlight + movement + mental clarity, all in one go!

Habit #3: We’re going to set our “Daily Non-Negotiable”

After your clarity walk/brain-dump, you’re going to take 5 minutes to decide: “What’s the one most important thing (up to 3 things) I could do today?”

Then pencil that into your calendar as a non-negotiable that’s going to get done without fail.

The Claude twist: This links in perfectly with a Weekly Debrief with Claude Cowork issue a wrote a couple of weeks back, where you track your goals week-on-week and use Claude to keep you on track. This will help you make sure your daily actions are aligning with your bigger goals.

Ok, so, your mission, should you choose to accept it:

1. Pick just ONE habit from this list to try this week

2. Add ONE Claude twist to make it easier or more valuable

3. Do it for a week before stacking the next one

That’s it! Small, sustainable, science-backed - get ready for a great 2026 lock-in season! 💜

This playbook is a bit different to my usual ones, because it’s not about making money. It’s about what I do with it after I earn it.

I’ve been investing in ETFs (exchange-traded funds) since 2022, and my portfolio (over $20,000 now) is currently showing a personal rate of return of over 68%.

When I checked that number recently, I was mega thrilled, because it’s genuinely amazing for something I spend maybe an hour a month on (and something I’ve managed to figure out by myself from scratch).

So let’s dive into how I’ve arrived here and how you can get into investing and growing your wealth using AI.

Super important disclaimer: I am not a financial adviser, and nothing in this playbook is financial advice. This is me sharing my personal experience and process for educational purposes only. Investing puts your capital at risk: the value of investments can go down as well as up, and you can get back less than you put in. Past performance (including mine) is never a guarantee of future returns. Please do your own research, and never invest money you can’t afford to lose.

Read more

I’ve sent a newsletter every single week since 12 April 2025. And as a recovering serial quitter, I’m so proud that I haven’t missed a single week in 16+ months!

That commitment has helped me create something genuinely incredible.

In that time I’ve grown the newsletter to over 30,000 subscribers, and it now earns money in five different ways: ads, paid subscriptions, digital product sales, brand partnerships and affiliate links. My best ad month alone was $1,048.64, from an email I was going to write anyway.

But the actual reason I’d tell anyone to start a newsletter is much less glamorous than the money: your social media can be taken away from you at any time. Your email list can’t.

Algorithms change, accounts get banned, reach dies overnight. Your list is the only audience you truly own, and it’s the single most valuable asset in your whole business.

This playbook is the exact setup I would follow if I was starting from scratch: how to pick a niche worth monetising, the platform I’d use, the cadence rule, my Human - AI - Human writing framework, how I grew to 30,000, and the maths of getting to $1,000+ a month.

Read more

The online discourse (and general widespread panic/anger/confusion) this week has been mainly revolving around the new EU AI transparency laws.

If you’ve seen the headlines, you’ve probably come across some version of “you now have to slap ‘Made with AI’ on everything you create with AI!”.

Regardless if you’re happy or annoyed about that, it’s an incorrect blanket statement.

So today, I want to break down what these laws actually mean in plain English.

Quick disclaimer: I’m sharing my understanding to make this simple, not giving legal advice. If your specific situation is unclear, please talk to someone qualified. 💜

The New EU AI Laws Explained (Without The Panic)

On the 2nd of August, the EU’s new AI transparency rules officially came into effect. And the internet immediately did what the internet does best: oversimplified it into hysteria. 😱

The viral take has been: “Label everything AI has ever touched!!!!”

But the reality is much more specific and sensible.

Also, an important heads up: even if you’re not in the EU, this can still apply to you. If your content, products, or AI systems reach people inside the EU, you’re potentially in scope.

The One Distinction That Clears Up 90% Of The Confusion

Most of the panic online comes from people missing one key thing: there’s a difference between being a provider and a deployer.

Providers = the companies that build the AI systems (e.g. OpenAI, Anthropic, Google).

Deployers = the rest of us - the people who use those AI tools in our work.

If you’re using ChatGPT, Claude, or any AI tool inside your business, you’re almost certainly a deployer. And that matters, because a lot of the technical, heavy-lifting responsibilities fall on the providers, not you.

The providers have to build safety and detection features into the tools themselves. You, as a deployer, just have to be transparent about certain specific uses.

The 4 Things The Rules Actually Cover

There are really only four areas to pay attention to:

1. People need to know when they’re talking to AI

If your customers are directly interacting with an AI - like a chatbot on your website - you need to make that clear. Not hidden in the terms and conditions. Just an obvious little heads up, ideally right at the start.

Something like: “Hey, this is a bot!” in your welcome message does the job.

2. AI tools have to mark their own outputs

Companies like Google and Anthropic have to build in invisible, machine-readable markers so a computer can detect their AI made something.

So yes, anything you create with these tool will have invisible watermarks that make them detectable from now on.

A lot of people have tried to explain what the watermarks are like this past week (it’s basically a complex system behind the choice of words).

For me personally, this makes no difference. I have and will continue to follow my Human-AI-Human framework. I input ideas, AI polishes them, I always have the last edit and rewrite most of the content in my own style.

3. Deepfakes need a visible label

This is MY FAVOURITE rule and I think most people will be celebrating with me:

If you publish AI-generated or heavily manipulated image, audio, or video that could realistically pass as real, you need to visibly disclose it. (check out my thumbnail on this newsletter).

The key question isn’t “did AI make this?” It’s “could this fool someone into thinking it’s real?”

For example: No one thinks a cartoon avatar of yourself is real - so no label needed. A realistic clone of your face or voice, a fake testimonial, or a synthetic “person” pretending to use your product - That absolutely needs a label. (and rightly so!!)

This was a grim corner of the industry that fuelled AI distrust, so I’m all for this.

4. Public-interest text needs disclosure

This rule only covers AI-generated text published to inform the public on serious matters - think politics, health, safety, legal rights, finance. Your AI-assisted Instagram caption or product description is not included.

And there’s a loophole: even for serious public-interest content, if a human properly reviews and edits it and takes responsibility for publishing it, you don’t need the disclosure.

