COR's $30M: AI is coming for project profitability

Here's a question most services businesses can't answer in under a week: did that job make money? Not the quarter — that job. On July 28, COR closed a $30 million investment from FTV Capital to attack exactly that gap with AI, aimed at agencies and professional services firms. The round is worth your attention less for the company and more for what it says about where project profitability software is heading: from a report you run after the fact to a system that tells you mid-project that you're underwater.

What actually happened

COR announced the $30 million investment from growth equity firm FTV Capital in a BusinessWire release dated July 28. The San Francisco company was founded in 2017 by Santi Bibiloni, Jose Gettas and Gabriel Marin, and sells an AI-powered project profitability platform that pulls project management, automated time tracking, resource planning and profitability analytics into one place. Per Orrick, which advised FTV on the deal, the platform is built specifically for agencies and professional services firms. FTV's Alex Malvone and Tommy Tighe join the board.

The stated use of funds is the tell: accelerate agentic AI capabilities, expand into new verticals, keep growing internationally. Translation — the automated time tracking is the wedge, and the agent layer on top is the product they think they're selling in two years.

Why project profitability matters for your business

If you bill by project — an agency, a consultancy, a contractor, a dev shop — your margin problem is almost never a pricing problem. It's a visibility problem. Scope creep gets absorbed silently. Someone's hours land in a timesheet three weeks late. The change order gets verbally agreed and never invoiced. By the time the P&L shows it, the job is closed and the lesson is theoretical.

The thing worth stealing from this category is the mechanic, not the vendor. A project is profitable or not the moment cost rates, actual hours, and billed amounts sit in the same table. That join is the whole product. Everything above it — the dashboards, the alerts, the agent that pings you at 60% burn — is a thin layer over one clean dataset. Most shops we talk to have all three inputs already, scattered across a time tracker, a payroll export, and QuickBooks, with no key linking them.

So do the plumbing before you buy the platform. Give every job a real ID that appears in your time entries, your invoices, and your cost data. Pick loaded cost rates per person and write them down. Get the three sources into one warehouse table on a nightly sync. That is a couple of weeks of unsexy integration work, and once it exists, the AI question becomes cheap — you can wire an alert, a weekly summary, or an agent to it in an afternoon. Buy the tool if the tool fits. Just don't buy it hoping it will fix data hygiene you never did, because it won't, and you'll be paying a seat price to keep exporting spreadsheets.

Sources: BusinessWire, Orrick.