All analysis was generated autonomously, without human review. Scores are analytical opinions drawn from the cited public sources, without hands-on testing. They are not audits, certifications, investment reports, purchasing advice, or evaluations of quality.
Reken offers an on-device security platform and its first application, Northstar, which blocks phishing, business email compromise and impersonation in email and chat, to corporations, government agencies and universities in an early-access program. Its differentiator is where the analysis runs. Small AI models Reken trains work on the user's own device rather than in a vendor's cloud. Fortune reports that rivals Abnormal and Doppel analyze communications with AI, and that most such services send data to the cloud. It also reports that Google and Apple are both said to be working on on-device phishing detection, with no product shipped from either. Reken has published no benchmark, case study or independent evaluation, so a buyer has nothing measured to weigh against Abnormal and Doppel.
| Description | Reken is an AI and cybersecurity company whose on-device Reken Platform and first application, Northstar, protect organizations against AI-enabled fraud, phishing, and social engineering, available through an Early Access Program. | [f1] |
|---|---|---|
| Founded | 2024 | [f2] |
| HQ | San Francisco, California, United States | [f3] |
| Funding | $10M total | [f4] |
| Latest funding | $10M oversubscribed seed (Jan 2024), led by Greycroft and FPV Ventures, with Firebolt, Fika, Omega, Homebrew, JAZZ | [f5] |
| Product | What it does |
|---|---|
| Reken Platform | On-device security subsystem, the Reken Private Core: locally run AI models (Model Zero), a human-verification Trust Sensor, and an API for applications to connect. In Early Access. |
| Northstar | Active AI defense against social engineering and AI-enabled cybercrime, built on the Private Core and integrated into email, direct messages, and group chat. In Early Access. |
Cyber Defense Matrix
| Identify | Protect | Detect | Respond | Recover | |
|---|---|---|---|---|---|
| Devices Workstations, servers, phones, tablets, storage, network devices, IoT infrastructure, and similar hardware. | |||||
| Applications Software, interactions, and application flows on the devices. | |||||
| Networks Connections and traffic flowing among devices and apps, plus communication paths. | |||||
| Data Content at rest, in transit, or in use across devices, apps, and networks. | |||||
| Users The people using the devices, apps, networks, and data. |
Northstar detects phishing, impersonation and other social engineering aimed at an organization's people, and blocks those messages inside email, direct messages and group chat. The Reken Platform supplies the on-device analysis behind it. These capabilities are mapped to the Cyber Defense Matrix. [f6]
How well the company can compete in its security market, scored across eight dimensions against public evidence.
| Dimension | Score | Rationale |
|---|---|---|
| Problem Clarity How precisely the company defines its problem, with evidence the problem exists at the scale claimed. | 3/5 | Reken names its buyer as corporations, government agencies and universities, and its problem as phishing, business email compromise and impersonation inside email and chat. Non-vendor sources quantify that pain, with Fortune citing $20.9 billion in FBI-logged cybercrime losses for 2025 and an MIT CSAIL event page describing a randomized experiment across more than 19,500 employees that found anti-phishing training unlikely to offer significant protective value, but that measures the category rather than a named buyer persona. [s4, s7, s12, s6, s3, s8] |
| Capability Depth How specific the technical capabilities are, with evidence beyond marketing claims such as docs and third-party validation. | 3/5 | The platform page describes a concrete architecture, the Private Core with Model Zero, a Trust Sensor and an API running local models on devices without GPUs, and Fortune and Biometric Update both describe that design in their coverage. Biometric Update also reports that Reken has published no benchmarks, customer case studies or independent evaluations, and Reken's navigation exposes no documentation or API reference, so the capability detail is on the vendor's own pages and no external validation point appears in the record. [s2, s7, s8, s3] |
| Market Timing Whether the market is ready for this product, with evidence that buyers are actively seeking solutions. | 3/5 | The enabler is wider access to generative AI, which Reken's problem page says lets unsophisticated criminals craft flawless phishing and lets advanced actors automate hyper-personalized attacks, dated by Fortune's report that the FBI added AI-related crime as a category it had not reported before in its 2025 figures. Buyer-side demand appears only as threat statistics, launch press and unnamed Fortune 500 testing rather than procurement signals, so the timing is plausible and the demand evidence is indirect. [s7, s4] |
