Can You Really Build an Automated Investment System With No Experience? I Tried It

Can You Really Build an Automated Investment System With No Experience? I Tried It

Six months ago, I would have laughed at anyone who suggested I could build my own automated investment system. I’m not a developer. I don’t have a computer science degree. My background is in marketing, and the most technical thing I’d done was set up some basic WordPress sites and fumble through Google Analytics.

But here I am, running PocketBots, writing about the automated systems I’ve built, and watching them execute trades while I sleep. And I’m going to be completely honest with you about how I got here — the scepticism, the frustrations, the genuine surprises, and the real costs involved.

The Scepticism Phase: Does This Actually Work for Normal People?

Let me start where most of you probably are right now. I’d seen the YouTube videos. The Twitter threads. The “I built a trading bot and made £10,000” posts. And my reaction was always the same: yeah, but you’re probably a software engineer with ten years of experience.

The people sharing these success stories always seemed to gloss over the actual building part. They’d show the results, maybe share some code snippets, but never really explain how someone without their background could realistically follow along.

I was convinced this wasn’t for people like me. Automated trading systems were for hedge funds, quant traders, and that one friend who always talks about Bitcoin at parties. Not for a bloke in his kitchen who still Googles “how to exit Vim” every single time.

But I kept coming back to one thought: if the tools exist, and the APIs are public, and people claim it’s possible… what if I actually tried?

What It Actually Took: The Honest Breakdown

I’m not going to pretend this was easy. It wasn’t. But it also wasn’t the insurmountable technical mountain I’d imagined.

Here’s what it actually required:

• Time: About three weeks of evenings and weekends. Maybe 40-50 hours total to get the first working version.
• Tools: A laptop, a few free accounts, and eventually a cheap virtual private server.
• AI assistance: This was the game-changer. More on this shortly.
• Patience: Loads of it. I broke things constantly.

The mistakes I made along the way could fill their own article. I accidentally created infinite loops that crashed my computer. I exposed API keys in code I shared (quickly fixed, lesson learned). I spent four hours debugging an issue that turned out to be a missing comma. Just one comma. Four hours.

There were moments where I genuinely considered giving up. Around week two, I hit a wall with authentication that seemed completely impenetrable. I couldn’t understand why my code wasn’t connecting, why the error messages were so unhelpful, why something that seemed so simple was so frustrating.

The Moment It Actually Worked

And then, at around 11pm on a Tuesday, I ran the script one more time. I wasn’t even expecting it to work — I’d just made another small change based on something Claude suggested, and I hit enter more out of habit than hope.

The terminal sat there for a moment. Then:

“Connected! Portfolio value: $100,000”

I actually said “no way” out loud to an empty room. It was paper trading — simulated money, not real — but the connection was real. The bot was reading actual market data. It was calculating positions based on the strategy I’d defined. It was working.

I sat there for probably twenty minutes just watching it run, making decisions, logging its reasoning. Something I’d built. Something that would keep running whether I was watching or not.

That moment changed my perspective entirely. Not because I’d suddenly become a master programmer, but because I’d proven to myself that the barrier to entry wasn’t as high as I’d believed.

What Genuinely Surprised Me

The biggest surprise was how accessible this had become thanks to AI assistants. I used Claude extensively throughout this process, and I want to be transparent about that because it fundamentally changed what was possible.

When I hit the authentication wall I mentioned earlier, I didn’t have to spend days reading documentation I barely understood. I could describe the problem conversationally — “I’m getting this error when I try to connect to the Coinbase API, and I don’t understand what it’s asking for” — and get explanations in plain English, along with specific suggestions to try.

AI didn’t write everything for me. I still had to understand what I was doing, make decisions about strategy, and debug problems. But it was like having a patient tutor available at any hour who never made me feel stupid for asking basic questions.

Five years ago, this project would have taken me six months of learning prerequisites. Today, the prerequisites can be learned contextually, as you need them.

What Was Harder Than Expected

Not everything was smoothed over by AI assistance. Some things were genuinely difficult, and I want to be honest about those.

