Building a robo-advisory engine for a neobank: our approach

What a modern robo-advisor needs

A useful robo-advisor isn't just "fill out a risk quiz, get a portfolio." That's a 2015 product. A modern one needs:

• Dynamic risk profiling that updates from behaviour instead of a one-time quiz
• Cross-domain context so investment decisions know about your banking data
• Awareness of life events so the portfolio adjusts when something changes
• Micro-investing built into the same flow
• Explanations so users understand what the system is doing and why

The data advantage of a neobank

Building it inside a neobank gives you a real data advantage. Traditional robo-advisors like Betterment or Wealthify work with isolated investment data - they know your declared income and risk tolerance, and that's about it.

A robo-advisor inside a bank can see:

• Your verified income from actual deposits
• Your monthly expense baseline from transaction history
• Your upcoming large expenses from recurring patterns
• Your emergency fund status
• Your business cash flow if you're an SME user
• Tax signals like year-end expenses or self-employment patterns

That changes recommendation quality a lot. If the system knows your rent is due in 4 days and your salary hasn't landed yet, it shouldn't be telling you to invest this week. If your income jumped 30% this month, it can bump up your contribution automatically.

Architecture

Our architecture has the banking data layer feeding a single real-time financial model of the user, which feeds the investment engine - risk profiling, portfolio construction, rebalancing triggers, the micro-investing flow, and the explainability layer.

The key decision: the investment engine reads from the shared user model. It doesn't keep its own separate profile.

Regulatory compliance

Investment services in the EU fall under MiFID II, which adds requirements around suitability, best execution, cost disclosure and appropriateness testing. We're building MiFID II compliance in from the start, which means the explainability layer isn't just nice UX - it's a regulatory requirement.

Micro-investing

One feature people underrate: micro-investing. Round-up investing sounds like a gimmick. The data says otherwise. Users who turn on micro-investing invest 3–5x more over a year than users who start investments manually, because removing friction works.

Our flow:

1. A transaction completes (say €4.60 coffee)
2. The round-up gets calculated (€0.40)
3. Round-ups build up in a buffer
4. At a threshold the user sets (say €5) the investment runs automatically

The user watches the portfolio grow alongside their spending, in the same session, same context.

What we haven't solved

Honest note on what we haven't solved: real-time rebalancing at scale is still a work in progress. The challenge isn't the algorithm - it's executing trades efficiently across thousands of accounts at once without slippage or latency problems. We're looking at execution partnerships for that part.

Y-tech Bank is pre-seed. We're building this stack and looking for engineers who want to work on hard fintech problems. ytechbank.com