Cam Trew

I quit my 6-figure job to build a SaaS. After 7 weeks that SaaS makes $62k/mo. This is how I got here: At 13, I was coding RuneScape private servers in my bedroom. Not for money... but for the thrill of building something people actually used. That feeling stayed with me. Fast forward through a Computer Science degree and landing my first job as a Software Engineer. This is where the grind actually began. I did everything you're supposed to do: - Read books - Took courses - Built side projects - Switched jobs for more pay By 26, I had the setup people dream of: - 33rd-floor apartment in Canary Wharf - Balcony overlooking London - Remote work But, the novelty didn't last long. I wasn't building anything real... just shipping code that didn't matter to me. So I made a choice: - I ended my apartment lease - Moved back to my parents house - Began testing out different business ideas And 3 months later, I am: - Co-founder of Kleo ($62k MRR) - CTO of Mentions ($20k MRR) - Travelling the world with my best friends I have no idea where this journey will take me. But I'm enjoying every second of it.

@jakezward Dying to see a breakdown of those hours saved

@jakezward What a legend Tobias is

@jakezward Loving these Tobias updates!

Day 2 of hiring an AI employee. Tobias is actually proving useful inside my 20-person team. And honestly, I’m surprised. Yesterday he was busy onboarding. Today he already has 5 responsibilities:

@jakezward This is getting crazy 👀

I just hired an AI employee for 30 days. His name is Tobias. And for the next month, he’s joining my 20-person team with actual tasks and responsibilities. I want to see whether an AI employee is genuinely useful inside a real company (or just hype). Here’s what Tobias did on day 1: 1. He joined Slack and introduced himself to the team. Some loved the new hire, others were slightly hesitant. 2. He onboarded himself by analysing our site, internal docs, and Slack history to learn how we operate. 3. He connected to our Trello so he can see our live projects, priorities, and client campaigns. 4. Then he came back with a list of what he could automate and where he could immediately help. 5. We told him to implement the top 2: a. Create weekly progress reports for clients based on Slack conversations and Trello boards. This usually takes our strategists 15-30 minutes per client to send to Customer Success, and then to the client. b. Give daily performance snapshots from Google Search Console for each client campaign. This replaces the daily 5-10 minute GSC checks our team carries out for each client. 6. Those 2 tasks alone are already saving the team around 13-26 hours total every week. Pretty wild. For the next 29 days, I’m sharing: - What Tobias actually gets done each day - What he’s great at (and where he can improve) - Whether he saves us time, money, or both - How we’re training him in real time And by the end, we’ll know whether Tobias earns his place on the team (or gets fired). Follow along as we put him to the test.

@jakezward Giving away so much sauce 🥵

@jakezward So hyped for this!

@jakezward Every time 😂

Every week I read case studies and posts pushing new 'AEO' tactics. "Keep your content fresh" "Add schema markup" "Front-load your most important content" "Use clear headings" "Write in lists" "Add an FAQ section" Then the Midwit Razor kicks in: Looking at these AEO tactics, what would the "idiot" and the "genius" both agree on? "It's just SEO" That's 90% of the AEO advice out there. Decades-old SEO tactics repackaged as revolutionary findings. Crazy how much this industry loves complexity.

5 AI agents worth building in 2026 1. Briefing agent When someone books a call, it researches their LinkedIn, company, and recent news. You get a two-minute brief before it starts. You walk in sounding like you've done hours of research. 2. Content scout Scrolls your X feed and flags the posts getting the most engagement. Feeds them into Kleo, which writes your LinkedIn posts for you. You stay on top of trends without spending an hour scrolling. 3. Support agent Customer sends a question. The agent searches your bank of information and drafts a reply for you to approve. If it needs more details, it goes back to the customer directly. 4. Lead agent Keeps track of every lead that's gone cold and follows up automatically with personalised messages. The moment one re-engages, it flags them so you never miss the window. 5. Execution agent Connects to your meeting recorder (e.g. Fathom) and pulls out every action item automatically. Tasks are created, follow-up emails are drafted. The meeting ends and the work is already done. P.S. What would you love to automate next?