A Simple Checklist Before You Publish

1. Am I the provider or the deployer here? (You’re almost certainly the deployer.)

2. Is someone interacting with AI without realising it? (If yes, tell them.)

3. Could this image, video, or audio pass as real? (If yes, label it.)

4. Is this text informing the public on a serious matter? (Most content isn’t.)

5. Did a human properly review it, and who’s taking responsibility? (Own your work.)

You do NOT need to go back and re-label years of old content (anything before August 2nd is fine), and enforcement is meant to be proportionate - one missing label isn’t an instant fine.

Why I Actually Think This Is A Good Thing

It rebuilds trust. Both as consumers AND as creators, we’ve been drowning in a sea of “is this real or is this AI?” More transparency means we can trust what we’re seeing again. Yay!

It’s the slowdown we all needed. The pace of AI has been dizzying. This feels like a collective deep breath - a chance to step back, reconnect with our own creativity, and remember that the human touch remains essential. Again, yay!

And here’s my personal lens: as a UGC creator, I’d already been watching brands shift towards AI-generated content and AI avatars. But with these transparency laws in place, I genuinely believe brands will keep choosing to work with real people for this kind of content. Triple yay!

So across the board, I see this as a win. More trust, more creativity, and more value placed on being genuinely, unmistakably human. Cheers to that 🍾🥂

Starting UGC has been one of the best things I’ve done this year.

I’ve made multiple 5 figures with it ($40K+) in just a few months, with July being my biggest month yet with $5040.

After testing about 20 side hustles, I really believe this is the easiest and quickest way to make some form of online income.

But I also notice so many people saying “UGC is saturated” or “UGC doesn’t work anymore,” what they almost always mean is that ONE type of UGC didn’t work for them (or they haven’t even tried it at all).

Because there are actually 6 different types of UGC. They pay differently, they suit different working styles, and most people only ever try one.

I’m now actively doing 4 of the 5, and it’s the stacking of them that’s made UGC one of my most reliable income streams. So in this playbook I’m going to walk you through all 5, tell you which ones I do, exactly where I find the work for each, what they pay from my experience, and how I’d stack them if I were starting today.

And no, you don’t need a following for any of these. Some of them need zero audience, zero personal brand and even zero portfolio to get going.

Read more

AI is the best tool we’ve ever had for doing the stuff we always meant to do: Stay informed, stay organised, actually keep track of our own lives…

So today we’re using Claude (or ChatGPT) to do a weekly debrief that helps you close your week properly, spot what’s actually working, and start Monday with a plan instead of a vague sense of dread. 😰

We’re building out your weekly debrief process…

Here’s what we’re building in this issue - a weekly review that:

• Logs what actually happened in your week (pulled from your calendar and inbox, not your memory)

• Gives you an honest read on what moved forward and what you avoided

• Asks you 4 reflection questions that take 10 minutes to answer

• Sets your top 3 priorities for next week WITH you (not at you)

• Remembers everything, week after week, and holds you to it

The setup takes 10 minutes. Then every Friday at 4pm, your debrief is waiting for you with the first question ready for you to answer.

Your past attempts at doing this haven’t worked because:

1. Logging your week is homework. Sitting down to remember what you actually did for 7 days feels tedious so you skip it & eventually quit.

2. Your journal has no memory. You can write “next week I’ll finally fix my pricing page” every single Sunday and the journal will never once say “babe, this is the fourth time!!!”

This setup kills both problems. Claude does the logging for you, and Claude remembers.

How it works: 2 phases

Phase 1 runs automatically at 4pm on Friday. We’ll set up a Claude Cowork Scheduled Task that will:

• Read last week’s debrief and note what you said you’d do

• Pull every meeting and event from the past 7 days of your calendar

• Check your inbox for an overview

• Check your goals doc for what was planned vs what got done

• Preview next week’s calendar for anything big coming

Then it writes your “week in receipts” with an honest read of the evidence. What progressed, what stalled, what you appear to have been avoiding.

And then it stops, asks you the first question, and waits.

Phase 2 is you, in the chair, for 10 minutes. You open the task, the debrief is prepped, and Claude walks you through 4 questions, one at a time:

1. What’s the one thing from this week you’re genuinely proud of?

2. What did you avoid this week, and what was avoiding it protecting you from?

3. If next week went exactly like this one, would you be okay with that?

4. What’s one thing that, if it happened next week, would make everything else easier?

Then Claude proposes your top 3 for next week, you push back until you agree, and it saves the whole thing as a gorgeous magazine-style file in a debriefs folder in Notion, Google Drive or on your computer.

The prompt tells Claude to read last week’s debrief before starting this week’s. Which means from week 2 onwards, it’s working off of real info about you.

⚠️ Before you start: you need Claude Cowork

Same deal as the morning magazine - this runs as a scheduled task, and scheduled tasks only live inside Claude Cowork.

Important note: Cowork needs your computer on for tasks to run. Which is exactly why I chose Friday at 4pm - you’re at your desk anyway, finishing the week. By the time you’re ready to close the laptop, your debrief is ready too.

‼️ And the same warning as last time: connecting your calendar and Gmail means Claude is reading them to build this. Totally fine for most people, but you should know you’re switching it on, not find out later.

OK, let’s build this thing…

Step 1: open Cowork and create the task

Go to “Scheduled” in the left-hand sidebar and click “New Task”.

Step 2: personalise your prompt

1. Your lens. Who you are, what you’re building, what a good week actually looks like for you.

2. Where your goals live. A Notion page, a Google Doc, a folder on your computer. Don’t keep goals anywhere? Just write “nowhere, just ask me” and Claude will ask you at the start instead.

Step 3: copy/paste this prompt

Build me a weekly debrief that runs every Friday at 4pm.

This works in 2 phases. Phase 1 you do automatically. Phase 2 is a conversation with me. Do NOT move to Phase 2 until I reply.

MY LENS: I'm [WHO YOU ARE: your work, your life, what you're building or working towards]. Judge my week against this, not against some generic productivity standard.

