Transmit Security Appoints Matt Thompson CEO and Launches Agentic Trust Solutions for Banks and Merchants
Transmit Security, a provider of customer identity and access management technology, has appointed Matt Thompson as its new chief executive officer and introduced its Agentic Trust capabilities to help financial institutions and merchants manage the security challenges associated with autonomous artificial intelligence agents.
Thompson, who previously served as president and chief commercial officer at Socure, co-founded ID.me, and held identity leadership roles at Capital One and IDEMIA, will lead Transmit Security through its next phase of growth. His appointment comes as AI-powered assistants increasingly perform digital activities on behalf of consumers and businesses, creating new requirements for identity verification, fraud prevention, access management, and cybersecurity.
The company’s Agentic Trust initiative is designed to help enterprises identify AI agents interacting with their digital services, understand their behavior, and determine what actions they should be permitted to perform. The capabilities extend Transmit Security’s established focus on human identity and access management to include autonomous software agents that may initiate transactions or perform tasks on behalf of legitimate customers.
As AI adoption accelerates across banking, payments, e-commerce, and other digital industries, organizations face a growing challenge in distinguishing legitimate customer activity from automated actions. Traditional security systems have largely been designed around interactions involving human users and conventional bots. AI agents introduce another category of digital actor that may use valid credentials and perform activities that resemble normal customer behavior.
Transmit Security aims to address this challenge by combining identity orchestration, agent detection, behavioral intelligence, and continuous governance. The company’s approach is intended to help organizations adopt AI-driven services while maintaining visibility and control over sensitive transactions.
AI Agents Create New Challenges for Digital Identity Security
AI agents are increasingly capable of carrying out complex digital tasks with limited human intervention. Depending on their permissions and capabilities, these systems may help consumers open accounts, transfer funds, dispute transactions, compare products, and complete purchases.
For financial institutions and merchants, these developments could create opportunities to improve customer convenience and automate routine processes. However, they also introduce security questions that conventional identity systems may not adequately address.
When a person logs into a banking application, existing security tools can evaluate credentials, authentication factors, device information, transaction history, and other signals to determine whether access should be granted.
An AI agent acting on behalf of that person may present a different challenge. It could interact with the same application, use authorized credentials, and initiate actions that appear consistent with a legitimate customer’s behavior. Yet the institution may not have enough information to determine which agent is acting, who authorized it, or whether the requested transaction falls within the customer’s intentions.
This distinction is important because a valid credential does not automatically establish that every action performed with it is authorized or appropriate.
An AI agent may be legitimate but overprivileged, misconfigured, compromised, or manipulated into performing an unintended task. Similarly, an automated system could initiate a high-value transaction without sufficient safeguards to confirm that the action reflects the customer’s intent.
Transmit Security argues that enterprises need a way to identify these interactions and apply controls based on the agent’s identity, purpose, authorization, and risk level.
The company’s Agentic Trust initiative is designed to provide that additional layer of visibility and governance, helping organizations assess AI-driven activity without treating every automated interaction as either inherently safe or automatically malicious.
Agentic Trust Extends Identity and Access Management
Transmit Security’s Agentic Trust capabilities focus on helping organizations understand and manage interactions involving autonomous AI agents.
The framework follows four principal stages: detecting agent traffic, classifying the agents and their activities, investigating their capabilities, and governing their access to applications and transactions.
The first stage involves identifying AI agents operating across digital channels. Some agents may clearly identify themselves, while others may interact with websites and applications without explicitly declaring that they are automated systems.
The second stage focuses on determining which agent is involved and what it appears to be attempting to accomplish. This can include recognizing leading AI assistants and custom automation frameworks, as well as interpreting the intended activity.
The third stage supports investigation into agent behavior and capabilities. By examining available information about an agent, organizations can better understand its functionality and evaluate whether its activities are consistent with permitted use cases.
The final stage involves establishing and enforcing access policies. Rather than relying exclusively on a simple decision to allow or block an interaction, enterprises can apply different levels of permission according to the circumstances.
For example, an organization could allow an agent to retrieve account information while restricting its ability to initiate a payment. A more sensitive transaction could be held until the customer approves it, while an agent that presents unacceptable risks could be blocked.
This approach is intended to provide a more flexible security model for environments where AI agents may perform different tasks on behalf of the same customer.
