Verifying an attribute instead of a person means checking a fact, such as a password, token, card, or voice, and then assuming the right human is behind it. Security programs show this pattern in nine places, including the service desk, MFA recovery, session lifetime, and standing third-party access. One internal question per place shows whether you have a working control or an assumption.
- Each of the nine blind spots verifies something about a person at one moment, then assumes it stays true for hours, days, or years.
- Credential abuse at any point in the breach progression still sits at 39% (Verizon, 2026 Data Breach Investigations Report, page 16), so the gaps around sign-in matter as much as sign-in itself.
- Third-party involvement reached 48% of breaches in the 2026 report’s dataset, up 60% from the prior year (Verizon, 2026 Data Breach Investigations Report, page 11).
- Presence-bound sign-in helps directly with two of the nine, and this post says plainly where it does not apply.
Why do programs that pass their audits still verify attributes instead of people?
Because most controls were designed to check a fact once, and a fact is easy to check. A password is correct, a token is valid, a voice sounds familiar. None of these confirm that the right human is using it, and none keep checking.
Credential abuse at any point in the breach progression still sits at 39% (Verizon, 2026 Data Breach Investigations Report, page 16).
The full report is listed in Sources. Attackers do not need to defeat your strongest control. They look for the place where an attribute stands in for a person and nobody is watching.
Here is the pattern to look for. In each of the nine places below, the organization verifies something about a person at one moment, then assumes the answer stays true. For each one you will find what it is, why it is missed, a question to ask inside your own organization, and a practical fix.
Which blind spots sit around the front door?
Three of the nine sit in the processes that create, restore, and remove access. Years of hardening went into the sign-in screen, and far less into the people and procedures around it.
1. The service desk
What it is. The team that resets passwords and re-enrolls MFA devices can effectively create an identity from a phone call.
Why it is missed. The desk is budgeted as a cost center and measured on handle time. It is rarely modeled as the highest-privilege identity system in the company.
Ask internally. “What does our help desk accept as proof of identity before it re-enrolls someone’s MFA, and does any of it survive a caller who sounds exactly like that employee?”
Practical fix.
- Remove knowledge-based questions and voice recognition as proof for resets.
- Confirm privileged resets through a second channel the caller did not supply, such as the manager’s directory-listed contact.
- Give agents a written “no” path so refusing never counts against their metrics.
2. MFA enrollment and recovery
What it is. The path someone uses to add a new device or recover a lost one.
Why it is missed. Hardening effort went into the sign-in step. Enrollment and recovery often rely on email links, texts, or a manager’s say-so.
Ask internally. “Walk me through what happens when an executive calls in at 11 PM having lost their phone.”
Practical fix.
- Hold recovery to the same strength as the strongest factor it replaces.
- Require two approvers for privileged accounts.
- Make temporary access expire on its own, and alert the account owner on every new device enrollment.
3. Onboarding and offboarding windows
What it is. The days when a new hire has no behavioral baseline, and the days after a contractor leaves while access remains.
Why it is missed. Policy says access ends at departure, but few teams measure how long it takes.
Ask internally. “How many days pass between a contractor’s last day and their access actually being gone, measured, not policy?”
Practical fix.
- Trigger access removal from the HR or vendor-management record, not from a ticket someone must remember to file.
- Give contractor accounts an expiry date by default.
- Report the measured gap to leadership each quarter.
Which blind spots appear after sign-in?
Two of the nine live in the hours after a successful sign-in, when the organization still treats the early check as current.
4. Session lifetime
What it is. The time between a successful authentication and the next time anything confirms who is there.
Why it is missed. Nobody owns the eight hours after authentication. Identity teams own sign-in, application teams own their own timeouts, and the gap belongs to no one.
Stolen session cookies make the gap real: SpyCloud’s 2026 Identity Exposure Report recaptured 8.6 billion of them in 2025 (SpyCloud, 2026).
Ask internally. “What is our maximum session lifetime on privileged SaaS, and what re-verifies the human mid-session?”
Practical fix.
- Set shorter session lifetimes for privileged applications than for general ones.
- Require fresh authentication before high-risk actions, such as changing payment details or exporting records.
- Bind tokens to the device where possible and revoke them when a device changes state.
- On workstations, lock the session when the person leaves, not after a fixed idle period. Our post on MFA and session hijacking covers the web-session side in more depth.
Where presence-bound sign-in fits. On a Windows, RDP, or VDI session, Proximia works like this: “Identity is verified with a live biometric at sign-in and bound to the user’s own XiFi Card. The card’s proximity keeps the session open. When it moves out of range, the workstation locks, and signing back in requires the biometric again.” It does not reach into a federated cloud application’s own session, and sessions still expire and require reauthorization. The provider of that application owns its token lifetime.
5. Physical-cyber convergence
What it is. Door access and workstation identity are separate systems, owned by different departments, with no shared record.
Why it is missed. Facilities and IT each answer their own audit. Nobody is asked to join the two.
Ask internally. “Can we produce one record showing which individual was physically at that terminal when that command was issued?”
Practical fix.
- Join door access logs and workstation sign-in logs on a common person identifier.
- Require per-person sign-in on shared stations, because a generic account breaks the record.
Where presence-bound sign-in fits. A sign-in that requires a live biometric plus the user’s own registered device ties each session to a verified individual, which gives the workstation half of that record. It applies equally to shared stations and single-user desktops. You still need to correlate it with door access data.
Where does a voice or a public record stand in for a person?