MY GOALS LIVE HERE: [LINK OR NAME YOUR NOTION PAGE / GOOGLE DOC WITH YOUR TASKS OR GOALS - or write "nowhere, just ask me"]

PHASE 1 - PREP (do this automatically, before I arrive):

1. Read my most recent debrief in the debriefs/ folder. Note what I said I'd do this week. If the folder is empty, this is week 1, skip this step.
2. Pull the past 7 days from my Google Calendar: every meeting and event.
3. Check my Gmail from the past 7 days: what I sent, what important emails are still sitting unanswered, and any threads that went quiet.
4. Check my goals doc (linked above) for what was planned vs what got done.
5. Pull NEXT week's calendar and flag anything big coming.

Then write me a short "week in receipts" summary: just the facts of my week, no opinions yet. Follow it with your honest read of the evidence: what moved forward, what stalled, and what I appear to have avoided. If I made commitments last week, say plainly whether the evidence shows I kept them.

End Phase 1 by asking me Question 1 below. Then STOP and wait for my reply.

PHASE 2 - THE DEBRIEF (only after I reply):

Ask me these questions ONE AT A TIME, waiting for my answer each time:

1. What's the one thing from this week you're genuinely proud of?
2. What did you avoid this week, and what was avoiding it protecting you from?
3. If next week went exactly like this one, would you be okay with that?
4. What's one thing that, if it happened next week, would make everything else easier?

Your tone: friendly but unflinching. You're a supportive reviewer, not a cheerleader. If my answers contradict the evidence, say so. If I'm avoiding the same thing for the third week running, name it.

After the questions, propose my top 3 priorities for next week based on everything above. Discuss them with me until we agree. Don't just generate them at me.

FINAL STEP - THE DEBRIEF FILE:

Write it all up as a single self-contained HTML file styled like an editorial magazine (Fraunces and Inter via Google Fonts, big display typography, distinct sections). Sections: My Week in Receipts / The Scoreboard / What I Said (my reflections, written up) / Next Week at a Glance / My Top 3.

Save to a debriefs/ folder, named YYYY-MM-DD.html. Next Friday, you'll read this file before we start again.

Schedule this to run every Friday at 4pm.

You can of course choose todo this on Saturday, Sunday or even Monday as a start to the week instead. And don’t worry if you miss a week - Nothing breaks. Claude just reads the most recent debrief, whenever it was, and picks up from there.

Your weekend will now start with the week intentionally and strategically finished. Not just... randomly stopped.

I made $18,228.97 in July, and that’s officially my best month ever! Again!

Last month I told you June’s $17,066.28 felt surreal. So you can imagine how I’m feeling typing an even bigger number 4 weeks later. 🤯

But July wasn’t a fluke, and it wasn’t luck. It was compounding.

I had return UGC clients coming back, systems I set up starting to pay off and multiple income streams slowly starting to mature.

And in full transparency, it was also a seriously hard-working (i.e. exhausting) month. I’ll dive into that a bit more at the end, because I’m expecting August to be a much lower month, on purpose.

First, let me show you exactly how the $18,228.97 came together.

Read more

Quick question: have you ever given Claude (or ChatGPT) a really clear instruction… and then had it completely ignore you?

Turns out, one tiny word can completely change the result you get!

So this week I want to share the single smallest prompting tweak I’ve found that makes the biggest difference.

Here’s a scenario yo might have experienced:

You tell Claude: “Don’t em dashes”

And it listens! No em dashes.

But then you read back what it gives you, and something’s still… off. It’s now packed with phrases like “it’s not just X, it’s Y.” Every paragraph opens with “In today’s fast-paced world.”

The em dashes are gone. But it still sounds exactly like AI wrote it.

Here’s what actually happened: you banned one specific thing, so Claude stopped doing that thing. But you never explained what you were really trying to fix - which was sounding robotic. So every other AI tell is still fair game.

AI can take your rules incredibly literally.

But there’s one word that makes this better:

Because.

Here’ another example:

⛔ Weak: “Don’t use emojis.”

✅ Strong: “Don’t use emojis because this is going into a formal proposal for a corporate client who expects a polished, professional tone.”

See that?

In the first version, you banned one specific thing - emojis. And Claude will very likely obey that.

In the second version, you explained the goal - a formal, professional document. Now Claude understands that emojis, casual language, exclamation marks, and anything else that undermines a polished tone are ALL off the table.

You didn’t have to list every possible mistake. You gave it the content + the reason, and it figured out the rest.

How To Write A “Because” That Actually Works

“Because it’s important” does absolutely nothing. Claude can’t work with it as there’s not much context to that.

Here are a few good examples instead:

For how it sounds:

“Write short sentences because most of my readers open these on their phone and skim while doing three other things.”

(Bonus: this will likely also get you short paragraphs and front-loaded points automatically.)

For formatting:

“Don’t use tables because I paste this into my email platform, and tables break when it sends.”

For protecting trust:

“Don’t promise specific results because this goes to clients in a regulated industry, and anything that reads like a guarantee gets flagged.”

(Words like “guaranteed” and “proven” will likely vanish from results - without you having to list them.)

For keeping facts straight:

“Tell me plainly when you’re not sure because I repeat these facts to clients, and a wrong answer costs me more than a missing one.”

(Don’t rely on this though - always fact check results from AI.)

Make Your “Becauses” Permanent

Custom Instructions (for reasons about YOU and your audience)

Click your initials → Settings → General → “Instructions for Claude.” Anything you write here gets read at the start of every single conversation.

Project Instructions (for reasons about a specific project)

Inside any Project, click “Set project instructions.” These only apply within that Project.

Set these up once, and every future first draft comes back already checked against these criteria.

Hope you enjoyed this hack -tiny tweak, massive difference!

Out of all 8 of my income streams, digital products is the one I get most excited about, and it’s the one I’d recommend to pretty much anyone.

In June, my little digital product Voice Match made $3,589.51. It’s made well over $5K since I launched it in May.

I built it once and I make sales from it almost every single day. No trading hours for money or chasing invoices. It’s truly amazing to see those little sales notifications coming up on your phone while you’re just living your life.

So if you’ve been circling the idea of digital products but felt overwhelmed by which one to make, how to actually build it, and (the bit nobody explains properly) where to sell and deliver it, this is the issue for you.