Detecting and Classifying AI Agent Activity
The ability to detect automated activity is a foundational part of Transmit Security’s approach.
Digital businesses already use security tools to identify bots, prevent automated abuse, and protect online services. However, AI agents can operate differently from conventional bots, particularly when they use sophisticated reasoning capabilities, interact dynamically with applications, and perform tasks that resemble ordinary customer journeys.
As a result, organizations may need additional context to distinguish legitimate AI-driven activity from unauthorized automation.
Transmit Security’s Agentic Trust capabilities are designed to identify agent traffic across online channels, including interactions in which an agent does not explicitly identify itself.
After detection, classification can help enterprises determine which agent is involved and infer the purpose of its activity. This information can support decisions about whether an interaction should proceed, require additional verification, or be restricted.
For a bank, for example, an AI assistant retrieving publicly available product information presents a different risk from an agent attempting to transfer funds between accounts. Similarly, a retail merchant may have different requirements for an agent searching for products, placing an order, changing delivery details, or initiating a refund.
Understanding these differences can help organizations establish policies that reflect the sensitivity of each activity.
Classification alone does not establish that an agent is trustworthy. Its identity, authorization, behavior, and requested actions must also be evaluated in context. However, reliable classification can provide an important starting point for determining how an agent should interact with enterprise systems.
Investigating Agents Before Granting Access
Transmit Security also emphasizes the importance of investigating agent activity and understanding an agent’s capabilities before deciding what permissions it should receive.
As businesses adopt AI tools from different providers, they may encounter a diverse ecosystem of consumer assistants, enterprise agents, and customized automation systems. Each may have different capabilities, operating methods, and access requirements.
An agent designed to summarize financial information may not require the same permissions as one capable of initiating transactions or modifying customer records.
Organizations therefore need to understand what an agent can do and whether those capabilities align with its intended role.
Transmit Security’s approach includes tools and research intended to help enterprises investigate agents and assess their capabilities. This information can support risk-based decisions about the level of access an agent should receive.
Such assessments are particularly relevant in financial services, where applications may contain sensitive personal information and support transactions with direct financial consequences.
Banks and payment providers must consider not only whether an agent can access a system but also whether it should be allowed to perform a particular action under specific circumstances.
Investigating agent behavior can help organizations establish more precise policies, although effective security also requires ongoing monitoring, appropriate authentication, secure system design, and clear accountability for actions performed through AI systems.
Governing Agent Permissions and Sensitive Transactions
The governance stage is designed to give enterprises control over how AI agents interact with applications and digital services.
Transmit Security’s framework supports scoped and time-limited permissions, allowing organizations to determine which activities an agent may perform and for how long.
This is relevant because authorization should not necessarily be treated as an unlimited grant of access. An agent may need permission to perform a specific task without requiring broader control over a customer’s account.
For instance, a customer might authorize an AI assistant to review transaction history or prepare a payment. That authorization does not necessarily mean the assistant should be able to execute every transaction, change account settings, or access unrelated financial information.
Scoped permissions can help restrict access to the functions required for a particular task. Time limits can further reduce the period during which an authorization remains valid.
Transmit Security’s Agentic Trust capabilities are designed to support several possible policy outcomes. Enterprises can allow an agent to proceed, limit its activity, hold a risky action until the customer provides approval, or stop the interaction entirely.
These controls could be particularly useful for high-impact activities such as fund transfers, account changes, purchases, and disputes.
A bank, for example, might permit an authorized agent to collect information and prepare a transfer while requiring the customer to approve the final transaction. A merchant could allow an agent to browse products but require additional verification before processing a purchase or issuing a refund.
By applying controls according to the activity and its associated risk, organizations may be able to support useful AI automation without granting agents unrestricted access.
Leadership Focuses on the Future of Digital Trust
Matt Thompson said AI agents are changing how consumers interact with financial institutions and merchants, creating new challenges at the intersection of identity, fraud prevention, and cybersecurity.
He emphasized that businesses need to determine who or what is accessing their applications, understand the risks involved, and control the actions agents are permitted to perform.
Thompson also highlighted the limitations of systems designed primarily for human interactions, noting that organizations need greater visibility as consumers increasingly rely on AI agents to act on their behalf.
His appointment reflects Transmit Security’s intention to expand beyond traditional human identity management and address the security requirements emerging from autonomous AI activity.
