Designing an AI mentor that helps first-time investors take their first step

I read 32 public studies on retail investors to check whether the pattern held beyond one platform. It did. Three barriers kept coming back, and all of them had to do with feelings or knowledge.

Access was already there, and people still didn't feel ready. So I designed a mentor that makes the first steps feel safe.

• 01A mentor that never recommends

  Fear was the main barrier in research, and in a regulated market an assistant that only educates earns more trust. So the mentor explains concepts and scenarios and suggests the next question, but never says what to buy or sell.

  Gave up

  Direct buy or sell recommendations.

  Got

  Trust, with every decision left to the person.

• 02It starts with how you feel

  43% of retail investors name fear of losing money as their main barrier. That's why the first question is how you feel about investing, the language puts no pressure, and information comes in small steps.

• 03Two characters

  Navi leads the explanations. Meanwhile Artemis, inspired by my own cat, notices hesitation and celebrates small wins, so beginners feel accompanied while they learn.

• 04Every pattern traces to research

  When to ask a follow-up, slow down or show a disclaimer came from studies on fear, drop-off and trust. As a result, every choice targeted a known problem.

  Gave up

  Nice-to-have features.

  Got

  Focus on the confidence gap, and fast decisions in 48 hours.

I tested it the way a real beginner talks, changing direction mid-conversation and asking follow-ups like “and what if I changed my goal?”. Each failure became a design rule for the next version.

• V1One generalist agent

  One agent answered every message, with no routing and no memory.

  Got

  Good answers to clean, one-off questions.

  Limit

  It only remembered one message at a time, so the context got lost.

• V2Specialized agents

  To fix that, I split it into instruction, decision and clarity agents, picked by keyword.

  Got

  Better answers than one generalist agent.

  Limit

  Without the expected keywords, people hit a generic error.

• V3Router and memory

  Then I added intent detection and short-term memory of goals, amounts and doubts.

  Got

  Follow-ups connected back to the original plan.

  Limit

  It still missed edge cases and emotional nuance, with no clear line on what it could say.

• V4Safety locks

  Finally, safety prompts blocked direct investment recommendations, while disclaimers explained risk and length limits kept each answer short.

  Got

  Plain answers that explain risk and leave every decision with the person.

  Limit

  Real investors and compliance teams haven't tested it yet.

The result is a working prototype, built in 48 hours. The next step is testing it with real investors and compliance teams.

In a conversational product, the dialogue is most of the design. Every line is a product decision, so I planned the empathy from the first message.

Guardrails can be part of the value. In a regulated market, an assistant that only educates earns more trust.

Knowing the backend widened my toolkit. Wiring agents, live market data and memory let me invent patterns like contextual follow-ups.

Talk to Navi in your own words. Ask what a word means, look up a stock, or share a goal and then change your mind.