I’ve gone deep on the best digital product creators out there, pulled the genuinely useful tips, thrown out the fluff, and stitched it all into one plan I’d follow if I were starting from zero today. Then I’ve added the exact tool I’d use to host, sell and deliver it, which is HighLevel.

Let’s get into it.

Read more

POV: you KNOW you asked Claude the perfect thing about 3 weeks ago… and had that answer that made you go “oh my god, yes!!”.

But now you’re scrolling through 200 chats called “Untitled”, “Draft”, or “Marketing ideas”, and it’s just... gone. Swallowed by the sidebar.

Your Claude sidebar isn’t a chat log. It’s a second brain. Or it could be!!

But for most people right now, it’s a junk drawer where their best thinking is rotting away.

Every chat you can’t find is a prompt you have to rewrite, a decision you have to make twice, or an idea you had once and then lost forever.

So let take 5 minutes together to turn your Claude sidebar into a second brain.

Let’s organise your Claude sidebar…

Layer 1: Pin your VIPs

Let’s just start with the quickest, highest-impact move.

You can pin any chat or any whole project to your sidebar so it’s always quick to reach, no scrolling required.

How to do it:

1. Click the 3 dots on the right-hand of any chat or project in the sidebar, then click pin:

2. Or from inside a chat - click the down arrow next to the chat title - then click pin:

Now pick your 5 to 10 absolute go-tos: the client you work with every single day, the chat you keep reusing, your current main project. Pin them and they’ll sit together in your Pinned section.

This one move fixes about 80% of your daily “where did it GO!“ frustration, in roughly 30 seconds.

I would suggest looking through and updating/removing pinned chats at the end of every week.

Layer 2: Projects are your folders

Projects act pretty much like folders inside Claude.

Think of a Project as a “room”. It holds its own chats, its own instructions, and its own files, all cut off from the chaos of your main sidebar.

Set up one “room” for each big, ongoing thing in your life:

• One per client (🤝 Acme, 🤝 BigCo)

• One for Content (📱Instagram, 📱TikTok)

• One for Money + admin (💰 Budgeting)

• One for whatever you’re currently learning (🧠 AI, 🧠 Copywriting)

You can also give each type of project its own emoji, so things jump out at a glance in your sidebar. For example:

• 💰 for anything money-related

• 🤝 for anything client-related

• 📱 for content and social

The trick is consistency: one emoji per category, used every time, on your projects AND on any loose chat that lives outside a project.

Your eye starts hunting the symbol instead of reading the words, so finding things gets so much faster.

(Skip it inside a project though. Everything in your 🤝 Acme room is already an Acme chat, so an emoji in there is just noise.)

Layer 3: Decide on a naming convention

Not everything deserves its own project. For the loose, one-off chats, the one thing you fully control is what the chat is called.

Claude auto-names your chats something generic and instantly forgettable. But you can rename any chat in 1 click: click the title at the top of the chat, hit Rename, done.

Use this format:

[emoji] Category - specific thing - Mon ‘YY

So “Untitled” becomes 💰 Tax - Q3 expenses breakdown - Nov ‘26.

Two rules make it actually work:

1. Start with the word you’d type into search. The important bit goes FIRST, because the sidebar is a narrow little column and it chops the end off anything long.

2. Use emojis as we covered above.

Layer 4: Use Search + Memory (paid plan)

On the paid plans, Claude can now search and reference the actual content of your past chats, not just the titles. Switch on “Search and reference past chats” in your settings and it becomes your personal librarian.

I personally still don’t love the search feature in Claude, so naming still matters for the fast eyeball-scan down your sidebar. Starring keeps your VIPs on top. Search + Memory do the deep digging. Different jobs, all covered.

Layer 5: The 5-minute weekly reset

A system you don’t maintain quickly goes back to the mess you started with.

So once a week, do a tiny reset:

• Rename the 3 to 5 chats actually worth keeping

• Pin anything new you’ll want fast

• Bulk-delete the dead chats you don’t need anymore

Want Claude to do all this for you?

Paste this into a NEW Claude chat (not inside a Project) and it’ll build the whole system around how you actually use Claude:

I want you to turn my messy Claude history into an organised second brain. Do it in 4 steps and keep it tight.

STEP 1 - AUDIT: Use your tool for searching and referencing my past chats to look at my most recent chats (aim for the last 50 to 100). Note what I actually use Claude for, and flag the 5 chats or projects I return to most so I can star them for instant access. If you can't access my past chats, stop and tell me to switch on "Search and reference past chats" in Settings first. Also flag that any chats living inside a Project won't show up here, so this only covers my loose chats.

STEP 2 - BUILD MY ROOMS: Based on what I actually do, suggest 4 to 7 Projects I should create (think one per client, plus areas like Content, Money, Personal, and whatever else keeps showing up). For each, give me a name, a one-line description of what belongs in it, and 2 to 3 lines of custom instructions I can paste straight into that Project.

STEP 3 - BUILD MY LABELS: Give me a naming format for the loose chats that don't live in a room, using this pattern: [emoji] Category - specific thing - Mon 'YY. Assign one emoji to each of my main categories, spelled and used identically every time. Front-load the word I'd actually search for, and keep every title under 45 characters so it doesn't get cut off in the sidebar.

STEP 4 - SET AND FORGET: Write me a short block to paste into Settings, under "What preferences should Claude consider in responses?", telling you to suggest an optimised title in my format at the end of your first reply in every new chat, using my exact categories and emojis, and to remind me to rename the chat myself (since you can't do it for me). Put that block in a code box so it copies cleanly.

Then, as a starter, rename my 10 most recent loose chats into the new format and give them to me as a simple 2-column list: current title, new title. No em dashes anywhere.

Last month I made $780 from tech UGC (also known as canvas UGC), from an app I’d only signed up to a few weeks before.

I created my SideShift account in mid-May, and June was my first proper month actually posting. That led to $780 for small tasks I was fitting around everything else I do, mostly posting simple little videos that take a few minutes each.

I’d been hearing about tech UGC everywhere, so I thought, I’m going to stop wondering about this and actually try it properly.

