You’ve been told to “improve neurodiversity hiring”—but the funnel still has to ship hires
You get the message: “Improve neurodiversity hiring.” Then you look at your open reqs, the backlog, and the hiring managers who already think the process is “too slow.” You can’t swap in a whole new system and hope it works later. The funnel still has to produce hires this quarter.
What usually happens is small, well-meant tweaks—one training, one line in the job ad—while the real filters stay the same. The risk is also real: interviewers worry they’ll “say the wrong thing,” and candidates worry that any signal will cost them the offer.
Where are we accidentally selecting for “good interviewing” instead of job performance?
It shows up when a candidate can’t “tell a clean story” in 60 seconds, or doesn’t make strong eye contact, and the panel quietly labels it a lack of ownership or confidence. In practice, you end up rewarding fast verbal processing, polished social timing, and comfort with ambiguity—skills that matter in an interview room, not always in the role.
Look for the steps where you rely on vibes because you don’t have shared yardsticks. Unstructured “culture fit” chats, rapid-fire behavioral questions, and group panels tend to favor people who can read the room and improvise. If the job is writing specs, debugging, or handling tickets, then a candidate who asks for clarification or pauses to think may be doing the work, not failing it.
Building a clear rubric and aligning interviewers takes real hours you don’t have. But without it, you keep selecting for performance in the interview, not on the job—and you won’t fix that without changing the format.
Job ads that don’t quietly demand a personality type
Changing the format starts earlier than the interview: most job ads quietly describe a person, not the work. “Fast-paced,” “high-energy,” “people person,” “self-starter,” “thrives with little direction.” In practice, those phrases screen for social style and tolerance for chaos, not whether someone can ship a feature, close tickets, or run clean ops.
Replace personality signals with observable outcomes. If you need independence, say what “good” looks like: “Breaks down unclear requests into written questions and proposed next steps within 24 hours.” If the job involves lots of talking, specify the channel: “Runs a weekly stakeholder check-in with an agenda and notes.” Add a short “how we work” block that names routines candidates can picture, like “Work is tracked in Jira; priorities change weekly; most communication is in writing.”
You can’t list every edge case without turning the ad into a novella. Pick the top three work conditions that trip people up, and be plain about them—then let the interview test those conditions directly.
When the “standard interview” isn’t fair: picking alternatives you can actually run
Let the interview test those conditions directly, because the standard format often tests something else: speed, social timing, and comfort being judged in real time. You see it when a strong candidate freezes on a vague behavioral question, but later sends a clear written follow-up. If the job is mostly written work, deep focus, or careful problem-solving, then you want signal from those modes, not just the room.
Start with one alternative you can run without rebuilding the whole funnel. Swap one live round for a structured work sample: a short ticket triage, a debugging exercise, a writing prompt, or a prioritization scenario. Time-box it, give the same inputs to everyone, and score it with a simple rubric tied to the outcomes in the job ad. For roles heavy on collaboration, add a “pairing” task with explicit rules: shared doc, clear goal, and permission to ask clarifying questions.
Work samples take design time, reviewers need calibration, and candidates will push back if it feels like free labor. Keep it small (60–90 minutes), pay when appropriate, and publish the evaluation criteria up front. Then you’re set up for the next hard question: what happens when a candidate asks for an accommodation, or discloses why they need it?
Disclosure, accommodations, and the fear of doing it wrong
What usually happens when a candidate asks for an accommodation is a scramble: the recruiter pings legal, the hiring manager worries about “setting a precedent,” and the candidate gets silence. That delay does more damage than most accommodations ever will. Most requests are simple—extra time, written prompts, a quieter setting, a different format—and you can handle them with a repeatable path instead of a case-by-case debate.
Set one intake channel and one owner. Make the ask about the process, not the diagnosis: “What would help you show your skills in our interview steps?” Offer a short menu (written questions in advance, breaks, captions, camera optional, work sample instead of rapid-fire) and commit to a response window, like 24–48 hours. Keep notes minimal, store them tightly, and only share what the interviewer needs to run the change.
When a req is urgent, people slip back to “just do the standard interview.” Lock the workflow before offers go out—because the same clarity is what keeps a great hire from exiting in week three.
Onboarding and early feedback: preventing the ‘great hire, quick exit’ pattern
That same clarity is what keeps a great hire from exiting in week three—because week one is where “we’re flexible” turns into “figure it out.” The new hire gets a laptop and a pile of links, meetings appear with no agenda, and the first real feedback arrives only after something goes wrong. If someone does best with written expectations and steady check-ins, they can look “behind” while they’re actually trying to find the rules.
Make onboarding measurable. Give a 30-60-90 plan with examples of “good,” a written “how we decide” doc (who owns what, how priorities change), and a default communication norm (“If it’s not in the ticket, it doesn’t exist”). Schedule two short feedback points in the first two weeks with three prompts: what’s unclear, what’s blocking you, what should we change in the process.
If they can’t commit to those touchpoints, the safest move is to slow the start date rather than hire fast and lose faster.
Manager readiness is the bottleneck—what do they need before the first 1:1?
Slowing the start date is often easier than admitting the real bottleneck: the manager isn’t ready to run the first month in a predictable way. So day one becomes a social test—reading between the lines, guessing priorities, and absorbing unspoken rules—right when a new hire needs the rules written down.
Before the first 1:1, give managers a small “operating kit” they can actually use: a one-page role scorecard (top 3 outcomes, examples of good work, what “done” means), a default cadence (weekly 1:1 plus one async check-in), and a script for clarity (“Here’s what I’ll watch for in the first 30 days; here’s how to tell me you’re stuck”). Add one accommodation rule: changes are about the work and the process, not a diagnosis.
This forces managers to commit to routines and to write things down. If they won’t, no hiring tweak will hold.
A workable next 30 days: one hiring change, one workflow, one norm
If they won’t commit to routines and write things down, keep the next 30 days brutally small. Make one hiring change: replace one live “tell me about yourself” round with a 60-minute work sample scored on a two-criterion rubric tied to the role scorecard.
Put one workflow in place: an accommodations request form routed to one owner, with a 48-hour response SLA and a short menu of options (written prompts, breaks, camera optional, alternative format). The hard part is coverage—when that owner is out, requests stall—so name a backup now.
Set one norm: “No agenda, no meeting.” If a manager can’t send a two-bullet agenda and the decision needed, it becomes async by default.