Mickey Boodaei, co-founder of Transmit Security, said the company sees a significant market opportunity as enterprises seek ways to manage AI agents without having to choose between allowing all automated activity and blocking it entirely.
Boodaei said Thompson’s experience and leadership make him well suited to guide the company’s next phase of growth.
The leadership transition comes as organizations across financial services and commerce evaluate how to incorporate AI into customer-facing processes while maintaining security, compliance, and operational control.
Industry Perspectives Highlight the Need for Agent Governance
Travis Jarae, chief executive officer of Liminal, said organizations need to understand who AI agents represent, what activities they perform, and which actions they are authorized to take.
Jarae noted that Thompson’s experience building identity businesses, combined with Transmit Security’s background in identity orchestration, positions the company to address the emerging requirements for visibility and governance over agent activity.
His comments reflect a broader challenge for businesses adopting AI: determining how to extend existing trust and authorization models to software that can act with a degree of autonomy.
Traditional identity management systems help organizations verify users and control access to applications. However, agent-based interactions can introduce additional questions about delegation, the relationship between an agent and its human principal, and the boundaries of permitted activity.
Enterprises may need to establish who authorized an agent, which resources it can access, what actions require human approval, and how to review activity after a transaction has taken place.
These questions become especially important when agents operate across multiple services or interact with systems that process financial transactions and sensitive customer data.
Thompson Brings Identity and Financial Technology Experience
Thompson brings experience across identity verification, financial services, and digital identity businesses.
Most recently, he served as president and chief commercial officer at Socure, where the company said he helped grow the business by more than 500% over five years. He also co-founded ID.me and served as its president, contributing to the development of a digital identity business.
His previous experience includes identity initiatives at Capital One and IDEMIA, giving him exposure to identity requirements across financial services and technology.
Thompson’s leadership experience was also shaped by more than a decade of service in the U.S. Army, including four combat deployments and service as a Special Operations Officer.
His background in building identity businesses and working across complex operational environments is expected to inform Transmit Security’s strategy as it expands into agent identity and governance.
The company has also received industry recognition for its work in identity and access management. It was named a recipient of the 2026 Morgan Stanley Innovation Award for its work with Morgan Stanley in this area.
That recognition provides context for Transmit Security’s existing identity management capabilities as it seeks to extend its platform to new forms of digital interaction.
Implications for Banks, Payments, and Merchants
The emergence of AI agents could change how customers interact with financial products, online stores, and digital services. Consumers may increasingly delegate tasks such as researching financial products, managing purchases, organizing transactions, and handling routine service requests to AI-powered assistants.
For banks, this could create opportunities to simplify customer journeys and automate selected financial activities. For merchants, AI agents may offer new ways for customers to discover products and complete purchases.
However, these opportunities depend on the ability to maintain appropriate controls over access and transactions.
Financial institutions must account for risks involving unauthorized transfers, account misuse, fraud, data exposure, and mistaken or manipulated agent actions. Merchants may face related challenges involving unauthorized purchases, account changes, refunds, and misuse of customer credentials.
Agentic Trust is intended to help address these challenges by giving organizations additional tools to identify agents, assess their behavior, and govern their access.
The effectiveness of such solutions will depend on factors including detection accuracy, the quality of available signals, integration with existing security systems, and the clarity of authorization policies. Enterprises will also need to maintain appropriate human oversight for sensitive actions and establish processes for reviewing unusual activity.
Advancing Trust in an AI-Driven Digital Economy
Transmit Security’s appointment of Matt Thompson and introduction of Agentic Trust mark a strategic expansion of its focus on digital identity and access management.
As AI agents take on more tasks on behalf of consumers and businesses, organizations will need to understand not only who is accessing their services but also which software agents are acting, whose authority they represent, and what those agents are allowed to do.
By combining detection, classification, investigation, and permission governance, Transmit Security aims to help financial institutions and merchants manage these interactions while supporting the responsible adoption of AI-powered services.
The company’s strategy centers on extending digital trust beyond human users to include autonomous agents operating within defined boundaries.
For banks, payment providers, and merchants, the challenge will be to balance the convenience and efficiency of AI-driven interactions with security, accountability, and customer control. As agent-based commerce develops, the ability to verify delegated authority and manage agent permissions could become an increasingly important component of digital identity infrastructure.
Source link: https://www.businesswire.com
