Case Study
Why a Global Automaker's Foresight Team Chose Crustdata Over Crunchbase and PitchBook for Stealth Startup Tracking
Company
A Global Automaker's Innovation Center
use case
Stealth Startup Tracking & Partnership Sourcing
company overview
This global automaker operates three innovation centers worldwide as part of its Group Innovation division. The Silicon Valley center, a three-person corporate foresight team, scouts early-stage technology companies for co-development partnerships, product integration, and venture investment across the automaker's portfolio of 10+ brands and multiple venture arms.
The foresight team's mandate is to find emerging technology companies before competing automakers reach them. Crunchbase and Pitchbook only showed companies that had already raised publicly, and their three-person team couldn't scout fast enough to cover their mandate.
The team scouts stealth and pre-seed companies building technology relevant to the automaker's product roadmap.
Crunchbase Pro and PitchBook only tracked companies that had already exited stealth or raised publicly. By the time a startup appeared in either database, rival automakers had already initiated conversations
PitchBook sourced much of its early-stage data from Crunchbase, meaning both subscriptions shared the same coverage gap
Every missed stealth-stage company was a co-development partnership or venture opportunity that went to a competitor OEM by default, delaying access to technology that could have shaped upcoming vehicle programs
The Silicon Valley team operates at 50% capacity on scouting, splitting time across other innovation projects.
Evaluating a single startup required the engineer to look it up on professional profiles, track down each founder individually, and check for a company website that often did not exist
Coverage was limited to however many companies one engineer could research in a week, while the mandate required monitoring hundreds of startups across multiple technology verticals
The engineer was also fielding internal IT support requests, further reducing available scouting hours
The automaker's IT security requirements created a barrier most data vendors couldn't help the team clear.
Cloud-hosted AI platforms like Claude and ChatGPT were not approved due to security vetting, so the team needed a data provider whose APIs worked within their already-approved tools, specifically VS Code with OpenAI's Codex
Any vendor that required IT to whitelist a new cloud platform was off the table, which ruled out most data providers that relied on their own front end or dashboard for access
Crustdata's Company Search API with 95+ filters surfaces companies at the earliest stages of formation, including those that have not appeared in Crunchbase, PitchBook, or any traditional startup database
During the evaluation, the team ran searches for emerging companies in specific technology verticals and found startups they were already in conversation with (validating accuracy) alongside companies they had not yet discovered through their manual process
The real-time search indexes companies from professional profiles, social activity, and web signals, catching startups before they have press coverage or established profiles
The engineer connected Crustdata's MCP server to Codex in VS Code and pointed it at the team's tracking list. In one session, the integration enriched every company with headcount data, founder profiles, social posts, and company details
The foresight lead began building a Python parser in VS Code to call Crustdata's APIs from natural language queries, turning an engineer-dependent workflow into something the whole team could operate
With 5,000 credits per team member and three credits per 100 results, the team had capacity for over 166,000 results, compared to a competing platform that exhausted its trial in two to three searches
The team scoped Crustdata's Watcher API for push-based notifications when companies matching their criteria emerge or change status, replacing the manual monitoring that only caught signals for startups they already knew about
The foresight lead outlined criteria for passive filters, including new stealth companies in target verticals, founders departing large companies, and first key hires at recently formed startups
This monitoring layer lets the three-person team cover their scouting mandate continuously rather than in weekly bursts limited by the engineer's availability
The engineer enriched the team's entire tracking list through the MCP integration in a single session, replacing a process that previously required opening each profile individually and manually recording details for companies that often had no website
The recovered hours went toward the internal engineering requests that had been accumulating across the innovation center, clearing a backlog that grew while scouting consumed the engineer's time
The foresight lead and analyst gained the ability to run their own queries, so three people could scout in parallel instead of funneling through one technical bottleneck
Crustdata returned early-stage companies in target technology verticals that did not appear in either Crunchbase Pro or PitchBook. The team now knows about startups that were previously invisible to them
With those companies on their radar, the foresight team can reach out before competing automakers become aware these startups exist, instead of arriving after the opportunity has already been shopped around
The team also validated accuracy by checking results against companies they were already in conversation with. The data matched, which confirmed the new companies Crustdata surfaced were worth pursuing
The Silicon Valley team tested Crustdata, found value, and began building the case to roll it out to the full German platform team of 10-15 engineers who run the tech intelligence platform for the entire automotive group
The foresight lead started drafting a presentation showing specific companies Crustdata surfaced that Crunchbase and PitchBook missed. Concrete examples of coverage gaps made the internal pitch far simpler than a generic vendor comparison
Once the German team integrates Crustdata, stealth-stage scouting scales from three people in Silicon Valley to every engineer and affiliate across the group, which means more startups found earlier and more technology options available for upcoming vehicle programs