Fingerprint expands device intelligence for the AI-driven web
Artificial intelligence is changing the makeup of internet traffic, making AI identity an emerging concern for websites and applications. They are no longer interacting only with humans and traditional bots. AI assistants are retrieving content directly over HTTP, autonomous agents are navigating browsers and completing tasks on behalf of users and AI systems are increasingly connecting to enterprise tools and operational data.
For e-commerce and financial services companies, that creates a new identity problem. Blocking all automation risks disrupting legitimate AI-driven interactions, while trusting anything that claims to be an AI assistant or agent creates an opening for scraping, fraud and impersonation. FingerprintJS Inc. is expanding its platform to address that challenge through Authorized AI Agent Detection, AI Assistant Detection, its Automation Intelligence API and the Fingerprint MCP Server. Together, these products are designed to help organizations identify different forms of AI traffic, distinguish verified systems from spoofed automation, and connect AI assistants with trusted fraud and device intelligence.
The expansion is set to assist organizations in moving from conventional bot detection to governing a web increasingly shared web by humans and machines.
From bot detection to AI identity
Traditional bot management has largely been built around a binary question: Is the visitor human or automated? That distinction is not as useful when some of the automation is desirable.
An AI shopping agent comparing products for a consumer may represent legitimate purchase intent. An assistant such as ChatGPT or Gemini retrieving information may represent a new referral channel. Meanwhile, an attacker can impersonate those same systems to evade bot controls or scrape sensitive content.
Fingerprint’s Authorized AI Agent Detection addresses the first category by identifying signed AI agents operating through browsers. The company says its ecosystem can cryptographically verify agents from OpenAI, AWS AgentCore, Browserbase, Manus and Anchor Browser.
AI Assistant Detection operates at a different layer. Rather than identifying an AI system controlling a browser, it examines direct HTTP traffic from assistants like ChatGPT, Gemini and Claude.
These distinctions are critical since many existing security products rely heavily on client-side JavaScript while AI assistants often never execute it.
Instead, Fingerprint looks at the claimed user-agent, originating IP address, reverse DNS and network information published by AI providers. Those signals help establish whether traffic claiming to come from a particular assistant is actually associated with that provider.
The result is that not every bot has to be treated the same way. Verified assistant traffic might be perfectly acceptable on a public page. The same may not be true for an unverified request claiming that identity, or for an agent trying to complete a sensitive transaction.
Browserless AI changes the security perimeter
Fingerprint’s Automation Intelligence API extends that strategy beyond browser-based detection.
The API, currently available in preview, is designed to classify automated traffic without requiring client-side JavaScript and can operate at the content delivery network edge, in middleware or on the backend. Along with automation classification, it can provide context around IP and network risk, including proxy, virtual private network, Tor and geolocation signals.
AI is essentially altering where identity decisions are made.
Historically, much of digital identity and fraud prevention assumed that meaningful customer interactions would arrive through a web browser or mobile application. AI assistants increasingly bypass that layer and communicate directly with websites, APIs and backend services.
That changes what security teams need to know about incoming traffic. It is no longer just a question of whether a request is automated. Teams may also need to know which AI system is making the request, whether it is really who it claims to be and what it is trying to do.
The shift arrives as enterprises are already becoming more comfortable with autonomous systems. theCUBE Research’s 2025 AI Builder Summit research found that 55% of respondents had deployed autonomous AI agents, while 60.5% expected to do so within the next 18 months. Multi-agent systems were already deployed in 41.8% of cases, with 50.9% planning adoption.
As that adoption expands, machine identity moves from an emerging security concern toward a production architecture requirement.
MCP brings AI to the other side of fraud prevention
Fingerprint’s MCP Server approaches the AI transition from the opposite direction, allowing authorized AI assistants and agents to interact with Fingerprint’s own device intelligence.
The MCP Server, now generally available, exposes device signals, fraud events, workspace management capabilities and integrations through the Model Context Protocol. Developers can also connect compatible coding environments such as Claude Code and Cursor.
For fraud analysts, that could change the interface used to investigate suspicious activity. Instead of manually navigating dashboards and correlating device identifiers across events, an analyst could ask an AI assistant whether several suspicious accounts are connected or what changed during a sudden increase in checkout fraud. The assistant can query Fingerprint data through MCP and return an analysis. The trend shows that AI is becoming both something enterprises must identify and something enterprises increasingly use to operate their systems. That dual role creates its own governance requirements.
