Carbon RE

The Project

Carbon Residential is a rapidly growing property management and real estate investment firm. The project involved a comprehensive digital transformation, moving the firm away from fragmented legacy software and manual processes into a unified, AI-driven "Carbon OS." This system automates compliance, market research, and administrative tasks, significantly reducing operational overhead.

The Challenge

As Carbon expanded its portfolio, its operational infrastructure failed to scale. The firm faced several critical inefficiencies:

• Stale Market Data: Property managers relied on manual calls or delayed third-party reports (like CoStar, often three months behind) for rent comparisons, leading to sub-optimal pricing.

• SOP Non-Compliance: Critical processes like delinquency tracking and tenant application reviews lived in static Google Sheets or Word documents. There was no systemic way to ensure or monitor employee compliance across multiple properties.

• Data Silos & "Software Slop": Essential data was locked inside legacy systems like Yardi (which possessed an incredibly poor UI), forcing employees to manually extract, parse, and re-enter data across platforms.

• Institutional Knowledge Loss: Critical business logic and operational knowledge lived entirely in the heads of key executives, creating massive risk during employee turnover.

  
  

The Solution

Codelitt partnered with Carbon to architect an end-to-end, AI-native operating system designed to automate repetitive tasks and enforce process compliance.

• Live AI Market Research (The "Rent Comps" Tracker): We built a fully automated competitive analysis tool. Using a combination of web scraping (bypassing Cloudflare bot detection) and NVIDIA’s voice API via Twilio, the system autonomously calls competing properties. It interacts with human leasing agents to gather real-time data on unit availability and pricing, aggregating the results into a centralized dashboard.

• Automated SOPs & Task Governance: We integrated data directly from Yardi and CRM platforms into BigQuery and visualized it using Coda and low-code environments. The new system automatically generates daily delinquency queues and task lists, forcing employees to interact with the data and ensuring 100% visibility into SOP compliance for management.

• AI-Driven Leasing & Administrative Automation: We implemented a "Lease AI" bot that handles overflow leasing calls and deployed AI agents capable of reviewing tenant applications (credit reports, pay stubs) against Carbon's strict criteria, reducing approval times from days to minutes.

• Claude Skills & Knowledge Capture: We established an internal AI org structure with custom Claude Skills. This included an automated AI exit-interview agent designed to extract and document institutional knowledge from departing employees directly into GitHub repositories.

  
  

The Result

• Real-Time Revenue Optimization: The Rent Comps tracker allowed Carbon to accurately price renewals and new leases based on live market data, directly impacting revenue efficiency and minimizing occupancy loss.

• 100% Operational Visibility: Management transitioned from guessing about employee performance to having a comprehensive oversight dashboard that tracks SOP compliance across five core domains and dozens of workflows.

• Dramatically Reduced Overhead: By automating tenant screening, market research, and data entry, Carbon achieved significant labor savings, allowing the existing team to manage a growing portfolio without a proportional increase in headcount.

• Systematic Governance: Established strict code-drift and AI-slop prevention protocols, ensuring that non-technical employees could leverage AI tools without compromising the firm's codebase or data integrity.

  
  

Tech Stack

• Claude / Custom AI Skills & Agents

• NVIDIA Voice API / Twilio

• Yardi API Integration / BigQuery / Google Cloud

• Coda / Blocks DIY / Custom Python & JS Dashboards