Charm Security

Charm has no product apart from its models. The company describes an agentic AI workforce in which agents assess risk levels, analyze alerts and deploy real time countermeasures while a scam is unfolding. Its thesis is that traditional fraud systems intervene too late because they read transactions and devices rather than the manipulation of the person. The entire product is a model reading human behavior and acting on that reading within the window of an active attack. Strip the models and there is no rule set, no case management system and no archive left behind.

The product acts on a live customer in the moment and no boundary on that action is published. Agents are described as intervening in real time to disrupt deception and manipulation while a scam is unfolding, transforming a scam in progress and deploying countermeasures. Whatever the intervention is, whether interrupting a payment, injecting a warning or escalating a call, it happens at machine speed inside a customer interaction.

No approval gate, confidence threshold or escalation path appears in public material, and neither does any analyst review or sampling regime. Nor is there any description of what the customer experiences or is told.

The metric the market rewards and the metric that protects the customer diverge. A scam prevention product is naturally measured by scams stopped. The cost that goes unmeasured is legitimate payments delayed or blocked and customers wrongly treated as compromised, which the institution absorbs as service failure rather than as fraud loss. Nothing published addresses that second metric.

There is no accuracy figure, detection rate or false positive rate, and no validation evidence or model documentation. For a system making a psychological inference, the missing account of how the inference is validated and how often it is wrong is the whole gap.

No financial institution is named anywhere. Charm was founded in 2024 and has one funding round of $8 million, led by Team8. The credentials are real: the chief executive is a third time founder with two decades in cyber intelligence, including ten years in a national signals unit. The chief technology officer led data science at Transmit Security and built fraud detection for global banks, and Team8's managing partner on the deal is a former chief executive of Israel's largest bank.

Three named relationships exist: a deepfake detection partnership with Reality Defender, a credit union member protection initiative with Curql, and a partnership with a mental health organization. One award was won at an anti scam summit. None of that is measurement. Buyer caliber and founder pedigree describe who is interested, not whether the product works. There is no customer, no volume, no detection rate and no case study.

Nothing published defines a data boundary. Scam intelligence has a strong natural pull toward pooling, because a manipulation script or a mule network seen at one institution is immediately useful at the next. Charm describes analyzing customer risk patterns and continuously adapting to evolving scam tactics.

It does not state whether patterns learned at one institution inform protection at another, whether any customer level signal crosses the boundary, or whether participation in any shared intelligence is optional. The stewardship question carries extra weight because the material being learned from is behavioral and psychological rather than transactional.

Charm infers how easily a named person can be manipulated, and no privacy account of any kind is published. The product assesses human vulnerability exposure and customer risk patterns using psychological insight, so the institution comes to hold a machine derived judgment about how susceptible an individual is to manipulation. That is an inference about cognitive and emotional state, not about a device or a transaction.

Inferring confusion, hesitation or coercion is categorically more sensitive than behavioral signals, and Charm makes that inference the product. There is no published privacy policy detail, retention schedule or data protection agreement, and no subprocessor list or deletion commitment. Nothing states whether a vulnerability assessment persists in the institution's customer record or expires, or whether the customer is told a susceptibility judgment exists about them. Nothing addresses lawful basis for profiling of this kind under European or state privacy law.

A security company has not yet published the evidence its own buyers demand of everyone else. There is no attestation, certification, trust center or dedicated security page, and no SOC report or ISO 27001 certificate is announced or offered on request. That gap is conspicuous for a company founded by cyber intelligence specialists and selling into bank vendor risk reviews, and it is the first item any financial institution's third party program would request.

One named regime is identified correctly, and it is the right one. The company states that banks face growing regulatory scrutiny and liability for authorized push payment fraud. That is the reimbursement regime that shifted the cost of scam losses onto payment firms, and it is the single change that created this market. Charm names the specific liability shift rather than gesturing at fraud regulation generally. Nothing maps the product to any other instrument.

There is no United States regime covering error resolution or unfair and deceptive practices, no elder financial exploitation reporting duties despite vulnerability being the core subject matter, and no admission process or accreditation. Charm has no license and needs none, which is the correct posture for a technology supplier.

The core mechanism is a susceptibility judgment about individual people, and nothing is published about how it is tested. Vulnerability to social engineering correlates directly with age, cognitive decline and bereavement. It also tracks social isolation, digital literacy, neurodivergence and operating in a second language.

A model built to identify who can be manipulated will concentrate its flags on older, disabled and less digitally fluent customers as a property of the technique rather than as a defect. Two failure modes follow and neither is addressed. A false positive means a customer is treated as incapable of managing their own money and subjected to friction or intervention on legitimate payments, which is paternalistic exclusion dressed as protection.

A false negative means the customer the product exists to protect is not protected. There are no demographic error rates and no accessibility analysis. Nothing states what a vulnerability score contains, how long it persists or whether it influences any other decision the institution makes about that customer, and there is no route for a person to know they carry one. The protected characteristic is not a proxy here. It is the thing being modeled.

The two harms specific to this system have no route to a remedy, and no guarantee, indemnity, falsifiable commitment or correction path is published. The reimbursement regime the company names does give a scam victim a real route, but it addresses the harm the product prevents rather than any harm the product causes. A customer carrying a vulnerability assessment is not told it exists and cannot see, correct or contest it.

A customer whose legitimate payment is interrupted because a model read manipulation into an ordinary interaction has no notice that an automated judgment was involved. Both are the institution's customers and neither has any relationship with Charm, and the gap is sharper because the judgment is about the person rather than the transaction.

One part of the chain is named in public. Charm identifies Reality Defender as the deepfake detection component embedded in its agentic workforce, announced as a partnership in March 2026, which places a specific third party model in the decision path. The rest of the chain is undisclosed. No foundation model provider, hosting arrangement or subprocessor is named, and nothing states whether customer interaction content passes to an external model provider during real time analysis.

Integration is implied but nothing is named. Protection is stated across digital, voice and in person channels, which necessarily implies integration into online banking, contact center and branch systems, and that spread would be a real differentiator if evidenced. No core banking platform, digital banking provider or contact center platform appears anywhere, and neither does any case management system or existing fraud engine. There is no developer documentation or API reference.

A buyer cannot determine what deploying this into a live branch or call center involves, which is a material question for the credit union segment the Curql initiative targets, where integration effort decides adoption.

Three jurisdictions are in play and nothing published says where anything runs. There is no hosting model, cloud provider or region selection, and no residency commitment or private deployment option. The question carries weight because real time intervention in a voice channel implies live processing of customer interaction data. The company operates across the United States and Israel while naming a United Kingdom originated liability regime as its market driver.

No pricing, packaging, basis of charge or trial route is published and every path into the product is a contact request. Nothing indicates whether charging is per protected customer, per intervention, per channel or as a platform fee, which matters for a product sold to institutions ranging from a small credit union to a national bank. Category norm rather than a specific failing.

The stated buyer set is broad and reads correctly for the problem: banks, fintechs, payment providers, and credit unions through the Curql initiative, with protection extending across digital, voice and in person channels. Published breadth describes what the platform is built to address, not who relies on it. No institution of any size or type is named as a deployment, so the coverage claim cannot be checked at a single point. Channel breadth across voice and in person would be a real differentiator if evidenced, since most scam controls work only in the digital channel.