And I’m really glad I did. It’s turning into one of the easiest little side streams I’ve added all year.

Very important to note: this is not the same as the “normal” UGC I wrote about in my previous playbook. With that kind, you create content and the brand posts it on their own channels.

With tech UGC, you create your own social accounts and you post simple videos on them every single day for an app or SaaS tool, and you get paid for posting. You don’t need a following and you don’t need a personal brand to do this.

That’s exactly why I think this one is such a good side hustle for anyone who wants to add some form of income, especially if you’ve still got a 9-5. It’s low-lift, you can do it around a full day, and the videos are genuinely quick once you’ve got your system going.

Let me show you exactly how it works.

Read more

I share a lot that for 7 years I stayed stuck in exactly the same spot - setting goals, then overthinking and procrastinating.

But 15 months ago I just decided to stop trying to make perfect decisions and start making fast ones I could correct later instead (and I used AI to help me every step of the way).

That one shift did more than 7 years of “planning and researching” ever did for me. Since then I’ve grown an audience of 100,000+, grown my income significantly, and landed amazing opportunities.

Overthinking kills more dreams than failure ever will.

So today, we’re covering 3 ways to stop overthinking and make faster, better decisions - each backed by real research, and turned into a Claude prompt you can steal.

1. Get out of your own head

Think of the last time you were helping a friend who was struggling with a decision you could see the answer to clearly in about 4 seconds. For you it was obvious.

But answers to your own decisions are never obvious.

There’s a name for this: Solomon’s Paradox - we reason wisely about everyone else’s problems and completely lose the plot on our own.

This is something Claude is amazing for!

But don’t ask it to decide for you. Make it pull your own wisdom out:

I'm stuck on a decision. I'm going to describe it, then I want you to reframe the whole thing in the third person, as if it's happening to a person called [your name] who I'm advising.

Don't give me your opinion. Ask me what I'd tell them to do, then keep asking until I've talked myself into the actual answer.

Here's the situation: [brain dump everything]

2. Fast-forward to the disaster

Most overthinking isn’t really thinking. It’s actually just vague fear replaying on a loop.

In 1989, researchers found that imagining an event has already happened (rather than might happen) makes you around 30% better at correctly naming why it turned out that way.

So before you commit, you pretend the decision already failed spectacularly, then write the story of how.

It helps you spot the real risks you were too optimistic (or too panicked) to see. Plus the fear shrinks, because a named, specific worry is always smaller than the nameless dread you’d been marinating in.

It's 12 months from today. I went ahead with [decision] and it failed badly. Write the story of exactly how it went wrong, in detail - every reason, including the uncomfortable ones I'd rather not look at.

Then for each reason, give me the single thing I could do right now to stop it happening. Be blunt with me.

3. Set the bar before you look

Psychologists split people into 2 types: Maximisers hunt for the best possible option. Satisficers decide what “good enough” looks like, take the first thing that clears it, and get on with their lives.

In a 2006 study, graduates who maximised their job hunt landed salaries around 20% higher - and felt worse. More stressed, more regret, less satisfied with the job they actually got. They did better and felt worse. Because when your standard is “the best,” you’re never allowed to be happy. There’s always a maybe-better option haunting you.

Overthinking is just maximising with no off switch.

The fix is a 2-step you set up before you look at a single option:

1. Decide your “good enough” bar first (3 to 4 must-haves).

2. Ask: is this reversible? Most decisions are a door you can walk back through, and reversible decisions don’t deserve your worry. They just need you to act with speed.

Before I look at any options for [decision], interview me to lock in my "good enough" bar - the 3 to 4 things any option MUST have. Then ask me one question: is this decision reversible?

If it is, hold me to this rule: I pick the first option that hits all my must-haves, and I stop looking. If I start chasing a "slightly better" one, call me out on it.

And be ruthless about picking 3 goals to focus on. I know 3 feels stingy when you’ve just written a gorgeous long list. But 3 goals you actually hit in 12 weeks will do more for you than 9 you half-chase and feel guilty about all autumn.

In conclusion…

Pick a decision you’ve been circling for weeks and run it through just one of these today.

Then go and do the thing. Momentum is the whole point.

Hey future-proof friends 💜

Cast your mind back to January for a sec…

Maybe you were planning your best year ever… dreams, goals, exciting plans…

We’ve just hit the half way point of the year, so it’s a great time to check back in. Review where you’re at, celebrate the wins or maybe make a new plan to get back on track.

So that’s what we’re doing today. A powerful mid-year review using Claude that you can turn into a plan you’ll still be referring to in September.

Your life can genuinely look different in 6 months (it actually took me a lot less than 6 months to start seeing big results after I decided to get serious about working on my dreams).

You have 6 months left in 2026 - plenty of time to make this year amazing. So it’s the perfect time for a mid-year review that does 3 things:

1. Look back honestly.

2. Decide where you actually want to be by January.

3. Set up the goals and weekly actions that keep you moving there after the excitement wears off.

And Claude makes this incredible powerful and easy.

Step 1: Reflect

We’re starting with a proper look back at your last 6 months. Full honest here, not just the highlight reel or the wins. We’re really reflecting on both good and bad.

The trick is to not let yourself off the hook, which is exactly what Claude’s good at if you tell it to be. Paste this in:

You're my thinking partner for a mid-year review, 
and your job is to get the honest answers out of me, 
not the comfortable ones.

Interview me one question at a time. Cover: what actually 
went well in the last 6 months, what didn't, what I learned, 
what gives me energy vs what drains it, and what would make 
the next 6 months a genuine success.

Rules for you:
- One question at a time. Wait for my answer before the 
next one.
- When an answer is vague, a humble-brag, or a story I'm 
telling myself, call it out and ask again. Push for the 
specific thing underneath it.
- If something I say contradicts something I said earlier, 
point it out.
- Don't reassure me or tidy up my answers as we go.

When we're done, summarise it back as a short, honest 
reflection in my own words, and ask me what you got wrong 
before we lock it in.

By the end of this you should have real clarity on the life you’re actually building and the person you want to become.