Access to fraud data does not have to mean access to everything. An assistant might be able to review fraud telemetry without having the ability to change rules, block an account, or take other action. All of those permissions can be handled separately, with human approval retained for decisions that require it.
There is still a gap between that kind of governance and the way AI is used today. theCUBE Research’s Agentic AI and Trust research found that only 20.2% of respondents had enterprise-wide AI deployments built on governed frameworks. By comparison, 50.7% said their organizations primarily relied on public AI tools.
The adoption curve is moving faster than the governance curve.
E-commerce and financial services face the identity question first
E-commerce and financial services are likely to be among the first industries to face the consequences of machine-mediated interaction at scale.
In commerce, AI agents could increasingly research products, compare prices and eventually transact on consumers’ behalf. A retailer may need to distinguish a legitimate shopping agent from a scraper or an automated fraud attempt without disrupting the customer experience.
Financial services raise the stakes further.
An agent interacting with a bank or fintech application may eventually assist with product selection, account management or transactions. In those environments, simply knowing that an interaction came from an AI agent will not be enough.
Organizations will need more context about the agent itself. Is it really who it claims to be? Who authorized it, and what is it allowed to do? Those questions start to put AI identity alongside authentication, fraud prevention, API security, and zero trust within the application’s security architecture.
How organizations can move forward
Companies do not need to overhaul their identity and fraud systems just because legitimate AI traffic is starting to appear. But it is worth asking how well those systems handle traffic they were never originally built to recognize.
There are five places to start:
- Look for AI traffic across customer-facing systems. Start by figuring out where assistants and agents are showing up today. This could be on a website, through an API, during login or checkout, or while accessing product information. Knowing where these interactions are happening gives teams a baseline to work from.
- Confirm who is behind the traffic. Detecting automation only answers part of the question. If an assistant or agent claims to represent a known AI service, teams also need a way to determine whether that claim is legitimate.
- Check what existing fraud and identity tools can actually see. AI activity will not always come through a browser. Teams should understand what happens when an assistant connects directly over HTTP and whether their existing tools can still capture the device, network, behavior and transaction information they rely on to assess risk.
- Give fraud teams room to use AI, with limits. Tools such as MCP can make fraud data easier to work with and may reduce some of the manual effort involved in an investigation. The important part is deciding what an AI system is allowed to see or recommend, and what it can actually change or automate.
- Plan for AI identity as a permanent requirement. Agentic commerce and machine-mediated financial interactions are still developing, but the underlying identity problem is unlikely to disappear. Policies created now should be designed to evolve as AI systems gain more autonomy.
Fingerprint’s product direction highlights how the market is heading toward more granular identification of AI traffic, verification of claimed machine identities and policy decisions based on the context of each interaction.
The larger issue goes beyond any one vendor. If AI accounts for more legitimate online activity, companies will have to accommodate it without granting malicious automation the same access.
The bottom line
The web is evolving from an environment dominated by humans and unwanted bots toward one shared by people, traditional automation, AI assistants and autonomous agents. Treating every automated interaction the same will become increasingly impractical.
The more important capability will be identifying the type of machine interacting with an application, verifying that it is what it claims to be and applying policy based on context and risk.
Fingerprint’s Authorized AI Agent Detection, AI Assistant Detection and Automation Intelligence API address different points where AI traffic enters the application stack, while its MCP Server addresses how authorized AI systems can interact with fraud and device intelligence from the inside.
For technology leaders, the issues extend beyond any single product portfolio. AI identity is becoming part of the digital trust architecture. Organizations preparing for that shift should begin by determining where assistants and autonomous agents already interact with customer-facing applications, whether existing fraud and identity systems can distinguish trusted AI traffic from impersonation attempts and which policies should apply as machines take on more actions traditionally performed by people.
As agentic commerce and AI-assisted financial interactions expand, we will move past the question of whether a visitor is human or automated and focus more on whether that visitor, human or machine, can be trusted.
Image: SiliconANGLE
Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities.
- 15M+ viewers of theCUBE videos , powering conversations across AI, cloud, cybersecurity and more
- 11.4k+ theCUBE alumni — Connect with more than 11,400 tech and business leaders shaping the future through a unique trusted-based network
Are you an AWS customer? Support SiliconANGLE financially by buying your AWS services from our Marketplace portal page and links: https://siliconangle.com/aws-marketplace/
About SiliconANGLE Media
theCUBE AIand theCUBE SuperStudios — with flagship locations in Silicon Valley and the New York Stock Exchange — SiliconANGLE Media operates at the intersection of media, technology and AI.
Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.