| Team Credibility Demonstrated domain expertise with public signals such as prior exits, publications, and industry recognition. | 4/5 | SecurityWeek reports that Shuman Ghosemajumder was Google's global head of product for trust and safety and later head of AI at F5 and CTO at Shape Security, and Fortune reports F5's billion-dollar acquisition of Shape in 2020, a verifiable prior build with an exit in AI-based defense against fraud. An InfoQ podcast published in December 2025 introduced him as one of three keynote speakers at QCon AI that month, so the record carries the exit and a speaking record outside the company. [s10, s7, s11, s5] |
| GTM Proof Evidence of actual traction (customers, revenue signals, partnerships) beyond stated intentions. | 2/5 | Reken has named no customer, design partner or pilot since coming out of stealth, and Biometric Update reports that it has published no customer case studies. What the record carries is the unnamed Fortune 500 testing Reken described to Fortune, which is the unnamed-customer level rather than the named-reference level above it. [s8, s7, s6] |
| Funding Efficiency Whether funding matches go-to-market ambition, with signs of capital-efficient growth. | 3/5 | SecurityWeek reports the $10M oversubscribed seed round announced on 31 January 2024, and two years of research produced a shipped platform and a first application on 13 July 2026, so the raise is proportional to the stage and the motion with visible shipping. Reken discloses no revenue, margin or customer count, so the efficiency itself is unconfirmed. [s10, s9, s7, s2] |
| Category Clarity Whether the company creates or fits a recognizable category that buyers can quickly place in their stack. | 3/5 | Biometric Update describes Northstar as an application that identifies AI-powered fraud and social engineering attacks, and Fortune sets Reken beside Abnormal and Doppel, so buyers can place the product in a recognizable budget line without vendor coaching. Reken also frames the Private Core as a new trust layer for the Internet, a label a buyer still needs the vendor to explain, so the placement is adequate rather than clean. [s8, s7, s1, s2] |
| Incumbent Defensibility How vulnerable the core value proposition is to absorption as a feature by a platform vendor. | 2/5 | Fortune reports that Google and Apple are both said to be working on AI-powered phishing detection that would run on device, and that neither has rolled out a product yet, so two vendors already adjacent to the buyer are reported to be pursuing the placement Reken presents as its edge. What keeps it off the floor is that no incumbent ships it today, and that Reken says two years of research and development went into the on-device design. The reviewed record documents no proprietary data asset, integration burden or switching cost that would slow absorption further. [s7, s2, s8] |
Reken targets the moment an employee has to judge whether a message is real. Its problem page argues that attackers blend malicious actions into normal behavior, leaving threat detection caught between false positives and false negatives. The same page argues that phishing training does not work and that deepfake detection rarely settles the question, because legitimate content is now often AI-generated too.
Independent sources quantify both halves of that argument. Fortune reports that the FBI's Internet Crime Complaint Center logged $20.9 billion in cybercrime losses in 2025, a 26% rise, and added AI-related crime as a category it had not reported before. An MIT CSAIL event page describes a study, presented by a University of Chicago researcher, that ran an 8-month randomized experiment across more than 19,500 employees and concluded that common anti-phishing training programs are unlikely to offer significant protective value.
The buyer is an institution. The early-access form asks for a name, title, organization and work email, and Fortune reports that the program opened to corporations, government agencies and universities. Reken's product page carries an endorsement from Jim Routh, a former chief information security officer at American Express and MassMutual, which aims the pitch at the security office. [s4, s7, s12, s6, s3]
Reken describes its architecture on its own pages. The Reken Private Core is an on-device subsystem with three named components: Model Zero, which combines proprietary models Reken trains with ensemble learning to analyze incoming data, a Trust Sensor that verifies human activity on outgoing communication, and an API that lets applications connect to the core. The vendor says those models run on average devices without GPUs.
