From customers to employees to agents: the path to KYA

Throughout this series, we have explored the architecture and risks of agentic AI. But solving those risks requires understanding how identity verification actually works, and why the frameworks that protected customers and employees are not built for autonomous software.

Identity verification frameworks exist because trust cannot be assumed, with each one emerging to close a specific accountability gap that prior systems left exposed.

KYC verified customers to prevent financial crime. KYE verified employees to stop internal fraud. Now, KYA verifies AI agents to govern autonomous actions.

Understanding how these frameworks connect shows us why organizations deploying AI agents need all three to work together.

KYC: How financial institutions learned to verify customer identity

KYC is a regulatory framework requiring financial institutions to verify customer identities before establishing business relationships. It was created to prevent money laundering, terrorist financing, and fraud by ensuring customers are who they claim to be.

Origins and regulatory foundation

The U.S. Financial Crimes Enforcement Network established KYC as a mandatory standard in 1990 under the Bank Secrecy Act. Financial institutions must collect and verify four core data points:

• Full legal name

• Date of birth

• Physical address

• Government-issued identification number

KYC applies to banks, credit unions, fintech platforms, insurance companies, and any business handling financial transactions. The framework solved a critical problem: anonymous actors exploiting financial systems to move illicit funds.

The gap KYC could not fill

KYC verified customers, but didn’t validate employee identities, leaving organizations vulnerable to insider threats.

The framework assumed employees were trustworthy once hired, and that assumption broke down when organizations realized insiders posed equal or greater risk than external bad actors.

In April 2025, Sentara Health disclosed a privacy breach involving a former remote employee who had improperly accessed patient records including names, birth dates, and Social Security numbers. Management raised doubts about the worker's identity during a virtual meeting, and an investigation revealed the employee may have been part of a job-sharing scam.

The company could not confirm whether the person who accessed patient data was even the individual they hired. KYC processes never touched these individuals because they were employees, not customers.

Know Your Employee: Applying verification rigor to the workforce

KYE is the practice of verifying employee identities and backgrounds both during hiring and throughout employment. It applies the same verification logic that KYC brought to customer relationships, but focuses on the workforce.

What KYE validates

KYE covers the full employee identity lifecycle, including:

• Identity documents and biometric verification during onboarding

• Background checks and employment history confirmation

• Ongoing identity verification for sensitive transactions

• Biometric authentication at high-risk access points

KYE prevents hiring fraud, stops social engineering attacks targeting help desks, and ensures access privileges remain tied to verified individuals throughout employment. Without real-time identity verification, an attacker can call the help desk, impersonate an employee, and request a credential change with nothing to stop them.

The threat is not hypothetical. In July 2025, the FBI issued updated guidance warning that North Korean IT workers continue to use fraudulent identities, AI-manipulated photos, and voice-changing software to secure remote positions at U.S. companies. Microsoft's Threat Intelligence team has also noted how these workers have embedded themselves within legitimate firms since at least 2020, generating revenue for the North Korean regime while potentially accessing proprietary systems.

KYC vs. KYE: Key differences

Both frameworks validate identity at the point of entry but govern different populations. KYC stops external bad actors from entering the system, and KYE stops internal bad actors from exploiting access.

Why traditional identity frameworks fail for AI agents

KYC and KYE share a foundational assumption: a human is always present to authorize high-risk actions. But that assumption breaks when software acts autonomously.

The core problem with applying KYC and KYE to agents

AI agents interpret prompts, make independent decisions, and execute actions at machine speed without waiting for human approval. Neither KYC nor KYE was designed for this:

• KYC cannot govern agents because agents are not customers

• KYE cannot govern agents because agents are not employees

• Both frameworks verify identity at entry, but neither validates what happens at execution

When developers leave an organization, the agents they created keep running on persistent credentials with no accountable human attached.

The identity frameworks that governed the developer do not extend to the software they left behind.

KYA: The identity framework for autonomous AI

Know Your Agent (KYA) is an identity framework that validates which AI agent is acting, under whose authority, and within what scope at the moment of execution. It extends verified identity principles to autonomous systems by requiring runtime authorization for every consequential action rather than trusting a credential issued at registration.

The fundamental shift: from registration to execution

The move from KYC and KYE to KYA represents a shift from registration-time identity checks to execution-time action validation. Where KYC and KYE verify who someone is when they enter the system, KYA validates whether a specific action should be allowed right now, under current conditions, with explicit approval from the human responsible for the agent taking it.

Policy thresholds determine when that human approval is required. Routine low-risk operations flow without friction, while consequential actions are intercepted and held for real-time human review before the tool is ever reached.

KYC vs. KYE vs. KYA: How the frameworks connect

Each framework builds on the same principle: identity must be verified at the point of consequence.

• KYC established that principle for financial transactions

• KYE applied it to workforce access

• KYA extends it to autonomous systems

Organizations deploying AI agents need all three working together: KYC ensures customers are verified, KYE ensures the employees managing agents are verified, and KYA ensures the agents themselves operate under verified human authority.

Industries where the KYC-to-KYA progression matters most

Financial services

Financial services organizations already implement KYC for customer onboarding and KYE for employee verification. Adding KYA allows them to deploy AI agents with payment system access while maintaining human oversight for every transaction that exceeds defined risk thresholds, satisfying both internal governance requirements and the human validation standards that cyber insurers are beginning to require.

Enterprise IT

Enterprise IT teams use KYE to verify workforce identity and prevent help desk fraud. KYA extends that governance to the hundreds of AI agents developers create to automate infrastructure tasks, ensuring no agent can provision resources, modify systems, or access sensitive data without a credential chain traceable back to a verified human.

Regulated industries

Organizations subject to GDPR Article 22, SOC 2, and similar frameworks must prove human review of automated decisions. The KYC-to-KYA progression provides a unified governance model that satisfies agent compliance and governance risks across customer, employee, and agent populations without requiring three separate governance architectures.

Why runtime authorization is the missing link

KYC and KYE verify identity at entry, but neither framework was built to validate what happens at execution time, and that is exactly where autonomous AI agents operate.

Runtime authorization closes this gap by shifting the security perimeter from who logged in yesterday to what is being requested right now.

By intercepting agent actions at the execution plane before tools are reached, KYA ensures every critical action traces back to a verified human authorizer, even when an agent makes the initial decision.

Without it:

• A customer with valid KYC credentials can still attempt financial fraud

• An employee with valid KYE credentials can still abuse their access

• An AI agent with a valid registration can still hallucinate and act beyond its authorized scope.

Closing thoughts

The progression from KYC to KYE to KYA is not a coincidence. It’s the natural extension of a single principle: trust must be verified at the point of consequence, not assumed from a prior check.

As AI agents gain the ability to take consequential actions autonomously, that principle becomes more important, not less.

Where 1Kosmos fits in the KYA framework

But principles require infrastructure. 1Kosmos built the Agent Identity Control Plane to put KYA into practice at the execution layer. It intercepts agent actions in real time, keeping routine operations frictionless while ensuring every critical tool call is bound to a verified human authorizer.

See how you can secure your autonomous systems without slowing them down. Visit our AI agents page or book a demo to see it in action.