Built by people who sat in the committee room.
We didn't build this from a product roadmap. We built it because we watched firms with excellent consultants lose margin and people to an allocation process that turned a 90-second scoring problem into a two-week committee exercise.
Why Nicole Built Crewpath
Nicole Yamamoto spent eight years in professional services operations at mid-size consulting firms in the Southwest before starting Crewpath. Her job was to make allocation decisions — and she was good at it. She kept mental models of which consultants had chemistry with which clients, who was approaching burnout, whose bench window was closing in the next three weeks.
The problem was not her judgment. The problem was that she was the only one holding that model in her head. Every time the firm grew past 80 consultants, every time a new partner joined with different preferences, every time she took a two-week leave — the institutional knowledge either disappeared or got diluted into a committee that talked over each other for three hours and defaulted to the same 15 names.
Crewpath started as a personal project in 2021. Nicole wanted to know whether the signals she was reading intuitively — project-type recency, post-engagement feedback, bench timing, daily rate economics — could be formalized, weighted, and scored. The first version was a Python script running against a CSV export from the firm's PSA. It disagreed with the committee's top choice on 27% of engagements. When she reviewed those divergences, the model was right more often than the committee.
She rebuilt it as a proper scoring engine, connected it to PSA exports from other firms, and started showing it to operations leads she knew. The pattern repeated: the data was already there. The gap was a tool that could process it in the time it took a committee to find a conference room.
Crewpath does not replace professional judgment. It replaces the two-week cycle that separates professional judgment from a ranked candidate list.
In early 2026, Crewpath raised angel funding to accelerate PSA integration development and expand the firm onboarding team.The team
Operating Principles
The engine recommends. Partners confirm or override. Every override is logged with a reason code, and over time those overrides calibrate the model toward your firm's actual priorities. The decision stays with a human. The data accumulates with the model.
The composite score breaks down to four dimensions. Each dimension breaks down to input signals. A partner can trace why Consultant A ranked above Consultant B before confirming. No black boxes — because a model your firm can't interrogate is a model your firm won't trust.
The default weight model is a starting point calibrated on professional services allocation data. It is not the right model for every firm. M&A boutiques weight domain expertise differently than change management generalists. The admin panel is there because we expect firms to tune it.