The Flight to Quality: Why Google and McKinsey Are Bringing Back In-Person Interviews
A pattern that hiring-press has started calling the "flight to quality" is visible in 2026 enterprise hiring practice: Google, McKinsey, and a growing roster of tier-one employers are explicitly reintroducing mandatory in-person final-round interviews, particularly for technical and senior-consultant hires. The rationale, as disclosed in various recruiting-operations posts and conference talks, is direct: the rise of AI interview fraud — deepfake video, voice-cloned proxies, real-time coaching tools — has made remote-only vetting insufficient for high-stakes hires.
This is notable for two reasons. First, it's the first meaningful reversal of the remote-first hiring trend that dominated 2020–2024. Second, it's an enterprise-scale acknowledgement that the current trust stack for remote AI-era hiring is broken. Both matter for how employers should think about the next 24 months.
What's actually happening
Public signals from 2026:
- Google reinstated mandatory in-person interviews for SWE candidates beyond L5, explicitly citing deepfake incidents in remote rounds.
- McKinsey moved final-round case interviews back to in-person for all associate-track consultants globally.
- Several US tier-one tech companies (names vary by source) have reintroduced "one in-person round minimum" policies.
- A wave of smaller employers — including scale-ups that were 100% remote-interview pre-2025 — are adding at least one in-person touchpoint for senior roles.
Anecdotally, recruiting ops leaders at these firms describe catching active deepfake incidents monthly, and multiple firms describe blocking candidate-side coaching tools (real-time interview assistants, voice modulators) as an explicit security task.
Why this is a signal, not a solution
The in-person move catches a real problem but doesn't fix it at scale. Three reasons:
1. In-person rounds don't scale with modern applicant volume. When a role attracts 95 applications and 40% are partially AI-augmented, the funnel logistics break down long before in-person. The defence has to work earlier.
2. Travel imposes selection bias. Mandatory in-person interviews disadvantage candidates by geography, caring responsibilities, disability, and financial flexibility. This is a DEI regression most organisations have spent years trying to reverse.
3. Most of the fraud happens pre-interview. The flagged cases are what gets caught at final stage; the unflagged majority — inflated CVs, AI-drafted application content, ghost-written coding tests — have already moved through the funnel before anyone gets to a video call.
The in-person move is therefore a last-line defence, not a primary defence. It's the enterprise-scale acknowledgement that trust infrastructure is missing, but the actual infrastructure still needs to be built.
What durable trust infrastructure looks like
If in-person is the patch and not the solution, what's the solution? Three layers that together scale in a way in-person rounds can't:
Layer 1: Candidate identity verification at application
Cryptographic identity at submission time — not at interview stage. Verifiable Credentials (W3C VCs) are the standard; the practical implementations today are LinkedIn's identity verification, government-ID-verification vendors (Persona, Onfido, Stripe Identity), and specialist hiring-fraud-detection firms. None are universal yet.
The right state: candidates carry a verified identity that transmits across hiring systems. Fraudsters can't easily forge it because the verification path includes hardened steps (biometrics, liveness, government ID, cross-reference). Employers accept the verified identity rather than re-verifying at every stage.
Layer 2: Employer verifiability (the other side)
The half of the problem that gets discussed less. Fake employer listings, recruitment scams, AI-drafted "hiring manager" profiles, and ghost jobs form the mirror-image trust problem to candidate fraud. If employers become verifiably real — with schema, attestation history, cross-referenced identity, DMARC/BIMI, and dated audit trails — fraudulent employer theatre becomes expensive to fake.
This is the layer OpenRole operates in — the employer-side attestation stack via the audit, badge, and pixel. Not because it's the whole answer, but because it's the specific piece that the candidate-side and identity-verification vendors aren't building and that won't get built by an incumbent ATS.
Layer 3: Content attestation
The newest layer. In a world of AI-generated CVs and AI-coached interviews, "did the candidate actually write this?" becomes a first-class question. Provenance tools (like C2PA for media content) are being adapted to knowledge-work artefacts. Some organisations are requiring signed code commits or time-stamped writing samples produced during supervised sessions.
All three layers together are what makes remote, high-volume, AI-era hiring tractable again. None of them individually is sufficient.
The economics of the flight to quality
There's a second-order effect worth noticing. When tier-one employers reintroduce in-person rounds, they signal to the market: "we take fraud seriously." This becomes a candidate-side signal — strong candidates who are worried about being passed over in fraud-heavy automated funnels prefer employers with rigorous final-round vetting, because those employers are more likely to correctly evaluate them.
