Why I built Anna

I started investing at 19, buying my first Bitcoin when most people advised against it. I got hooked early, on the mix of intuition and using knowledge to predict what happens next. I’ve always been curious about how markets and code can work together.

So when ChatGPT came out in late 2022, my first thought wasn’t “where’s the business opportunity?” It was “how do I use this for the thing I already enjoy?”

The eleven sources that didn’t check out

I asked AI to help me understand a stock I was looking at. The answer was confident, clean, and well written. It had eleven sources at the bottom, so it looked trustworthy. Then I started checking them.

The first source didn’t contain the number the AI had quoted. The second didn’t mention the claim. The third was a dead link. Another was behind a paywall. Twenty minutes later, I still hadn’t verified the answer.

And it didn’t happen once. It kept happening, different companies, different questions, the same pattern: confident answers, weak sources, no real way to trust the output.

At the same time, I was hitting the other problem every serious investor knows: the market moves too fast. The ASX gets thousands of announcements and filings every week. I was trying to properly follow ten companies, and even that was hard. Most tools were too slow, not connected to live data, or only analysed what I manually uploaded.

That was the moment Anna started to become obvious to me.

Four things have to be true

AI in finance is only useful if four things are true at once:

• Real data, not second-hand scraped content.
• Verification, every claim links back to the actual source.
• Real-time, because in markets, a delay can matter.
• Full coverage, because the important announcement is often the one you forgot to check.

The idea: don’t make the model sound trustworthy. Make it verifiable. Wire every answer back to the filing it came from, so trust is earned, not performed.

So I went directly to the source. We got a direct ASX real-time data licence and built Anna around the actual filings. Every announcement, filing, and market update lands in our system as it’s published. When you ask Anna a question, it doesn’t search random web pages, it runs an analysis straight from the source that you can see and validate.

That part matters to me. If Anna gives you a number, you should be able to see where it came from. If Anna isn’t sure, it should say so.

A confident wrong answer is worse than no answer.

That’s the gap I’m trying to close.