Northstar applies that core to messages as they arrive. The product page says it blocks threats instead of generating alerts for investigation, works inside email, direct messages and group chat, and needs no server integration. Fortune reports that the central research problem was making the models fast enough on standard corporate laptops to block a threat in real time.
No independent measurement of the product appears in the reviewed record. Biometric Update reports that the company has published no performance benchmarks, no customer case studies and no independent evaluations of detection accuracy. Reken's own navigation offers a problem page, a platform page, a product page and an about page, and exposes no documentation, API reference or self-serve test path. [s2, s3, s7, s8]
Reken's differentiator is where the analysis happens. Fortune reports that competitors such as Abnormal and Doppel also use AI to analyze communications, and that most of them transmit data to cloud-based services. Reken runs its models on the device instead, and Fortune reports Ghosemajumder's argument that sending data to the cloud adds breach risk and lag.
Reken is not alone in that placement. Fortune reports that Google and Apple are both said to be working on AI-powered phishing detection that would run on device, the same placement Reken presents as its edge, and that neither has rolled out a product yet. If either ships it, a buyer has one fewer reason to buy the capability separately.
The Reken Network is the one documented mechanism that could get harder to match as adoption grows. The platform page says organizations using Reken join a network without exchanging confidential data or configuring anything between them, and Fortune describes suppliers and partners joining a widening protected circle as they adopt Northstar. No organization inside that network appears in the reviewed sources, so Reken has not yet shown the adoption that argument needs. [s7, s2, s3]
Reken offers Northstar through an early-access application. The form asks for a name, title, organization and work email, and Fortune reports that the program opened to corporations, government agencies and universities. The reviewed pages offer no trial and no self-serve purchase alongside that form, which makes this an enterprise qualification motion.
The commercial proof on the record is an endorsement. Reken's product page quotes Jim Routh, a former chief information security officer at American Express, DTCC, Aetna/CVS Health and MassMutual, saying there is nothing like this available to CISOs. Reken told Fortune it had tested the product in Fortune 500 environments and would not name them.
No named buyer appears in the reviewed record. Biometric Update reports that Reken has published no customer case studies, and no pilot count, contract or revenue figure appears in the reviewed sources. Investor backing from Greycroft and FPV Ventures is the financial proof on the record, and the record carries no named adoption proof or disclosed commercial result beside it. [s6, s7, s3, s8, s9]
The team is the strongest verifiable thing about Reken. SecurityWeek reports that Shuman Ghosemajumder was Google's global head of product for trust and safety and later head of AI at F5 and CTO at Shape Security, and that Rich Griffiths was VP of product at Shape. Fortune reports that F5 acquired Shape for a billion dollars in 2020, and published a correction clarifying that Ghosemajumder was an early Shape employee rather than a co-founder. An InfoQ podcast published in December 2025 introduced him as one of three keynote speakers at QCon AI that month.
The domain fit is direct. Reken's about page says the co-founders led the creation of Shape Security products that used AI to defend large banks, airlines and federal agencies against bots and cybercrime. That is an institutional buyer overlapping the corporations, government agencies and universities Northstar now serves, though the adversary there was bots using fake accounts and stolen passwords rather than social engineering aimed at people. Ghosemajumder also founded Google's Trust and Safety product group and helped launch Gmail.
Reken now names staff beyond its two founders. Reken's about page names Sébastien Soudan as lead architect and Suraj Venkata Raman as head of design, and a published Reken patent application lists Ghosemajumder, Griffiths, Soudan, Raman and Christian Glauch as inventors. [s10, s7, s5, s13, s11]
Reken's answer to the third-party-risk question is architectural. The platform page says data is kept on the device and confidential content is not sent to third-party clouds or frontier models, which is the structural answer to what a security team would otherwise ask of a product that reads employee messages.
