Newshot

PsyFi checks AI-generated dollar amounts against bank records

Original text(via HN AI)

Our Approach

  • We established the application architecture, user authentication, financial personality onboarding, and account management workflows. We also developed the initial subscription, trial, and credit-based access features to support PsyFi's commercial model.
  • We integrated Plaid for financial account connectivity and developed backend services to manage transactions, investments, budgets, and spending information across CAD and USD currencies. We also built a dedicated service to retrieve live market data for investment-related features.
  • We helped build Fy - an AI financial wellness coach, using Amazon Bedrock and Claude, with specialized coaching modes, proactive insights, and access to users' financial records. We also developed a grounding layer with 15+ AI insights agents, including Cash Flow, Category Spending, Savings Rate, Investment Insights, Asset Allocation, Mood Spending, and Loss Aversion.
  • We integrated Flinks, added AI-powered statement processing, and developed investment tracking, financial goals, and insights connecting spending patterns with users' moods. These insights help identify spending categories associated with difficult emotional days and behavioral patterns such as impulsivity, loss aversion, and overconfidence. We also introduced saved AI summaries, habit tracking, and financial scenario modelling.
  • We conducted unit tests, AI response evaluations, staging reviews, and client acceptance testing as a part of our rigorous QA process. We also strengthened security, monitoring, and deployment workflows and migrated the AI chat system to Vercel AI SDK through a gradual transition.
AI financial wellness platform foundation with onboarding, user authentication and subscription access features.
AI financial wellness platform foundation with onboarding, user authentication and subscription access features.

The text above was extracted automatically from the source page and may miss images, tables or code blocks. The original is authoritative.

Published by HN AI
This site only summarises and aggregates; copyright stays with the original author. Headlines and summaries follow the source — where they differ, the original wins.
Read the original ↗