Step 2: Turn it into goals

Clarity’s lovely, but it’s not a plan. So now we turn it into goals you can actually measure.

I group mine by life area (a habit I nicked from Chris Guillebeau, who’s run a public annual review every year for years), because “be healthier” isn’t a goal, it’s a vibe. “Move 4 times a week” is a goal.

You're my high-performance coach for this mid-year review. 
Be direct, not a cheerleader.

Here's my reflection: [paste your reflection from Step 1]
Here are the goals already on my mind: [paste anything you're 
already thinking]

Work through this with me. Don't dump it all at once:

1. Start with the goals I gave you. Before improving any of 
them, pressure-test each one: Is it actually mine, or one I 
think I "should" have? Is it measurable, or just a vibe? Is it 
realistic in 6 months alongside everything else? Tell me which 
ones are weak and why.
2. Then, with me, rewrite each goal worth keeping so it's 
measurable: a clear target, and where I am right now.
3. Now show me 3-5 goals I might be missing based on my 
reflection, as options. Tell me why each might matter. I'll 
pick which ones make the cut, don't assume.
4. Once I'm happy with the final list, sort it by life area.

Step 3: Pick 3 and build the systems

This is the step nearly everyone skips. And it’s the reason so many of us set goals every January and hit December with nothing to show for it.

Setting a goal is one thing. Building the daily and weekly actions that slowly drag you towards it is another thing entirely.

2 ideas make this work:

1. Plan in 12-week blocks, not 12-month ones: I pinched this from the 12 Week Year (Brian Moran’s book). A year is too far away to feel real, so most of us go hard in January, coast for 10 months, then panic in December. 12 weeks is close enough that this week matters every week.

2. Track the action, not the outcome: it’s tempting to stare at the follower count or the revenue number. But you can’t do an outcome, and watching a number you don’t control mostly just makes you anxious. So track the thing that’s actually in your hands. Not “hit 5k subscribers” but “publish 2 newsletters a week.” Stack enough of those and the outcome tends to sort itself out.

We're turning my goals into a plan I'll actually follow. 
Let's go goal by goal.

1. Help me pick the 3 goals to focus on for the next 12 weeks. 
Give me your recommendation and your reasoning, then let me 
make the final call. 
2. For each chosen goal, suggest 2-4 strategies (the real 
levers) and 1-2 lead measures. Lead measures are weekly actions 
I control ("post 4 reels a week", "pitch 5 clients"), not 
outcomes ("hit 50k followers"). Show me your thinking and let 
me adjust before you move on.
3. Then look at the whole list and challenge it. If I'm trying 
to do too much, say so. If 2 goals are fighting for the same 
hours, flag it.

And be ruthless about picking 3 goals to focus on. I know 3 feels stingy when you’ve just written a gorgeous long list. But 3 goals you actually hit in 12 weeks will do more for you than 9 you half-chase and feel guilty about all autumn.

Step 4: Review weekly

Reviewing is boring. There’s no dopamine in it, so it’s the first thing to fall off the second life gets busy.

But a tiny weekly check-in is the single habit I see in basically every person who actually gets where they said they would. It’s how you spot the patterns, stay accountable, and catch a bad week before it becomes a bad month and then a year off-track. Which is exactly why it has to be small and scheduled.

Once a week just spend 5 minutes in the same chat with Claude:

It's my weekly review. Ask me questions to understand the 
progress I've made towards my 3 focus goals.

Ask me, one at a time: what I actually did this week, 
what moved, what slipped. When I name the one change 
for next week, pressure-test it: is that the real lever, 
r the easy answer? If I had a low week, help me work 
out whether it was the wrong action or something I can 
genuinely remove, before I blame my discipline.

Give me an overview of everything at the end.

The low weeks are the most useful ones, by the way. If you only hit half your actions, the useful question isn’t “why am I so hopeless.” It’s “was that the right action, or did something get in the way that I can actually remove.”

In conclusion…

None of this works if you do it once and never open it again.

So if you only take one thing from this issue, make it this: take acton on this today, then schedule the weekly 5 minute check-in into your calendar and make it a non-negotiable now. That single recurring event is worth more than the fanciest goals doc you’ll ever build.

Last month my free weekly newsletter made $1,048.64 in ad revenue.

I didn’t have to create a course, or sell an anything or create lots of social media videos.

It was all just from ads that sit inside an email I was going to write anyway.

My first ever newsletter ad back in April 2025 paid me $2.80. I remember screenshotting it and thinking, WOW! I made money sending my first ever newsletter!

That number has continued to grow, and my best single issue since then has paid $414.06 (that was one send, in June). And last month was my biggest ad month yet at $1,048.64.

I’m 67 weeks into sending this thing every Tuesday, so this isn’t something that grows quickly, But it’s an additional income stream you can generate every week, with barely any additional effort. Just by sending a weekly newsletter (which you should be doing anyway as building an email list is hands down the best asset you can create for your business).

So with all that said, let’s get into how to make this happen using my favourite tool, Beehiiv.

Read more

Hey future-proof friends 💜

Last week I got invited onto an OpenAI pre-launch call where they walked us through their brand new tool: ChatGPT Work.

And my very first thought was: finally, OpenAI has an answer to Claude Cowork.

If you’ve been reading my newsletters for a while, you’ll know I’ve become slightly obsessed with Claude Cowork - the tool that lets AI actually DO work on your computer rather than just chatting with you. For months, it’s had no real competitor.

Now it does.

So today, I want to break down the two for you - what’s the same, what’s different, and (most importantly) which one you should actually use.

Let’s compare Claude Cowork vs ChatGPT Work…

Both of these tools do the same fundamental thing: they let AI take real action on your computer. Instead of just answering questions in a chat window, they can access your files, build documents, connect to your apps, and complete multi-step tasks for you.

ChatGPT Work is essentially based on OpenAI’s developer tool called Codex, but stripped down so regular humans (like you and me) can use it. It’s now available for you to access through the ChatGPT app on your desktop or phone.

Oh, and they also launched a shiny new model, GPT-5.6, which powers the whole thing.

