The Shift In Thinking: From Perimeter To Person

The zero-trust model, highlighted in NIST Special Publication 800-207, says you should never trust any user or device by default, inside or outside your network. Granting trust based on network location no longer works.

But even with zero trust, identity becomes the new choke point. If your identity system is compromised, the entire zero-trust framework collapses. Attackers know this, which is why they invest heavily in phishing and credential theft. A single database storing all your users’ secrets is a jackpot waiting to be hit.

Why Centralized Identity Is A Broken Model

We need to shift from perimeter security to identity resilience. For leaders, this requires three changes. First, stop storing sensitive identity data in a single database. Second, prioritize decentralized verification so no single system can leak every employee’s biometrics. Third, treat identity as continuous risk assessment, not a one-time check. Ask your team: If our identity provider’s data center is breached tomorrow, what would the attacker get? If the answer is more than meaningless proofs, then your architecture is not resilient.

Two schools exist: the centralized stronghold (vendors protecting their own databases) and the decentralized proof approach (no central store of sensitive data). The market is converging on the second because the first has failed repeatedly. Many identity solutions still store biometrics in a vendor’s database, turning personal information into a liability. That honeypot is broken.

A New Architecture For Identity

A better approach moves away from a centralized database to a decentralized one where the user controls their own information. This eliminates the single point of failure.

Here are four core principles any organization should evaluate when choosing an identity solution:

1. Keep The User In Control

Sensitive identity data should not live in a corporate or vendor data center. It should stay on the user’s own device in a secure digital wallet. When authentication is required, the user presents a cryptographic proof, not the raw data. A breach of your servers then yields no usable identities.

Ask any vendor: Where does the biometric template or verified credential reside at rest? If the answer is anywhere other than the user’s device, you are still building a honeypot.

2. Verify The Person, Not Just The Device

Passwordless authentication is only as good as the initial verification. A strong solution starts with verifying the person behind the identity using advanced biometrics and liveness detection, matching someone to a government-issued ID. This stops deepfakes and imposters. Insist on solutions that separate enrollment (proving you are a real human) from ongoing authentication (proving it is still you).

3. Distribute Trust To Remove Single Points Of Failure

A private blockchain or similar distributed ledger can create an immutable, auditable record of identity events without the cost of public networks. Biometric data and credentials are encrypted and distributed, making large scale breaches extremely difficult.

This model shines for financial services, healthcare or any business handling sensitive customer data. It ensures that even if an attacker compromises your corporate network, they cannot walk away with clean user identities. The only exceptions are very small organizations with no regulatory pressure, or workplaces using shared kiosk-style devices. For most modern enterprises, the trade-off is worth it.

​4. Use Continuous Risk Evaluation

Security cannot be static. An intelligent policy engine continuously evaluates risk in real time by looking at location, device, time of request and even behavioral patterns like typing speed. If a login attempt from a known user occurs from an unusual location at 3 a.m., the engine can step up security, asking for a fingerprint scan or blocking access entirely.

When evaluating identity solutions, prioritize those that separate policy decisions from identity storage. This lets you change risk rules without rebuilding your entire infrastructure.

Why This Matters For Your Organization

​Adopting a decentralized, identity-centric architecture changes the risk equation for your business. As a CISO, I recommend evaluating these specific risks:

First, assess the risk of a catastrophic identity breach. If your current provider stores biometrics in its own data center, what would happen if that center was breached? Would the attacker have everything they need to impersonate your users forever? If yes, that is unacceptable. Move to a model with no central store.

Second, evaluate your exposure to phishing and credential stuffing, the leading causes of breaches. Any system still relying on passwords is inherently fragile. Require true passwordless authentication backed by biometrics that are never transmitted or stored centrally.

Third, consider compliance. Regulations like GDPR, CCPA and HIPAA treat personal data breaches harshly. Decentralized models reduce your scope because you no longer possess sensitive identity data. Prioritize use cases like privileged access or customer-facing authentication.

Finally, look at operational efficiency. The hidden cost of password resets, lockouts and help desk tickets is enormous. A decentralized, biometric-driven approach eliminates most of that friction. Measure total cost of ownership, including help desk time and breach remediation. Modern identity often pays for itself quickly.

The Bottom Line

The old model of digital trust is finished. The perimeter has dissolved, and our approach to identity must evolve with it. Moving to a decentralized architecture that combines verified identity, biometrics and distributed storage gives you stronger security, protects user privacy and turns your identity system from the weakest link into the strongest foundation. It is time to stop guarding the castle and start securing the person.​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?