How Infinity Constellation Saved $100K a Year on Reporting
Infinity Constellation's 8 portcos ran on clashing data definitions, so leaders waited on stale reports. Now their team chats with live, trusted data inside Claude — an estimated $100K saved a year.
$100K
Saved a year on manual reporting
Ilan Man
Founder & CEO
Paradox Machines
Connect ↓
The Challenge
Infinity Constellation is an AI-services holding company with eight portfolio companies across different verticals. Each portco runs independently, with its own CRM, revenue management, and definitions for key sales objects like 'leads' and 'proposals' — so there was no shared language across the portfolio. Reporting was manual: by the time data reached HoldCo leadership it was stale and still required interpretation. Portcos spent meaningful time on reporting overhead instead of building their businesses, and without real-time visibility the HoldCo was stuck in reactive mode, unable to provide the value-added services it wanted to.
What They Built
Paradox Machines built Infinity Constellation a centralized, cloud-hosted data warehouse spanning its eight portfolio companies, deployed inside Infinity's own cloud. The team standardized business logic and definitions across the portcos, integrated each business unit's CRM and revenue systems via API, and exposed the clean, modeled data to Claude through an MCP custom connector. Rather than ship another portal, they met leadership inside Claude — the tool they already use — so stakeholders query live, trustworthy portfolio data in natural language and generate reports and insights on demand. The system is live in production.
Clean, modeled, trustworthy data exposed to Claude turns portfolio reporting self-serve: leadership moves from reactive, analyst-dependent reporting to proactive, real-time decision-making — optimizing capital allocation, spotting risks faster, and driving multiple expansion, at an estimated $100K/yr in reporting savings.
Paradox Machines ran the engagement as a lightweight, production rollout that leaned on tools Infinity Constellation already used. It began with about 30 minutes per portco to gather requirements and map how each company defined core objects like 'leads' and 'proposals.' The team then standardized that business logic across all eight companies into a shared data model, so figures could be trusted and compared. Deployment ran inside Infinity's own cloud and took a few days of configuration. Once each business unit provided API keys, wiring up each integration took minutes, with live data flowing into a centralized warehouse. The final step exposed the clean, modeled data to Claude through an MCP connector — configured in a few clicks inside Claude itself ('add custom connector,' paste the MCP URL). Rather than ship another portal, the team deliberately met stakeholders where they already worked: leadership now queries live, trusted portfolio data conversationally in Claude and generates reports and insights on demand. The result is real-time visibility across the portfolio and a standardized foundation that makes onboarding the next portco fast.
AI Role
Claude's agent sits on top of the centralized, standardized warehouse via an MCP connector. Non-technical leadership query live portfolio data conversationally inside Claude — surfacing value-creation opportunities, tracking KPIs, and pulling reports themselves, with no separate portal, no analyst in the loop, and no second-guessing the numbers.
Infrastructure
• Centralized cloud data warehouse, deployed inside Infinity Constellation's own cloud environment • Source systems: each portfolio company's CRM and revenue-management tools, connected via API • Standardized, shared data model across the eight companies
Integration Points
• Per-business-unit API integrations feeding live data into the central warehouse • Standardization layer mapping each company's definitions (leads, proposals) to a shared schema • MCP custom connector exposing the modeled warehouse data to Claude's agent
Impact
Estimated annual savings from eliminating manual reporting time across the eight portfolio companies.
~2 hrs/week saved per portco
Automated, real-time reporting removes roughly two hours of manual reporting per portfolio company each week.
Self-serve, trustworthy data
Non-technical stakeholders query live, trusted company data directly in Claude — no pinging analysts, no second-guessing the numbers.
Implementation Complexity
Standardized business logic across eight portcos, then ingested and centralized data from each company's separate CRM and revenue systems into shared datasets, and layered AI querying on top. Designed to be lightweight: ~30 minutes to gather requirements per portco, under a week to deploy the centralized platform, and under an hour to integrate each new data source, with data available in real time thereafter.
Best Fit For
Holding companies, private equity firms, and multi-entity operators managing several portfolio companies on fragmented CRMs and inconsistent data definitions — teams that need real-time, trustworthy reporting for leadership and LPs, and want self-serve access to portfolio data without standing up a new tool.
