Copy article

AI in portfolio construction

ended 24. April 2026

Is it possible for someone to structure an investment portfolio purely with AI? What specialist tools out there exist to this end or can you use the likes of well-known agents like Claude and ChatGPT? What are the risks of doing this? Can people save money on advice by using AI engines that will understand every single thing about established investment strategies and asset allocation? At the very least, can people use AI to assess the portfolio their adviser has created? Any thoughts, insights or intel, whether you're an adviser or AI expert, send them across ASAP as publishing this story TODAY.

5 responses from the Newspage community

Copy all

Copy

Editor
Anita Wright
Chartered Financial Planner at Ribble Wealth Management
AI can already do more than most investors realise. The technology has moved fast further and faster than most people tracking it from the outside appreciate. But AI cannot replicate the human element. Understanding a client's real relationship with risk not the answer they give on a questionnaire, but the one they give at 11pm watching markets fall requires empathy, experience and trust built over time. These are not variables an algorithm can reliably capture.
There are regulatory limits too. AI cannot take regulated financial advice. If it goes wrong, there is no ombudsman to call. The smarter use case may be as a second opinion. Uploading your adviser's portfolio to an AI and asking it to challenge the rationale, check for over concentration or flag high charges is entirely free and surprisingly powerful. AI will not replace good advisers. But it will increasingly expose the ones who are not adding value.
Copy

Editor
Scott Gallacher
Director at Rowley Turton
AI can be a useful tool, but it often makes basic errors while expressing a high degree of confidence. Without expert oversight, that’s a dangerous combination when applied to investment decisions. Building a portfolio isn’t just about theory — it requires judgement, context, and an understanding of individual circumstances. Used carefully it can support the process, but relying on it alone is a risk most investors shouldn’t take.
Copy

Editor
Graham Nicoll
Financial Planner, Chartered FCSI at NCL Wealth Partners
AI is a powerful tool that can support individuals to evaluate their options, evaluate advice they are being provided and prompt questions to ask an advisor. That said, it is fraught with challenges if being relied upon as the sole source for financial and investment decisions as the output can be wrong, outdated or confused.
Searching online for insight has been used by many people for a number years using the likes of google search, AI tools have prompted more people to use this route. Use these tools to support research but proceed with caution on the outputs.
Copy

Rohit Parmar-Mistry
Founder at Pattrn Data
AI can help with portfolio construction, but it is not a substitute for judgment, suitability or regulated advice. It can explain diversification, compare allocations, model scenarios and stress test ideas, yet the risk is mistaking polished output for fiduciary thinking. A portfolio is never just maths; it depends on goals, tax position, time horizon, liquidity needs and behavioural tolerance when markets turn volatile. While people with real technical expertise can now use tools like Claude to build and monitor far more advanced personal systems, including automated strategies in crypto or prediction markets, that is still high-risk territory requiring skill, constant oversight and a clear understanding of failure points. For most investors, AI is best used to challenge assumptions, improve understanding and help them ask better questions, not replace accountable advice outright, even if robo-advice may be getting closer. But it is not there yet for most people today, still, overall.
Copy

Colette Mason
AI Ethics Consultant at Clever Clogs AI
There is a widespread global misunderstanding of what "an AI agent" can do reliably. Those people who have tried to tame a wayward agent understand the risks. The people who listen to the 'guru hype' do not. Agents are simply a set of repeatable instructions sent to a tool like Claude. They aren't magic wands. Natural language is loose. The probabilistic nature of AI means it will interpret those instructions differently, but plausibly, every time. Plausible information is not accurate information. Using AI to sense-check what a regulated adviser has already built for you, fill in jargon knowledge gaps, is genuinely useful. However, a regulated adviser carries fiduciary duty and professional indemnity insurance. ChatGPT carries a disclaimer in tiny print: "AI can make mistakes." It is imperative to remember using AI is like pulling a fruit-machine lever, and there is a good chance you will be stung by the outcome. Artificial is correct. Intelligence is a gross overinflation of quality.