API keys and authentication: Every platform handles this differently. Coinbase was particularly frustrating — their Advanced Trade API has specific requirements for key generation, and their documentation wasn’t always clear about which permissions were needed for which actions. I probably spent a full day just on authentication issues.

Coinbase specifically: I’ll call this out separately because it was my biggest headache. Their API has gone through several iterations, and a lot of online tutorials reference older versions that no longer work the same way. Finding up-to-date, accurate information was genuinely challenging.

Getting articles to publish cleanly: This is a meta-frustration. Even writing about these systems and publishing to my blog involved technical hurdles — formatting issues, image hosting, broken links. The whole ecosystem of building and sharing required learning adjacent skills I hadn’t anticipated.

Psychology: Watching a bot make decisions with real money (even small amounts) is nerve-wracking at first. There’s a different kind of difficulty in trusting automation you’ve built yourself.

What It Actually Costs

I want to give you real numbers because I’ve seen too many “build passive income” posts that hide the costs.

• Virtual Private Server: £5/month (I use a basic DigitalOcean droplet)
• Domain and hosting: £3/month
• Trading platform fees: Variable, but Coinbase’s fees are built into spread
• AI tools: I use the free tiers mostly, occasionally pay for Claude Pro when working intensively

Total monthly overhead: approximately £8-10.

That’s genuinely it. The capital you trade with is separate, of course, but the infrastructure cost to run an automated system is less than a Netflix subscription.

What I’d Do Differently If Starting Again

Looking back, several things would have saved me time and frustration:

1. Start with paper trading from day one. I was too eager to test with real money (tiny amounts, but still). Paper trading lets you make all your mistakes without financial consequences.
2. Pick one platform and stick with it initially. I tried to build connections to multiple exchanges early on and it multiplied my headaches. Master one first.
3. Document everything as you go. I lost time re-figuring out things I’d already solved because I hadn’t written them down.
4. Join communities earlier. There are Reddit communities, Discord servers, and forums full of people who’ve solved the exact problems you’ll encounter.
5. Set realistic expectations about timelines. I thought I’d have something working in a week. It took three. That’s fine, but adjusting my expectations would have reduced frustration.

An Honest Assessment of Returns

I’m not going to promise you riches. That would be dishonest, and honestly, it would be irresponsible.

What I can tell you is this: my systems generate modest but consistent returns. We’re not talking “quit your job” money. We’re talking “covers my monthly subscriptions and then some” money. Over time, with compound growth, that becomes meaningful.

More importantly, I’ve built something that generates returns without requiring my constant attention. The value isn’t just in the pounds generated — it’s in the hours not spent checking charts, making emotional decisions, or trying to time the market.

Some months are better than others. Some strategies work in certain market conditions and not others. This isn’t a money printer. It’s a tool that, when built thoughtfully, can contribute to long-term wealth building.

The Answer to the Big Question

Can you really build an automated investment system with no experience?

Yes. Genuinely, yes.

But with caveats. It requires patience. It requires accepting that you’ll make mistakes. It requires being honest about what you don’t know and willing to learn. It requires managing expectations about both the building process and the returns.

The tools available today have genuinely democratised what’s possible. The same fundamental capabilities that used to require engineering teams are now accessible to individuals willing to put in the effort.

You don’t need to be special. You don’t need a technical background. You need curiosity, persistence, and access to the right guidance.

Want to Build Your Own?

That’s why I started PocketBots. I went through this process largely alone, piecing together information from dozens of sources, making mistakes that better guidance could have prevented.

I’m now documenting everything — the strategies, the code, the pitfalls, the solutions. If you’re interested in building your own automated investment systems, I’d like to help you avoid the frustrations I experienced and get to that “Connected!” moment faster.

Whether you want to automate cryptocurrency trading, build portfolio rebalancing tools, or create systems that generate genuinely passive income, the path is more accessible than you think.

Get in touch through PocketBots, and I’ll guide you through the same process. Not with hype or unrealistic promises, but with honest, practical steps that actually work for normal people.

Because if I can do this, I genuinely believe you can too.