Portfolio construction has long been a careful balancing act between risk, return, and investor psychology. Traditionally, it has been siloed, relying on historical data, human judgment, and static models. Classic portfolio construction follows a linear
path: where risk tolerance and investment goals are assessed first, assets across predefined categories are allocated, and market shifts dictate how the portfolio is rebalanced.
While this model has worked for decades, in today’s world, it is limited because of its reliance on historical averages and manual oversight. In 2025, portfolios must be capable of rapidly adapting to fast-moving markets or nuanced investor behaviours. Gone
are the days where clients can be treated as categories, rather than individuals.
In recent years, namely in 2022 and 2023, retail investors withdrew more money from mixed asset strategies than they invested. This shows that investors have been disappointed in their returns, and this can be put down to insufficient diversification or
a lack of evolution of markets amid uncertainty and turbulence. Modernisation of both traditional equity and fixed income portfolios is overdue.
According to T. Rowe Price, the “traditional equity and fixed income
‘balanced’ portfolio needs modernisation.” Further, this portfolio “has not evolved to reflect today’s investment environment.” Beyond diversification, investors must know to make the most of the opportunities available for return enhancement and risk reduction
for both equities and fixed income portfolios.
T. Rowe Price reiterates that portfolios should “include a variety of safe-haven assets with a view to increasing portfolio resilience in periods of market turbulence. More attractive yields within fixed income markets, in particular, mean that existing
portfolio allocations may need to be revised.
“Finally, as we look ahead to an uncertain path with the possibility of both geopolitical and economic shocks, it is worth reviewing if tools such as volatility management may have a wider role to pay.” It is evident that in 2025 and beyond that there must
be a shift from reactive strategy to proactive precision, while also reimagining what it means to invest.
This article will explore what is required for these changes to succeed and how AI can play a role in transforming how portfolios are built, managed and evolved.
Enter AI: Dynamic, data-driven, and personalised
Today, the investment management industry needs a competitive edge. AI can provide this by augmenting human judgement and basic qualitative models by rapidly analysing market data and risk factors through ML and predictive analytics. By picking up on subtle
asset correlations within a range of alternative investments, portfolios can be transformed based on predicted market volatility.
Reinforcement learning and neural networks can also help optimise portfolio allocation with risk tolerance, while AI and ML also offers a holistic and real-time view of potential threats to portfolio performance. AI-powered early warning systems can detect
subtle signals of impending market stress before they become apparent, ensuring portfolio managers are better prepared for adverse market conditions.
Optimisation algorithms can also enhance portfolio construction by establishing portfolios that are tailored to specific goals and risk profiles, without constant human intervention and in turn, without emotional biases. AI systems can also incorporate spending
habits and values to create customised portfolios that extend to tax optimisation.
AI changes the game by introducing: