How to Build a Recruiting Function From Zero at a Vehicle Startup

Most vehicle startups make the same first recruiting hire. They bring in a coordinator or a generalist, hand them a job board subscription and a spreadsheet, and point them at a stack of open reqs. Three months later the pipeline is thin, offers are slow, and agencies are back in the picture.

The problem is not the person. The problem is the architecture they inherited. A recruiting function that starts wrong takes 12 months to rebuild, and in a launch-stage business, you rarely have 12 months to spare.

I Aecthe recruiting function from zero for the U.S. launch of AFEELA at Sony Honda Mobility: 48 hires, 9 functions, 12 months, 98% offer acceptance, $1.5M+ in annual agency savings. Here is how that architecture worked and what I would do the same way again.

The Intake Before the First Req

When I joined AFEELA there were no open reqs, no ATS, and no sourcing channels. There was a business that needed to hire across engineering, manufacturing, supply chain, sales, marketing, finance, legal, HR, and operations before a product launch.

The first thing I did was not open a req. I spent the first two weeks in intake conversations, not about roles, but about the business: what was launching, in what order, where the timing dependencies were, what the candidate pool looked like for each function, and where the real risk was if we were understaffed at launch.

That intake work made everything after it faster. When a req opened, I already knew the business reason, the timing dependency, the must-haves versus negotiables, and what the offer trajectory would look like. Time to fill averaged 42 days against an industry average of 60 to 75. The intake is why.

Most startups skip this step because it feels like delay. It is actually acceleration. Every hour in intake saves three hours of back-and-forth when a candidate is in process and the hiring manager's expectations turn out to be different from what was in the job description.

ATSSelection and Ops Architecture

Before the first req posts, two decisions have to get made: where candidate records live, and how offers flow through the organization. These decisions shape everything downstream.

For early-stage vehicle companies, the ATS should optimize for speed of configuration and recruiter flexibility, not feature count. A platform that requires three months of IT setup and professional services is the wrong choice even if the reporting suite is superior. The intake template, req structure, and disposition logic all need to be live before sourcing begins.

This is also where the agency dependency question surfaces. If the ATS is not configured for direct sourcing, and if no sourcing channels are activated, the path of least resistance becomes agency submissions. Once that pattern starts, it is structurally difficult to stop. Each placement reinforces the dependency and the cost model that goes with it.

Sourcing Architecture for Mobility

Direct sourcing in mobility requires vertical-specific channels that generalist recruiting does not have. LinkedIn Recruiter is table stakes. What differentiates is sourcing depth: aerospace and defense talent communities for propulsion and certification roles, OEM alumni networks for launch-phase manufacturing and quality, university pipeline programs for battery and embedded software, and technical Slack communities and GitHub activity for AI and autonomy functions.

I built sourcing channel lists by function before the first req opened. Not because every channel would be active immediately, but because knowing where to go when a role opened eliminated the two to three day delay of figuring it out under deadline pressure. That delay compounds across a 48-hire program.

There is a secondary benefit to this architecture: it makes the recruiter irreplaceable in the right way. A sourcing system that lives in documented channel lists and search strings can be handed off or scaled. A sourcing system that lives only in one person's head creates fragility, not capability.

The Offer Process That Gets to 98%

Offer acceptance at an industry average of 85 to 90 percent means one in ten candidates who cleared your entire interview loop said no. In a launch environment, a single unexpected decline can shift a program timeline by weeks.

The pre-offer call is the single highest-leverage move in the process. Fifteen minutes before the formal offer goes out: What would make this a no-brainer? What would make it a no? Are there competing offers or active conversations we should know about?

At AFEELA, 48 of 49 offers were accepted. The pre-offer call surfaced every competing offer and unresolved objection before the PDF landed. It also created a relationship that holds through the onboarding gap, which matters at a launch-stage company where start dates can shift and candidates need a reason to stay committed through the uncertainty.

The other piece of the offer process that gets underestimated: same-day candidate debriefs after each interview round. If a candidate has concerns, you want to know in the loop, not after the offer. If the hiring manager has reservations, you want to know before the committee meeting, not in it. The debrief cadence is the early warning system for every potential decline.

The Agency Elimination Math

Agency fees in vehicle startups typically run 20 to 25 percent of first-year compensation. At a $150,000 average comp, that is $30,000 to $37,500 per hire. For a 48-hire launch across 9 functions, a fully agentized model would have cost $1.4M to $1.8M in fees alone.

What direct sourcing costs is a senior recruiter's time and the ATS subscription. The math is obvious to anyone who does it on a spreadsheet. The reason most startups do not do it is that building the sourcing function requires front-loaded investment: the intake process, the channel architecture, the offer process design. All of that happens before a single hire is made.

That is the real reason agencies persist in early-stage vehicle companies. They offer immediate capacitty with no infrastructure investment. The cost is not visible until you pull 12 months of invoices together and do the math. When I ran that exercise at AFEELA, the savings were $1.5M+ annually. That number had not been calculated before because no one had looked at it in aggregate.

The $243 cost per hire I averaged across 782 hires at Gurley Leep over 28 months tells the same story at volume: direct sourcing with the right architecture produces fundamentally different unit economics than agency dependency.

What to Build First

If I were walking into a vehicle startup today with a mandate to build recruiting from zero, the sequence would be: intake process before the first req, ATS before the second week, sourcing channel lists by function before the first search begins, and pre-offer call process baked into the workflow before the first offer goes out.

The coordinator hire, the agency contract, the job board subscription: those can come later or not at all, depending on volume and stage. The architecture comes first.

The vehicle companies that get this right early do not just hire faster. They build institutional knowledge about their own talent market that compounds over time. The recruiter who has run 48 hires across 9 functions in 12 months knows things about that market that no agency ever will.