Midtown Advisory Group is a 40-person wealth management firm in Midtown Manhattan. They were burning 30+ hours a week reconciling client reports across three custodians — Schwab, Fidelity, and Pershing — with manual compliance checks slowing every client review.
The Challenge
Every night, the ops team manually pulled data from 3 custodian portals, matched it against their internal books, flagged anomalies, and compiled compliance reports. One error could mean a missed trade, a fine, or a lost client.
The Solution
We built an RPA + AI hybrid pipeline: bots pull custodian data nightly, an AI agent flags anomalies against historical patterns, and a reconciliation bot generates review-ready PDFs. Compliance gets a full audit trail automatically.
The Results
30 hours/week saved, 100% audit trail coverage, client reviews ship in under 24 hours instead of 5 business days. The firm onboarded 8 new clients without expanding operations headcount.
The Stack
UiPath (RPA), Claude (anomaly detection), QuickBooks (accounting), and a custom compliance dashboard built on Next.js + Prisma.
Lessons for NYC Finance Firms
If you're doing manual reconciliation across multiple custodians, you're burning money. RPA + AI can cut reconciliation time by 80%+ while improving accuracy. Start with one custodian, prove the ROI, then expand.