So how do they actually compare? Let me break it down:

A Lot Is The Same

These tools are almost mirror images of each other. Both give you:

1. Skills - Reusable “recipe cards” that teach the AI to do specific tasks the way you like them done.

2. Live outputs you can interact with - Claude calls these “Artifacts.” ChatGPT calls them “Sites.” Same concept: working dashboards, documents, and tools created right inside the app.

3. Scheduled tasks - Both let you set up automations that run on a schedule (like my morning briefing magazine I’ve shown you before).

4. App connections - Claude calls these “Connectors.” ChatGPT calls them “Plugins.” Both let the AI read and act on your Gmail, Slack, Calendar, and more.

5. Folder access - Both desktop apps let you point the AI at specific folders on your computer to work within.

6. Mobile syncing - Both now let you start a task on your phone and pick it up on your computer.

Now For The The Key Differences

ChatGPT Work has native image generation (loooove!!)

This is a huge advantage. ChatGPT can generate images directly, built right in. With Claude Cowork, you’d need to connect an external tool to do the same. If visuals are part of your workflow, this is a huge point for ChatGPT.

ChatGPT Work lives in more places

You can use ChatGPT Work in the web app, the desktop app, AND the mobile app. Claude Cowork is a bit more desktop-focused (though mobile is now catching up).

Your Skills transfer between both

Here’s something genuinely useful: Skills are saved as simple files. So if you’ve spent months building brilliant Skills inside Claude, you can literally download them and upload them straight into ChatGPT Work. They work identically.

The models are different

ChatGPT Work runs on GPT-5.6 (with options to toggle between faster and more powerful versions). Claude Cowork runs on Claude’s models. Which one is “better” genuinely changes month to month as they leapfrog each other - so this shouldn’t be your deciding factor.

So... Which Should You Actually Use?

My take is: it genuinely doesn’t matter that much.

I know that’s not the dramatic verdict you might’ve wanted, but it’s the truth. Because they work almost identically, the “right” choice comes down to a few simple questions:

Pick ChatGPT Work if:

• You already live in the ChatGPT ecosystem

• You need built-in image generation

• You want to work seamlessly across web, desktop, and mobile

Pick Claude Cowork if:

• You’re already invested in Claude (like me)

• You prefer Claude’s writing quality and reasoning (I still find it stronger for creative work)

• You’ve already built a system that works

Personally? I’m sticking with Claude Cowork for now. I just prefer how it works, and I’ve built up a system I love. But I’m genuinely thrilled OpenAI has entered the ring - because competition means both tools get better, faster, and that’s a win for us.

What I Think You Should Do…

If you’ve been sitting on the fence about trying AI that actually does the work rather than just talking about it - now’s the time.

You’ve got two brilliant options, both racing to outdo each other, both getting more powerful by the week.

So this week:

1. Pick ONE (flip a coin if you have to - you can’t go wrong)

2. Try ONE simple task - you can try my morning briefing magazine hack here

3. See how it feels to delegate actual work to AI

And if you’d like more help with setting Claude Cowork (and now ChatGPT Work) up the PROPER way to get results that always sound like you, check out my step-by-step Voice Match guide.

Try Voice Match

How June even happened (the income dip I saw coming)

For most of this year I’ve been hovering around $13,000 a month, with the majority of that coming from freelance client work (and income from my creator business slowly growing).

Then back in May, because of a few unrelated things happening across different clients at once, I could see it coming: June was going to have a big dip in client work. So I braced myself. I knew I’d have to hustle to make up the shortfall.

And low and behold… All that freed-up time from the drop in client work, plus the pressure to cover the gap, made me more creative and more resourceful than I’ve been in ages. I went and found the money. And the result was my best month of the year so far.

Sometimes the dip (and the fear/kick up the bum that gives you) is the thing that forces the growth.

Read more

1. The morning briefing

What it does: A little digest every morning that pulls together your calendar, the emails that actually need you, and a few headlines from your world.

Connectors you’ll need: Gmail and Google Calendar connected to Claude Cowork with a scheduled tack (the news bit needs nothing, it just searches the web).

Set it up:

I actually wrote an entire in-depth article on how to set this up - you can read it here:

Claude builds me a personalised magazine every morning + plans my whole day. Here’s how...

2. The Sunday (or daily) reset

What it does: every Sunday evening it looks at your week ahead and hands you a plan.

Connectors you’ll need: Google Calendar & Google Mail connected to Claude Cowork with a scheduled task set up (plus your task list, for example if it lives in Notion or Google Tasks).

Set it up:

1. Set up your Google Calendar/Google Mail/Notion connectors in Claude.

2. Copy/paste this prompt into Claude Cowork:

Review my calendar, emails and task list for the coming week. 
Give me the 2 to 3 top priorities, flag any day with more than 
4 hours of meetings, and point out my best focus blocks.

3. Schedule weekly: /schedule every Sunday at 5pm: [prompt above].

You can also turn this into a daily reset and schedule it to happen every week day instead of every Sunday.

3. Inbox triage (with replies half-written for you)

What it does: each morning it reads your unread emails, sorts it into urgent / needs a reply / FYI / ignore, and drafts the replies that need writing. You skim, tweak and send.

Connectors you’ll need: Gmail connected to Claude

One important point: Claude drafts but never sends your emails, on purpose. You still need to review and click send.

Set it up:

1. Connect the Gmail connector.

2. Paste this into Cowork as your task, swapping out the bracketed bits:

Every morning, go through my unread emails from the last 24 
hours and triage it for me. Sort everything into four buckets:

Urgent: anything from [key clients or names], anything 
mentioning[an invoice, payment, deadline or contract], or 
a direct question that needs me today.

Needs a reply: real emails from real people that aren't urgent 
but do need a response from me.

FYI: things I should see but don't need to action (updates, 
receipts, confirmations).

Ignore: newsletters, promos and automated noise.

Then do three things:

Give me a summary at the top: how many emails in each bucket, 
plus a one-line note on every Urgent and Needs a reply email 
(who it's from and what they want).
For everything in Urgent and Needs a reply, draft a response 
in my voice and save it as a draft in Gmail. Keep replies short, 
warm and direct, the way I write. Nothing formal or corporate.
If you're not sure how I'd want to reply to something, don't 
guess. Flag it in the summary and leave it for me.

