One prompt is all it takes to get an answer to almost anything these days. Whether it’s digging into a random rabbit hole, researching a medical condition, or something less objective like financial advice, AI has made information feel instant and effortless.
When we are talking about financial and Investment Advice, the difference between a well informed answer and the right answer for you specifically is very different.
Don’t get me wrong, the language models’ outputs are coherent and well reasoned, but they respond to a prompt and give an answer, and the quality of that answer depends entirely on the quality of what you ask. In theory, a well constructed prompt that outlines your full financial picture could generate something that resembles personalised guidance. But resembling advice and actually being advice are two different things. Regulators draw that line clearly, and for good reason.
Many investment decisions still require a qualified human adviser because accountability and contextual judgment can not be replicated in a prompt. Even if you could, there are some human, irrational aspects to an investor that AI simply can not grasp.
Telling your AI to “act like a expert financial advisor” doesn’t change the fact that it’s just a next word predictor at the end of the day.
What does research tell us
Earlier this year, Stuart Kirk from the Financial Times wrote about an experiment that illustrates this. Dr. Bruce Lloyd, emeritus professor of strategic management at London South Bank University, took a single investment prompt and ran it across eight of the leading AI models: ChatGPT, Claude, Gemini, Grok, Copilot, Perplexity, Qwen and DeepSeek.
Every model came back with the same foundational principles intact: diversification, disciplined rebalancing, tax efficiency. It was all present and accounted for. Their output was actually quite reasonable.
Copilot and Gemini approached the task with the rigour of a seasoned analyst, quantitative, risk conscious, and focused on low-volatility quality shares. Gemini went further with style tilts and factor elements such as small-caps and growth. Grok and ChatGPT cut through the noise, favouring straightforward, low-cost implementation without overcomplicating things. Qwen kept a close eye on forex exposure. Perplexity hedged everything through a UK lens and recommended holding 10% in cash for what it called “psychological comfort.” DeepSeek was a committed believer in classic portfolio theory. Claude suggested keeping traditional fixed-income exposure lean at 20%.
Most suggested 60 to 65% in equities, with Copilot and Perplexity slightly lower at 55%. On fixed income, Grok and Qwen pushed as high as 35%, with around 15% in corporate credit.
When Dr. Lloyd pushed further and asked the models to blend everything into a single “best of all worlds” portfolio, the result was: 50% in a global equity ETF, 10% in UK equities, 25% split across gilts (UK government bonds) and a global bond fund, 5% in property and infrastructure, 5% in gold, and up to 5% in cash.
On paper, it was just your standard well balanced portfolio, built for anyone but no one in particular.
The perfect portfolio for the prompt, but you are more than prompt. Aren’t you?
So where does AI actually belong in your investing journey?
- Learning the basics: If you want to understand what an ETF is, how compounding works, or what expense ratio means, AI is perfect for the task
- Decoding financial documents: Fund factsheets and prospectuses are not written for the average reader. You can use AI to extract specific points or give you simple summaries.
- Preparing for conversations with your adviser: Using AI to get up to speed before a check in means you arrive with better questions and get far more out of the conversation.
- Building conceptual understanding: Asking things like “what tends to happen to bonds when interest rates rise” or “how has inflation historically affected long term returns” is education, and AI handles that well.
- Stress testing your assumptions: This is one of the best uses of AI in investing. You can ask it to play the devil’s advocate and question every assumption you make, minimizing your blind spots when you make investments.
All in all, investing decisions are never made in a vacuum. They are shaped by your full financial picture, your emotional relationship with risk, and the specifics of your life at this particular moment. An AI only ever sees the version that you choose to share in a prompt, and that is rarely the full picture.