Giani Brussich — Revenue Analytics & Decision Systems

I find the number everyone already trusted, and it wasn't true.

Nearly ten years building the analytics and decision systems behind lead routing, agent performance, and compliance at Morgan Stanley, Moody's, HomeServe, and now National Debt Relief. Outside of that I build software for real estate appraisers, an AI automation practice for small businesses, and I'm usually in the middle of evaluating a small business acquisition.

assumed:the pilot failed
actual:the phone carrier was flagging valid numbers as spam
Field Notes

Assumed vs. actual

A routing model

assumed:3–8% forecast error, good enough
actual:rebuilt to hold under 0.5% MAPE

Nobody had measured the old model against what was actually achievable. It hadn't been rebuilt in years. It now sets daily routing configuration.

A compliance clock

assumed:complete
actual:only recorded time for calls that connected

The clock excluded exactly the failures it existed to catch. Rebuilt on the right timestamp, it now flags potential violations across 4M+ dials, all 50 states.

A test that looked broken

assumed:the strategy failed
actual:a carrier was marking valid numbers as spam

Contact rates dropped across every group mid test. The test wasn't the variable that moved. Nearly half the numbers being dialed were flagged before anyone picked up.

A deal that looked clean

assumed:matched the listing summary
actual:verified against four months of raw statements

Sellers show you the number they want you to see. The real one is underneath it, in the statements nobody asks for.

Experience

Nearly ten years, four seats

2025—Present

National Debt Relief — Revenue Analytics Manager, Data & AI

Lead AI-enabled analytics across sales and operations on ChatGPT Enterprise and Claude Code. Built the lead-routing model above and the 50-state compliance monitoring system.

2024—2025

HomeServe USA — Manager, Revenue Management

Led forecasting, monthly reporting, and budgeting for a subscription business. Built an LLM-assisted Excel validation tool and streamlined the Alteryx/Dataiku pipelines feeding Tableau and Power BI.

2020—2023

Morgan Stanley — AVP, Strategy & Solutions

Led the rollout of automated, self-service Tableau reporting adopted by centralized teams. Automated revenue-cycle analytics in SQL for senior leadership.

2015—2020

Moody's Corporation — Associate → Senior Operations Analyst

Led a Six Sigma initiative that improved operational efficiency 50%. Delivered SOX and Dodd-Frank regulatory reporting and drove a firm-wide access program to 100% adherence.

Proof

What I've actually built

Acquisition · Lead analyst

Small business due diligence

Ran full financial due diligence on a family owned retail business: proof of cash reconciliation against raw statements, seller's discretionary earnings normalization, a twelve month cash flow model compared across financing structures, and a lease risk analysis that ended up being the reason to walk away.

Status: passed at the negotiated terms. The model stays ready if better terms come back.
Product

Property Appraisal Pro

A PostgreSQL and PostGIS platform for real estate appraisers, built against MISMO 3.6, UAD 3.6, and USPAP standards. Includes modules for comp justification, adjustment rationale, and automated QC flag detection.

Status: in development, active inbound interest from appraisers.
Practice

NYClaw

An AI automation and lead generation practice for small businesses in Westchester and NYC. Same workflow audit and systems approach used at scale, applied to businesses that have never had it.

Status: early, pre client.
Approach

The part that doesn't change

Set the number before you look

Every decision has a real constraint attached to it: a budget, a floor, a deadline. It usually surfaces halfway through instead of getting stated up front. Naming it first means the options that were never actually in play stop costing time before anyone touches them.

The honest fit

When something doesn't match exactly, say so before making a case for it. Then point to the closest real example instead of the best-sounding argument. That's what's still standing once someone actually checks.

Built

Technical work

SQL (Snowflake) · Python · R · Sigma · Tableau · Power BI · Git · MCP

Writing

Notes on building

Longer write ups on acquisitions, analytics, and building with AI, posted first on X and cross posted to LinkedIn.

Now

Currently

Running revenue analytics at National Debt Relief. Building out PAP's core schema. Staying in the market for the right acquisition at the right terms.