What thirty-seven years in leadership taught me about hiring autistic people

I spent most of my working life on the hiring side of the table. Four years as a Navy Surface Warfare Officer, then Dell and Tech Data, then two companies I founded and ran. Thirty-seven years of deciding who got the job.

Here is the part that matters for anyone reading this from the other chair. Culture fit never made sense to my brain. I could not read the signal that tells most interviewers this person is like us, so I never felt the comfort of similarity or the discomfort of difference that quietly decides a great many hiring decisions. I hired people who could do the work and judged them on what they delivered. I was diagnosed autistic at fifty-five, which explained a great deal about why my interviews ran the way they did.

So, I am not writing this as an advocate. I am writing it as someone who made the calls, and who now understands what most interview processes are really measuring.

I’ll be direct about the limits of this. I’m not a clinician and nothing here is advice about anyone’s diagnosis or support needs. It is advice about job hunting from a person who used to be the hiring manager.

Why are 85% of autistic adults unemployed?

They are not, and that figure is inflated. The honest numbers are lower, they come from government statistics and national registers, and they measure several different things, which is exactly why the single round number keeps winning.

I see the 85% number as getting attention, which the unemployment and underemployment needs.

What the measured evidence says, with the construct named each time:

• UK graduates. Of first-degree graduates who declared autism in the 2021/22 cohort, 12 percent were unemployed fifteen months out, 43 percent were in full-time employment and 63 percent were in highly skilled work. That is AGCAS, drawing on the HESA Graduate Outcomes near-census. Autistic graduates had the lowest full-time employment of any disability group, and 12 percent is a long way from any version of 85.
• UK working-age adults. 31.4 percent of working-age disabled people whose main long-term health condition is autism are in employment. That is the Department for Work and Pensions, from the Annual Population Survey 2024/25. It is an employment rate rather than an unemployment rate, and the difference is not pedantry: “not employed” includes everyone who is not looking.
• Over time, rather than at one moment. Across eight annual waves of the Netherlands Autism Register, about forty percent of the 2,449 autistic adults enrolled were employed at any given wave, and the largest group in the analysis was characterized by stable unemployment: 1,189 of the 2,307 adults with enough data to classify. Employment there counts one hour a week of paid work, and the register is voluntary, so the sample selects itself.
• The number that gets misquoted most. Drexel’s National Autism Indicators Report found that 14 percent had a paying job in the community. That figure describes adults on state developmental-disability service caseloads, a population selected for high support needs. It is a real measurement of a specific group and it is not a general employment rate for autistic adults.

Four sources, four constructs, one direction of travel. The outcomes are genuinely bad. They are also specific, and a specific number survives the moment someone looks it up.

Is it hard for autistic people to find a job?

The hiring process is harder than the job, and there is now measured evidence for that rather than only testimony. The instruments employers use to screen candidates produce worse results for autistic applicants even when the instruments were chosen precisely to remove human bias from the decision.

Whelpley and colleagues gave a general mental ability test, a Big Five personality inventory and a customer-service situational judgment test to 107 autistic and 99 neurotypical employed adults. All three produced results unfavourable to the autistic group. On the situational judgment test the gap was large, and being autistic still predicted a lower score after controlling for mental ability, personality, gender and age.

The finding underneath that one is the more useful finding, and it is the reason I bring the paper up at all. The personality inventory’s reliabilities in the autistic sample collapsed, in some cases to near zero. A reliability that low is not a low score. It is an instrument that stopped measuring anything coherent in the population it was given to. The screen was not grading those candidates harshly. It was not grading them at all.

State the limits plainly, because I would want them stated to me: the sample came from an online panel with self-reported diagnosis, the data were collected in 2022, the situational judgment test was built for customer service work, and the authors are careful about how far the personality findings go.

What this means for you is narrow and real. A rejection from a process like that is information about the process. It is not a measurement of whether you can do the job.

Why the job-title lists are the wrong answer

Almost every page that ranks for this question is a list of occupations. Data entry, software testing, library work, lab technician. The lists are written by people who have never made a hire, and they answer a question nobody actually has, because the occupation is not the variable that decides whether a job is survivable.

I have watched the same job title work for one person and destroy another in the same building. The title was identical. The open-plan floor, the manager who gave instructions verbally and then changed them, the three unscheduled meetings a day, the performance review that graded manner rather than output: those were the variables.

So the useful question is not what job. It is what environment, and the environment is almost entirely knowable before you accept an offer.