The probed surfaces carry no formal assurance. The trust, security and compliance paths on reken.ai return 404, the trust and security subdomains do not resolve, and a random control subdomain does not resolve either, so the missing subdomains are a real absence rather than a wildcard artifact. No SOC 2, ISO 27001 or FedRAMP reference appears on the home, platform, product, problem, about, signup or privacy pages.
A privacy policy is the one published commitment. It took effect on 2026-07-13, says Reken does not sell or rent personal information, and sets out access, deletion and correction rights for California, EU and UK residents. That is a stated commitment rather than an audited control, so a buyer evaluating a product that reads employee messages has nothing independent to review. [s2, s14, s16, s15, s1, s3, s4, s5, s6]
| Company | Relationship | Note | Compare |
|---|---|---|---|
| Abnormal AI | competes with | Fortune names Abnormal as a competitor that also uses AI to analyze communications. | N/AWe captured the evidence for these companies under different evidence-model versions (v1 vs v2), so the totals were scored under different conditions and are not directly comparable. |
| Doppel | competes with | Fortune names Doppel alongside Abnormal as a competitor that uses AI to analyze communications. | |
| adjacent | Fortune reports that Google is said to be working on AI-powered phishing detection that would run on device, the placement Reken presents as its differentiator. | N/AGoogle is scored by product line, not as a whole company, so there is no company-wide column to compare. Open its profile to compare a specific product. | |
| Apple | adjacent | Fortune reports that Apple is said to be working on AI-powered phishing detection that would run on device, and that it has not shipped a product. |
Add analyzed competitors to compare them side by side with Reken.
A closer look at the company's product strategy, measuring how defensible it is against market forces and examining the eight areas behind it.
pivot urgently
Reken offers Northstar, an on-device application that blocks phishing and impersonation in email and chat, and what a buyer can weigh today is Reken's account of its on-device models and the founders' record. Reken says two years of work produced small AI models that run on an ordinary laptop without a GPU, and Fortune reports Ghosemajumder's account that reaching that speed was the central research problem. The founders built AI bot defenses at Shape Security, which F5 bought for a billion dollars. Beyond those two the record documents no attestation, no named customer, no data history and no outside application on the Reken API, so a buyer who leaves loses only the protection. A published Reken patent application is pending, and the reviewed record shows no granted patent right.
| Dimension | Score | Rationale |
|---|---|---|
| Value Delivery Does the product sell software as the product, or judgment, trust, or accountability with software as the delivery mechanism. | 1/3 | Northstar is software an organization installs and runs on its own devices, and the reviewed record documents no managed service, no analyst layer and no accountability Reken accepts for outcomes. What is delivered is the software itself, the software-product level. |
| Switching Cost How expensive leaving is for a customer: data portability, integrations, learned workflows, network effects, regulatory data residency. | 1/3 | The reviewed record documents no data history, no accumulated policy base, no evidenced integration burden and no named deployment for a product still in early access, so nothing in it shows what a customer would reabsorb on exit. Reken's own product page says Northstar is immediately effective with no server integration required. |
| Compliance Moat Whether certifications, liability acceptance, or audit trails block an easy replacement. | 1/3 | Probes of reken.ai/trust, /security and /compliance returned 404, the trust and security subdomains did not resolve against a control subdomain that also did not resolve, and no SOC 2, ISO 27001 or FedRAMP reference appears on the home, platform, product, problem, about, signup or privacy pages. The privacy policy states data-handling commitments and no authorization or audit record that would block a funded replacement. |
| Problem Complexity Whether the product requires ML, optimization, real-time systems, or years of specialized expertise. | 3/3 | Fortune reports that Reken's central research problem was making proprietary models produce high-quality results on standard corporate laptops without GPUs, fast enough to block a threat in real time, after two years of research and development. That is machine-learning and real-time systems engineering. |
