Robo-Advice

Fewer than 10 per cent of UK adults have received regulated financial advice. The advice gap is not a mystery: a full financial review costs 1,000 to 3,000 pounds, which makes it inaccessible for anyone without substantial assets. Robo-advice offers a path to closing that gap, but the regulatory framework for automated advice is precise, and the firms that have succeeded built for compliance first and technology second.

What is robo-advice?

Robo-advice is the provision of automated, algorithm-driven financial planning and investment management with minimal human intervention. The underlying risk assessment models determine suitability. A customer completes a risk questionnaire, the system recommends a portfolio (typically constructed from index funds or ETFs), and the platform manages ongoing rebalancing, tax-loss harvesting, and reporting. The entire process, from onboarding to ongoing management, operates digitally.

The regulatory classification matters. In the UK, if the service makes a personal recommendation based on the customer's circumstances, it is regulated advice under FCA rules and the firm must be authorised accordingly. If it provides information and tools that help the customer make their own decision without a personal recommendation, it is a guided investment service. The distinction affects capital requirements, complaint liability, and the scope of the firm's duty to the customer. Most UK robo-advisers operate in the regulated advice space, which provides stronger consumer protection but higher compliance costs.

The business model depends on scale. Margins per customer are thin, typically 25 to 75 basis points of assets under management annually. To cover technology costs, regulatory capital, and compliance infrastructure, a robo-adviser needs tens of thousands of active customers with meaningful portfolios. Several UK entrants have struggled to reach profitability, leading to consolidation: standalone robo-advisers acquired by larger platforms, banks, or asset managers that can distribute the service to an existing customer base.

The landscape

The FCA's Consumer Duty adds new obligations for robo-advisers. The suitability of the recommendation must deliver good outcomes, and the firm must monitor those outcomes on an ongoing basis. A risk questionnaire that categorises a customer as "moderate risk" and places them in a standard portfolio is not sufficient if the firm cannot demonstrate that the portfolio has performed in line with reasonable expectations for that risk profile. Consumer Duty shifts the obligation from point-of-sale suitability to ongoing outcome monitoring.

The EU AI Act classifies AI systems used for creditworthiness assessment and financial product recommendation as high-risk. Robo-advice platforms operating in the EU must meet requirements for transparency, human oversight, and documentation. The risk questionnaire algorithm, the portfolio construction model, and the rebalancing logic all fall within scope. Firms operating across the UK and EU face dual compliance requirements.

Hybrid models are emerging as the dominant architecture. Pure robo-advice (fully automated, no human involvement) serves cost-sensitive customers with simple needs. Hybrid advice (automated analysis with human adviser review for complex situations) serves the mass affluent segment that wants digital convenience with the reassurance of human oversight. The most successful platforms offer both, routing customers to the appropriate model based on their complexity and preference.

How AI changes this

Risk profiling is moving beyond static questionnaires. AI models incorporate behavioural data, transaction patterns, and life event signals to build a richer picture of the customer's risk capacity and risk attitude. A customer who panics and sells during a market downturn has a different actual risk tolerance from the one they reported in a questionnaire completed during calm markets. Behavioural analysis captures this gap. This connects directly to credit scoring innovations that use transaction data to understand financial behaviour.

Goal-based planning replaces risk-level categorisation. Rather than asking "are you moderate or aggressive?", AI-driven platforms ask "what are you saving for, and when do you need the money?" The system models the probability of achieving each goal under different portfolio allocations, presenting trade-offs in terms the customer understands: "there is a 75 per cent chance of reaching your house deposit target by 2028 with this allocation." This is more meaningful than a risk score and supports the Consumer Duty requirement for consumer understanding.

Tax optimisation at the individual level is becoming automated. AI systems consider the customer's ISA allowance, pension tax relief, capital gains position, and income tax band to recommend the most tax-efficient way to invest each new contribution or handle each withdrawal. This was previously the domain of human financial planners and is where robo-advice can deliver value that justifies its fee even for financially literate customers.

Personalised communication uses generative AI to explain portfolio performance, market events, and rebalancing decisions in language tailored to the customer's financial literacy level. A first-time investor receives a different explanation from a sophisticated self-directed investor, even though the underlying portfolio action is the same. The same Consumer Duty analytics that monitor outcomes across product types should monitor the robo-advice channel, tracking whether customers understand and benefit from the recommendations. This supports customer engagement and retention, which are the metrics that determine commercial viability.

What to know before you start

Suitability is not a one-time check. The FCA expects firms to monitor the ongoing suitability of their recommendations, particularly when market conditions change materially. A portfolio that was suitable when recommended may become unsuitable if the customer's circumstances change or if market movements push the asset allocation outside the target range. Build ongoing suitability monitoring into the platform, not as an afterthought but as a core feature.

The risk questionnaire is a regulated document. Every question must be defensible, every scoring mechanism explainable, and every outcome demonstrably linked to the customer's circumstances. The temptation to "improve" the questionnaire with AI-generated questions or dynamic scoring must be weighed against the regulatory requirement to validate the suitability assessment process. The FCA has issued specific guidance on risk profiling tools. Read it before redesigning yours.

Customer acquisition cost is the metric that kills robo-advisers. The technology works. The compliance framework is manageable. The business fails because acquiring a customer who will invest 10,000 pounds costs more than the revenue that customer generates in the first two years. Distribution is the strategic question: integrate with a bank's existing customer base, embed within an employer's workplace savings platform, or partner with a financial planning provider. The same predictive analytics that improve portfolio construction can also identify which existing bank customers are most likely to benefit from and adopt the robo-advice service. Standalone customer acquisition via digital marketing is prohibitively expensive for most firms.

Start with the simplest advice journey: ISA investment for a single goal with a straightforward risk profile. Build the full compliance infrastructure (suitability assessment, ongoing monitoring, complaint handling, regulatory reporting) for that single journey before expanding to pensions, SIPPs, or multi-goal planning. The regulatory architecture is the hard part. The portfolio construction is the easy part.