Four environment questions worth asking in an interview

Ask these as a candidate. They are ordinary professional questions, every interviewer I ever sat opposite would have answered them, and the answers tell you more than any job description.

1. “How does work get assigned here — in writing, or in conversation?” A team that assigns work verbally and tracks it nowhere will cost you a running translation job on top of the work itself.
2. “What does the promotion criteria document look like?” If there is one, ask to see it. If there is not, promotion runs on impression, and impression runs on social fluency.
3. “Walk me through a typical Wednesday for this role.” You are listening for the number of unscheduled interruptions and who can generate them. Ask how many meetings were in last Wednesday’s calendar.
4. “How is performance reviewed, and what goes in the review?” Outcomes or manner. You will hear which one in the first two sentences, and you do not have to ask a second question about it.

None of those four require you to disclose anything. All four are questions a thoughtful candidate asks, which is the point.

What to ask for, and when

Disclosure is your decision and the trade-offs are real. What I can tell you is the sequencing I saw work from the manager’s chair.

Adjustments that are framed as working conditions get granted at offer stage far more easily than at any later point, because at offer stage the company has decided it wants you and has not yet set a precedent. Written instructions, a defined meeting load, a desk away from the main floor, a two-week notice period on travel: every one of those is a condition of work, every one is cheap, and every one is much harder to introduce in month seven than in week zero.

Ask for the thing, not the category. “Can we agree that project assignments come by email, with a short written brief?” is a request a manager can say yes to in a sentence. A general request to be accommodated puts the manager in a process they do not understand, and processes stall.

What I now know I was screening out

For years I believed the hiring systems I had designed were simply fair. They were explicit because I had no choice: written promotion criteria, structured feedback, visible decisions, because I could not run the implicit versions. The autism diagnosis at fifty-five reframed that from an accident into a design.

What I can also see now is what every conventional interview I ever sat in was really grading. Rapport. Whether the conversation felt easy. Whether the candidate’s eye contact, pace and small talk matched the interviewer’s expectations of a competent person. Every perceptive person I knew who got filtered out at that stage had been described, somewhere in the paperwork, as sensitive or difficult. The filter was not measuring capability. It was measuring similarity, and it did it very efficiently.

The critique of all of this came from autistic adults long before it had a journal citation attached, and much of what I have written here is their analysis arriving at a conclusion I reached from the opposite direction. The insight is theirs. The chair I was sitting in is mine.

If you are job hunting right now, the shortest version is this. Stop auditioning for the occupation and start interviewing the environment. The environment is what you will be living in, it is the thing the lists never mention, and it is the one variable you still control at offer stage.

Sources

• AGCAS. (2025). What Happens Next? A Report on the First Destinations of 2021/22 Disabled Graduates, 21st edition, 14 April 2025. HESA Graduate Outcomes near-census; UK first-degree graduates who declared autism, outcomes at fifteen months. This figure is verified from AGCAS written evidence to Parliament rather than from the report PDF — confirm against the report before any print use. AGCAS, What Happens Next?
• Department for Work and Pensions. (2025). The Employment of Disabled People 2025. Official Statistics, published 4 November 2025, last updated 24 March 2026. Annual Population Survey 2024/25, ages 16–64, main long-term health condition self-reported. GOV.UK, The Employment of Disabled People 2025
• Bury, S. M., Flower, R. L., Hedley, D., et al. (2024). Patterns of employment and unemployment among autistic adults. Autism, 28(9), 2322–2333. Eight annual waves of the Netherlands Autism Register, 2,449 autistic adults, mean age 42. Autism, Patterns of employment among autistic adults
• Whelpley, C. E., Fezzey, T., Batchelor, J. H., & Weekley, J. A. (2026). Using selection tests to hire autistic employees: An exploration of methods. PLoS One, 21(8), e0356539. 107 autistic and 99 neurotypical employed adults; open access, data deposited. PLoS One, Using selection tests to hire autistic employees
• Roux, A. M., Shattuck, P. T., Rast, J. E., Rava, J. A., & Anderson, K. A. National Autism Indicators Report. A.J. Drexel Autism Institute, Drexel University. The 14 percent community-employment figure describes adults on state developmental-disability service caseloads. Drexel A.J. Drexel Autism Institute, National Autism Indicators Report

Peter Allen Mann is the author of The Perception Revolution (Fast Company Press, 2026). He spent thirty-seven years in leadership, including four years as a Navy Surface Warfare Officer and two companies founded and run, and was diagnosed autistic at fifty-five. He writes for executives and system designers at SHIELD Perception.