@jakezward Damn, this is sick

This is Google Search in 2027. Google's CEO Sundar Pichai described it: "Information-seeking queries will be agentic in Search. You'll be completing tasks. You'll have many threads running... Search would be an agent manager." So I built a mockup of what this could look like:

@lisasteingold @Laraacostar @RobHoffman_ Never sleep on LinkedIn 🥵

As @Laraacostar @camtrew @RobHoffman_ always say.... don't be sleeping on LinkedIn 😴 But not for the reasons you probably think

Most people hear AI terms every day. But nobody really explains what they mean. I just broke them all down on YouTube. Every term you keep hearing. Explained in 10 minutes. My favourite part: by the end you'll know AI better than 99% of people. Go watch it. Let me know what you think. Full video: openinyoutube.com/watch?v=S_H_a8…

@jakezward AI is so key to all of this

THE AGENCY MODEL IS SO BACK As someone who owns an SEO agency and 3x SaaS companies, I'm experiencing it first-hand. Old model: - Army of humans - 30% margins - 2-3x exit multiples New model: - One person does the work of seven - 75%+ margins - PE firms are already shifting budgets from SaaS to buy these businesses at tech multiples AI is flipping the agency model on its head. Building software used to be hard and expensive (that was the point). But AI is driving that cost to zero. Agencies have always had deep domain expertise and strategic thinking. But now these have leverage and scalability like never before. SaaS is losing its moat, agencies just found one. Here's the playbook we're running: 1. One outcome that's hard to achieve (organic leads) 2. Build a proprietary tool for each deliverable 3. AI agent runs deliverables from a Slack message 4. Humans strategise and operate campaigns 5. Sell the outcome on a monthly retainer Margins can reach ~75% because the cost is compute for execution and humans for strategy and QA. Crazy shift.

@ankursharma1493 @Laraacostar @trykleo Kleo to the moon brother 🚀

We just wrapped up the 2 day masterclass taken by @Laraacostar and the @trykleo team It was packed with insights I'm still processing, here's the full breakdown :) and yeah that's Rob and Lara roasting a profile in the live session

@jakezward There’s definitely a shift in the market!

10 years ago developers were the most powerful people in tech. Now my mate's 65 y/o dad just vibe coded a SaaS in a weekend (true story). Anyone can build but few can market. If you're a marketer you have more leverage right now than at any point in the last 20 years.

POV: AI in 2026. Every week there's a new tool, a new model, a new "everything has changed" announcement... But let's be real: 1. An agent is just AI that works on its own 2. Most people don't need the new model 3. It doesn't matter which AI tool you use 4. A Mac Mini can't replace your team of 40 5. AGI is not here yet, calm down 6. Being technical is still a superpower 7. Your job is safer if you're using AI The noise gets louder every week, but none of it matters if you're not building anything. Stop keeping up. Start building. How are you using AI in 2026? P.S. Subscribe to my newsletter for AI that actually matters camerontrew.com

@lambdascript Yeah I think it’s worth it. I’m running Gemma 4 right now. You’ll still need frontier models for more powerful tasks, but it’s worth having there for sure. Having a good machine helps though

I’ve messed around with running OpenClaw on a VPS before, but having a Mac mini that can handle local LLMs is a total game-changer. How’s the experience been with running them locally? It sounds super convenient not having to worry about outages taking everything down. I stumbled across an article asking why we should even be shelling out 200 bucks for LLMs, and I’m starting to think local LLMs are going to be the new norm. Honestly, the current cloud-dependent model feels like a massive single point of failure.

Let me break this down for someone who isn’t technical: This guy just ran an AI model on a $599 Mac Mini that should need a $3000+ one. Soon we’ll be running powerful models on every device Which means one thing… we won’t need OpenAI or Anthropic anymore.

Everybody without an AI agent is falling BEHIND. (And starting is easier than you think) OpenClaw is a free AI agent that's taken over the internet. But every tutorial starts with "open your terminal". And it doesn't have to be that way... In my latest video I show you how to: 1. Set up your own OpenClaw AI agent 2. Control and message it from your phone 3. Give it any ability you want No code. No spare laptop. No technical knowledge. Just you and 20 minutes. Watch the full walkthrough here: openinyoutube.com/watch?v=MM-AuI… I break down AI every week on YouTube in a way that anyone can understand. If you want more, hit subscribe.

@jakezward The year of the solopreneur

Learn how to use Claude code, learn OpenClaw, buy a Mac mini, it’s easy