Conversely, employers stuck with fraud-heavy, low-rigor remote-only funnels increasingly attract lower-quality applicant pools (the real high-quality candidates route elsewhere, and the fraud-prone filter rewards bots over people). It's an adverse-selection spiral.
The employers who ride this well are the ones who communicate their rigor — not just "we do in-person" but "here's our anti-fraud stack, here are our verifiable credentials, here's why applying here means being evaluated fairly." That communication requires — you guessed it — verifiable employer content that AI systems can represent accurately to candidates.
What this means for mid-market employers
Most UK mid-market employers can't match Google's ability to fly candidates across the Atlantic for final rounds. The flight-to-quality move isn't directly copyable. But the principle is: be clearly distinguishable from fraud-prone hiring through structural investments that signal trust.
The practical playbook:
- Publish a public anti-fraud stance. A page on your careers site naming your verification steps (ID check, reference policy, role-specific safeguards). Candidates read this; so do AI systems.
- Ship schema. Complete
OrganizationandJobPostingschema with cross-references to Companies House and LinkedIn. Makes your listings distinguishable from fake ones at the machine level. - Build attestation history. Monthly AI visibility audits archived over 12 months are a trust-capital investment. Ghost jobs have no history.
- Lock down the sending domain. DMARC at p=reject, BIMI with VMC. Candidates' email clients will mark spoofed recruiter emails claiming to be you.
- One supervised stage minimum for senior hires. If in-person is impossible, a supervised video round with ID verification at the top does most of the work — and most of the adverse selection can be avoided if scheduled flexibly.
- Communicate the rigor. Candidates should find it easy to understand why your process is trustworthy.
The longer arc
The 2026 flight to quality is the mid-point of a cycle that starts with "remote hiring scales!" (2020–2022), hits fraud saturation (2024–2026), and resolves into "remote hiring scales once the trust infrastructure exists" (2027+).
In-person interviews won't fully replace remote-first hiring. They'll become the fallback for the cases where trust infrastructure is weakest — senior, high-stakes, specialist roles — while the bulk of hiring continues remotely on top of improving verification primitives.
The employers that benefit most through this transition are the ones who invest in verifiable identity, structured data, and dated attestation now. They're the ones the market labels "clearly real and clearly rigorous" — and that label routes disproportionate high-quality applicant volume their way while fraud migrates to easier targets.
The flight to quality is a category-validation moment for the trust-infrastructure thesis. The companies reintroducing in-person rounds are signalling the gap; the ones building the software layer to close that gap are what the rest of the market will eventually run on.
Frequently Asked Questions
Q: Are all companies reintroducing in-person interviews?
A: No. The move is concentrated at tier-one firms with high applicant volumes and high-stakes roles. Mid-market employers are largely continuing remote-first but adding verification steps at the top of the funnel.
Q: Does in-person actually prevent deepfake fraud?
A: For the final round, yes — showing up physically is hard to fake. But most of the fraud enters the funnel earlier (fake CVs, AI-coached earlier rounds), so in-person final is a necessary but insufficient defence.
Q: Is this a temporary reaction or a permanent shift?
A: Temporary-to-partial. Expect in-person to remain for highest-stakes hires permanently, but the mid-stakes roles will shift back to remote as verification infrastructure matures — probably 18–36 months.
Q: What's the DEI implication?
A: Negative, absent mitigation. Mandatory in-person rounds disadvantage geographically distant, caregiver-heavy, and disability-affected candidates. Employers adding in-person stages should explicitly fund travel (including time off for caregivers), offer alternative supervised-remote paths where feasible, and monitor representation metrics closely.
Q: Can mid-market employers replicate any of this?
A: Selectively. The directly usable moves: publish an anti-fraud stance, ship verifiable schema, build attestation history, lock down DMARC/BIMI, add one supervised stage for senior hires. All are cheaper than reintroducing travel-funded in-person rounds and collectively close most of the same trust gap.
Q: What's OpenRole's role in this?
A: OpenRole operates in the employer-side verifiability layer — machine-readable, time-stamped, cross-AI attestation of what public AI systems say about an employer. It's one specific piece of the broader trust stack the flight-to-quality cycle is validating the need for.
Start building your attestation history — monthly audits, archived with signed dates, are the trust capital that separates real employers from fraud theatre.
Related reading: