AI potential rests on the power of data

Will artificial intelligence transform wealth management? The potential is certainly there. But like AI use cases everywhere, the degree of impact will depend on the value of the outputs – which will rest in turn on the quality of data inputs. Garbage in, garbage out has never been so true.

As a recent Financial Times article observed, up to 80% of AI projects fail – in large part because of poor underlying data.

“Without good, accurate data, that is also well-organised, unbiased, un-siloed, and optimised in the right ways, [an AI project] rollout will probably fail to get past a basic level,” it warned.

AI in wealth management

We are already seeing the diverse benefits AI integration can bring to wealth management firms and their clients. It is enabling more efficient operations, better risk management and more personalised customer experiences, while allowing human advisors to focus on value-adding tasks such as complex strategic planning and relationship building. Current applications include:

Robo-advisors have long used AI to create and rebalance portfolios automatically based on client risk profiles and goals. Machine learning models and AI algorithms can increasingly analyse vast datasets – from economic indicators and market sentiment to company financials and customer activity – to identify market patterns, generate investment recommendations, optimise asset allocation and execute timely trades.

AI systems can evaluate portfolio risk more comprehensively by analysing correlations across asset classes, stress-testing portfolios against different market scenarios and monitoring real-time risk exposures. They can also assess client-specific risks by evaluating spending patterns, income flows and life circumstances on the fly.

AI enables highly-customised financial advice based on detailed client information, modelling different scenarios for retirement planning, tax optimisation, estate planning, etc. to provide personalised recommendations that adapt as circumstances change. Predictive analytics can also suggest proactive outreach opportunities to help wealth managers provide more timely and relevant advice.

AI-powered chatbots and virtual assistants handle an ever-widening selection of (increasingly complex) client enquiries and requests, freeing staff to focus on more value-adding activities. Natural language processing helps scrutinise client communications to identify concerns or opportunities to offer additional services.

AI systems can monitor trading activities in real time for regulatory compliance, detect suspicious transactions and ensure adherence to fiduciary responsibilities. They can also track regulatory changes and automatically update compliance procedures.

And we are only at the beginning. The lightning-fast pace of technology innovation means new and more powerful use cases are emerging all the time.

Treat data as a product

To stay competitive, every firm will need an AI strategy that encompasses some or all of these applications. More importantly, they need AI-powered processes to work. Building models on a bedrock of clean, clear, consistent, golden source data is fundamental to:

With poor data quality leading to flawed investment recommendations, incorrect risk assessments and suboptimal portfolio allocations that can translate into significant financial losses.

Stale, incorrect or missing data can cause AI models to reach erroneous judgements, miss opportunities or make poorly-timed trades that hurt portfolio performance.

That can tap into accurate historical and real-time data to identify potential threats and navigate volatility.

Based on a complete client profile and picture of their investment objectives and preferences, allied with up-to-date information on market conditions and economic indicators to deliver personalised investment strategies and financial planning. Poor data quality results in generic/inappropriate recommendations that fail to meet individual client needs.

Where high-quality, well-documented data is essential to ensure AI-generated reports are accurate, and AI recommendations are explainable and auditable, and so able to withstand regulatory scrutiny and preserve client trust. Incorrect or incomplete data risks compliance violations and knock-on reputational damage.

Poor quality data corrupts an AI system’s learning process and compounds mistakes over time, producing models that become progressively less accurate and useful.

Despite the profound implications poor data has for the viability of wealth managers’ AI rollouts, the data challenge remains underappreciated and overlooked. Every firm across the industry has (often unrecognised) data problems. We are on a mission to help institutions create a golden source of data they – and their AI models – can rely on. Wealth managers that embrace that mission will be far better placed to reap the many rewards AI has to offer.