Do not send anything. Drafts only. I'll review and hit send 
myself.

3. Schedule it daily with /schedule every weekday at 7am: [prompt above]. It saves the drafts to Gmail, ready for you to review and send.

(Heads up: the sorting’s about 80% right on day one and gets sharper once you tweak the rules over a week.)

Let me know if you enjoyed these tips in the poll below and I’ll give you more ideas!

Hey future-proof friends 💜

For the first time ever, UGC was my single biggest income stream this month with $4,700.

Some of it I’ve been building for months (more on that), but the real trigger was that my usual freelance client work suddenly dipped. So I thought, right, I’m going to dig in and actually go and find more UGC work.

And it worked! It worked so well that the thing I treated as a side income is truly becoming the main event.

With UGC, if you have a decent portfolio, a few sample pieces and a willingness to pitch relentlessly, you can start landing paid jobs from day one.

You don’t need a social media following. or to “build an audience” for a year first. You just need proof you can create content.

That’s the whole reason I think this is one of the best side hustles going for women 35+ who want to start something now, without a personal brand or a social presence to lean on.

Has my personal brand helped? Massively, and I’ll be straight with you about where. But it is not essential, and you can 100% get started without a single follower.

Let me show you exactly how my $4,700 UGC month came together.

One final note before we start: I do mostly tech, SaaS, AI and software UGC, so I’m writing from that lens. But almost all of this applies to any kind of UGC. The niche changes. The playbook doesn’t.

Read more

How many brilliant ideas, useful articles, and incredible insights have you saved over the years... and then completely forgotten about?

Podcasts, videos, articles, notes… all lost in a sea of screenshots, bookmarks, and “I’ll come back to this later” save attempts in random folders.

The most successful people don’t just consume information… they build a system to keep it, connect it, and actually use it.

It’s called a “second brain” - and until recently, building one meant fiddling with complicated apps, and manually organising everything yourself.

But Claude has completely changed this. You can now build a second brain that organises itself, makes connections you’d never spot, and actually gets smarter the more you use it.

It takes about 30 minutes to set up and it’s incredibly easy.

Set a 30-minute timer and let’s build your second brain with Claude…

Here’s the concept (explained simply)

Think of it like having a personal librarian who:

• Takes everything you throw at them (articles, notes, ideas)

• Organises it all beautifully without you lifting a finger

• Connects related ideas together

• And answers any question you have using YOUR collected knowledge

You do the easy part (dumping information in). Claude does the hard part (organising, connecting, and making it useful).

The trick that makes this a self-learning system is one special instructions file that sits at the top of your folder.

Think of it as the rulebook. It tells Claude:

• How your folders are structured

• Where to put things

• What to do when you add new material

• How to keep track of what it’s already organised

Once that rulebook exists, Claude knows how to run the whole system every time - you don’t have to re-explain it.

Let me walk you through it properly.

Step 1: Create Your Folders + The Rulebook

On your computer, create one main folder (call it “Second Brain”). Inside it, create three subfolders:

• Raw - your messy dumping ground for anything you want to save

• Wiki - where Claude writes the clean, organized version (you never touch this - it’s Claude’s job)

• Outputs - where Claude saves answers and reports when you ask it questions

Then, open Claude Cowork, point it at your main folder, and ask it to create your rulebook:

Just copy/paste this instruction:

I'm building a self-organising knowledge base in this folder. 
It has 3 subfolders: Raw (where I dump information), 
Wiki (where 
you organise it), and Outputs (where you save answers 
to my questions).

Create an instructions file that explains how this system works, 
so that every time I add new material to Raw, you know to:
- Only process the NEW items (not redo everything)
- Update the organised Wiki
- Keep a simple log of what you've already processed

Keep it clean and simple.

This creates your rulebook and a little “memory log” that tracks what’s been organised. That memory log is the key - it’s how Claude knows what’s new each time.

Step 2: Dump Everything In (The “Raw” Folder)

This is the easy, ongoing part. Whenever you come across something valuable, drop it into Raw. Don’t organise it or make it pretty.

• A great article? Copy it into a document and save it in Raw

• Podcast notes? Dump them in

• A screenshot? Drop it in

• Your own random ideas? In they go

Messy is the goal. Capturing is the only job here.

Step 3: Let Claude Build Your Organised Library

The first time you do this, point Claude at your folder and say:

Read everything in my Raw folder and build an organised 
knowledge library in my Wiki folder, following the 
instructions file.

Create one summary document per topic, link related ideas 
together, and make an index so I can find anything quickly.

Walk away, make a coffee, come back to a beautifully organised library - grouped by topic, summarised, and cross-referenced.

When you add more stuff later (which you will, constantly), you don’t redo everything from scratch. You just drop the new items into Raw, then say:

Check my Raw folder for anything new since you last organised 
it, and update my Wiki accordingly.

Because of that memory log from Step 1, Claude knows exactly what’s already been processed and only adds the new material. It slots fresh ideas into existing topics or creates new ones.

Step 4: Ask It Anything

Now you can question your second brain, and it answers using everything YOU’VE collected:

Based on everything in my knowledge base, what are the best 
strategies I've saved for [growing an audience / saving money]?

Claude pulls from your Wiki and gives you an answer rooted in your own curated knowledge - then saves that answer into your Outputs folder.

And finally, here’s the loop that makes it “self-learning”: Those saved answers become new material the system can draw on. Plus, every so often, you can ask Claude to review the whole thing:

Review my entire knowledge base. Flag any contradictions, point 
out gaps where I'm missing information, and suggest connections 
between ideas I haven't made yet.

This is what turns it from a static folder into something that genuinely improves over time. It spots holes in your thinking, surfaces forgotten gems, and connects dots you’d never see yourself.

One quick question: Want to train Claude Cowork to always sound exactly like you?

Check out Voice Match - my step-by-step guide to help you personalise your Claude Cowork to sound exactly like you, every single time it completes a task for you. Setup takes 60 minutes and saves you hours prompting + fixing bad AI outputs.

Try Voice Match Here