The Algorithm Doesn't Speak Veteran.

The Challenge

Last week I wrote about how volume recruiting optimizes for filled, not fit. A few days later, Korn Ferry published research making the same argument from a different direction, and the timing wasn't lost on me.

One AI screening platform alone, used by 60% of Fortune 500 firms, rejects nearly 9,000 applications every half hour, over a billion applicants across a few years, without a human ever opening the file. AI use in HR nearly doubled in a year, from 26% to 43%, per SHRM.

Korn Ferry's own people are raising the flag. Karena Man, who leads their Technology and Digital practice, says these systems find candidates who precisely match a profile and miss everything else: the mid-career pivot, the candidate whose two unrelated skills only make sense together, the qualification that never shows up as a keyword. Her advice to employers: filter in before you filter out.

Bryan Ackermann, the firm's head of AI strategy, is just as direct: never let one automated decision point make the call.

That's a talent firm, not a Veteran advocacy group, independently describing the problem I'd already named: a system built to reject exactly the kind of candidate a transitioning Servicemember is.

What Gets Lost in the Filter

A Veteran's resume speaks in MOS and AFSC codes, real capability that doesn't map word-for-word to a civilian job description. “Platoon Sergeant” and “Operations Manager” describe overlapping work, but they don't look alike to an Applicant Tracking System (ATS) built for the closest string match.

The deeper failure is that translation only works aimed at a specific reader. Too military and a civilian can't follow it. Too civilian and the service disappears entirely, a resume so scrubbed it reads as a generic career with no explanation for the gaps.

An ATS can strip the jargon. It can't know what a specific hiring manager, at a specific company, needs to hear to recognize the fit. That's judgment, not matching, and judgment is the one thing a filter was never built to make.

The cost is real. Veterans are underemployed at meaningfully higher rates than civilians with the same education, not for lack of capability. More than 85% served in roles companies already hire for: engineering, logistics, electronics, healthcare, HR, transportation. The other 15% made decisions under conditions no business school can simulate.

Veterans with a bachelor's degree bring nearly three times the work experience of civilian peers with the same degree, and when the role actually fits, they're promoted earlier.

The talent is documented. The question is whether anything in the process is built to see it before an algorithm throws it away.

What the Filter Can't See

Resilience. Leadership. Adaptability. Accountability. Every hiring manager wants these, and almost none can screen for them, because stating a trait proves nothing. A filter can't tell “proven leader” from evidence of one.

They show up as proof instead. A team that performs when conditions worsen. A manager who doesn't need managing. Someone who executes when the plan changes and the mission doesn't. A team that holds itself accountable without being told to.

Then there's what an ATS is worst equipped to find at all. Team before self, demonstrated, not declared: taking the harder assignment because someone had to. A definition of finished that doesn't move until the mission is done, not the clock. Problem-solving with no manual to check, building the answer in real time with the information available. Strategic thinking that holds two altitudes at once, the six-month plan and the decision in front of you, without losing either.

None of that is a keyword. It's a pattern, visible only to someone who reads the record closely enough to see it, exactly the read an algorithm skips.

Why Always Forward VTA

A résumé filter decides in milliseconds, off a string match. Knowing whether someone will actually thrive on a specific team takes months of paying attention, not a scan. That's not a knock on the technology, it's a different kind of question entirely, one nobody entered as data in the first place.

What actually succeeds or fails in a seat. What the team is really like under pressure. What this person will need from the people around them, and what those people will need from them. None of that exists anywhere an algorithm can read it.

That's the work I do. I grew up in a military family, served as an Armor Officer, and spent 26 years in corporate leadership without stepping away from the military community. Reading both sides takes a lifetime in both worlds, not a keyword list.

Every leader I've ever respected treats a responsibility as non-negotiable: finding, and then nurturing or developing, the best teammates for the people already on the team. That's proactive, or it isn't real. I work directly with installations and their Transition Assistance Programs (TAP) to meet Servicemembers 1 to 18 months before they separate, before there's a resume for an algorithm to reject.

I'm not incentivized to fill a seat. I'm incentivized to be right about one. I don't put someone in front of you because a req is open and I need to hand you a name. I put someone in front of you because I've watched, and waited, and I'm certain they belong on your team.

A filter closes a requisition. What I'm building is a partnership, not a transaction that ends the day someone starts.

The company that wants the person their team stops describing by title and starts describing by name needs a different filter. A human one.

The relationship doesn't end at the offer letter, and it doesn't start with an algorithm either. It starts with a conversation.

Always Forward.

Sources

Korn Ferry, “Rejected, Again,” This Week in Leadership, August 12, 2026.

SHRM, cited in Korn Ferry, AI use in HR, 2024 to 2025.

Bureau of Labor Statistics via Louis, 2023 (85%+ non-combat roles stat).

LinkedIn Veteran Opportunity Report (original edition): 2.9x work experience, 39% earlier promotion.