Two of the nine involve audio that attackers can now manufacture cheaply.
6. Executive digital exhaust
What it is. Earnings calls, conference keynotes, and webinars form a public library of an executive’s voice.
Why it is missed. Marketing wants executives visible, and security rarely treats that exposure as fraud input.
Ask internally. “How many hours of our CEO’s voice are publicly available, and what would stop a convincing version of it?”
Practical fix.
- Inventory public audio and video of executives who can authorize payments or access.
- Assume a convincing clone is possible, and never treat a recognized voice as an approval.
- Agree a procedure for urgent high-value requests that does not depend on recognizing anyone, and rehearse it.
7. Voice as proof, anywhere it survives
What it is. Any process where a voice on a phone counts as authentication, including IVR voiceprints, callbacks, and phone-based identity proofing.
Why it is missed. These processes were deployed when convincing impersonation was expensive. They kept working on paper after the cost collapsed.
Ask internally. “List every place in our business where a voice on a phone still counts as authentication.”
Practical fix.
- Build the list first; most organizations find more entries than expected.
- Replace voice checks with a possession factor plus a biometric on a device the person holds.
- Where a callback must stay, call a number already on file, never one supplied in the request, and add a second approver.
Which blind spots involve identities that are not your employees?
The last two are about access held by software and by other companies.
8. Non-human identities and AI agents
What it is. Service accounts, API keys, and AI agent credentials that hold standing access to production.
Why it is missed. This is the fastest-growing identity population, and it is governed by no one team. Rotation and review habits built for people do not carry over.
Ask internally. “How many service accounts and agent credentials hold standing access to production, and who attested to them this quarter?”
Practical fix.
- Build an inventory with a named human owner for every credential.
- Issue short-lived credentials and rotate the rest on a schedule.
- Limit permissions to what the job needs and review them each quarter.
Where presence-bound sign-in does not fit. Proximia does not govern non-human or AI-agent identities. This one needs a different set of controls.
9. Standing third-party access
What it is. Vendor tunnels, remote support accounts, and partner connections that stay open between uses.
Why it is missed. Vendor risk is run as an annual questionnaire, not a live inventory.
Breaches with third-party involvement rose 60% from the prior dataset, reaching 48% of total breaches (Verizon, 2026 Data Breach Investigations Report, page 11).
Ask internally. “How many vendor tunnels are open into our environment right now, and who is on the other end of each?”
Practical fix.
- Replace standing access with access that opens on request and closes on its own.
- Require named individual accounts at the vendor, not shared ones.
- Write MFA expectations into contracts and ask for evidence.
Where does presence-bound sign-in help, and where does it stop?
It helps where the question is whether a specific person is at a specific workstation, and it stops everywhere else. Said plainly, Proximia Connect does not detect deepfakes, does not stop an authorized insider acting within their access, does not govern non-human or AI-agent identities, and does not replace segmentation, EDR, logging, or backup.
What it does address is narrower. The session on a Windows, RDP, or VDI workstation stays tied to a verified person who is physically present, and it locks within seconds, at a distance and timing your organization sets, when that person leaves. Of the nine blind spots, that bears directly on session lifetime and the physical-cyber record. The other seven need process changes, inventories, and other tools, and the fixes above do not depend on any product.
Pick the three questions your team would least like to answer in front of an auditor, and start there. Our overview of zero trust security shows where these checks fit in a wider program.
Key Takeaways
- The nine blind spots share one pattern: a fact about a person is checked once, then trusted.
- Start with the service desk, MFA recovery, and any place a voice counts as proof, because they are cheap to fix and expensive to ignore.
- Presence-bound sign-in narrows the session and physical record gaps on Windows workstations; the rest of the list needs its own controls.
How Proximia Makes Identity Persistent
Proximia replaces point-in-time authentication with Persistent Identity Assuranceâ„¢, a new category of identity and access management. We continuously prove the right person is present.
- Confirm. A live biometric verifies it is you at sign-in, not just someone with a code. Images are never stored or transmitted.
- Bind. Your identity is bound through Proximia’s XiFi® technology. User and device become one.
- Persist. Your presence enables continuous access. Walk up work, walk away lock.
To walk through the session lifetime and shared workstation questions for your own environment, request a demo at proximia.com/contact-us.
Frequently Asked Questions
What does it mean to verify an attribute instead of a person?
It means checking a fact, such as a password, a token, a card, or a voice, instead of confirming the individual holding it. The fact can be copied, stolen, or imitated, so a correct answer does not prove the right person is present.
Which of the nine should we check first?
Start with the service desk and MFA recovery. They are the paths attackers use to obtain a legitimate-looking identity, and the fixes are procedural, so they cost little and take weeks rather than quarters.
Does passwordless sign-in solve all nine?
No. Passwordless sign-in removes a credential that can be phished or reused, which helps at the workstation. It does not fix help desk procedure, vendor tunnels, executive voice exposure, or non-human identities.
Sources
- Verizon, 2026 Data Breach Investigations Report, Verizon Business, 2026. https://www.verizon.com/business/resources/T343/reports/2026-dbir-data-breach-investigations-report.pdf (39% credential abuse at any point in the breach progression, page 16; 48% third-party involvement, page 11).
- SpyCloud, 2026 Annual Identity Exposure Report, SpyCloud, March 19, 2026. https://spycloud.com/blog/2026-annual-identity-exposure-report/ (8.6 billion session cookies recaptured in 2025; vendor measurement).