| Buyer Profile Whether buyers are SMB operators, mid-market IT teams, or regulated enterprises and governments with procurement gates. | 2/3 | Fortune reports that the early-access program opened to corporations, government agencies and universities, and Reken's product page carries an endorsement from a former chief information security officer at American Express, DTCC, Aetna/CVS Health and MassMutual. The reviewed record names no buyer of any class, and the Fortune 500 testing Reken describes is unnamed, so the evidenced buyer stops short of the regulated-enterprise class. |
| Layer Whether the product is an end-user application, a platform with application features, or infrastructure other applications depend on. | 2/3 | The Private Core is a subsystem with an API that Reken says lets any application connect to it, and Northstar is its first application. No third-party application or dependency appears in the reviewed record, and Fortune reports that opening the core to outside developers is a plan, so what is evidenced is a platform with one application rather than infrastructure others depend on. |
| Proprietary Data, Content, or IP Whether the product accumulates datasets, content licenses, or IP that a rival cannot recreate from scratch. | 1/3 | Reken names Model Zero as proprietary models it trains, and Fortune and Biometric Update both report them as small proprietary models, but the reviewed record never states what those models accumulate from. A published Reken patent application on detecting fraudulent electronic communications is pending with no grant recorded, and the Reken Network's compounding claim has no evidenced participant, so the record supports a technique in use rather than a retained asset. |
Reken addresses organizations whose employees are the target, and it now names the attacks plainly. Biometric Update lists the attacks Northstar detects as phishing, impersonation, business email compromise and other social engineering, and Reken's product page places them inside email, direct messages and group chat. Fortune reports that the early-access program opened to corporations, government agencies and universities, and the application form asks for an organization and a work email, so the served buyer is an institution.
The pitch is aimed at the security office. Reken's product page says Northstar replaces phishing training and blocks threats instead of generating alerts for investigation, and it quotes Jim Routh, a former chief information security officer at American Express, DTCC, Aetna/CVS Health and MassMutual. Both moves address the person who holds the security budget.
Adoption has not narrowed the segment yet. No named customer, design partner or vertical focus appears in the reviewed sources, and Reken told Fortune it had tested the product in Fortune 500 environments without naming them. Whether Reken concentrates on the banks its founders' prior Shape Security products defended, or sells broadly, stays unobservable while the product is in early access.
The shipped architecture is an on-device subsystem Reken calls the Private Core, with three named parts. Model Zero combines proprietary models Reken trains with ensemble learning to analyze incoming data and detect threats. A Trust Sensor uses signals to verify human activity on outgoing communication and transactions, and an API lets any application connect to the core.
The engineering claim is speed on ordinary hardware. The platform page says the models run on average devices without GPUs, and Fortune reports that proving high-quality results on standard corporate laptops, fast enough to block a threat in real time, was the central research problem. Northstar turns that into an active control, blocking threats rather than generating alerts for investigation, and the vendor says it needs no server integration.
Every capability claim traces back to Reken. Biometric Update reports that the company has published no performance benchmarks, no customer case studies and no independent evaluations of how accurately the system detects attacks. A buyer can read a concrete architecture and still cannot see how it performs outside the early-access program.
Entry runs through an application form, which makes this an enterprise qualification motion. The form asks for name, title, organization and work email, and Fortune reports that the program opened to corporations, government agencies and universities. The reviewed pages offer no trial and no self-serve purchase alongside that form.
The commercial proof on the record is an endorsement. Reken's product page quotes Jim Routh, a former chief information security officer at American Express, DTCC, Aetna/CVS Health and MassMutual, saying there is nothing like this available to CISOs. The reviewed record does not identify Routh as a customer, so his endorsement is not a deployment reference.
The launch itself was the go-to-market event. Fortune covered it as an exclusive on 13 July 2026, Biometric Update wrote it up two days later, and Reken issued its own wire release naming Greycroft, FPV Ventures and the rest of its seed investors. Those are launch-cycle signals, and no named customer, pilot count or revenue figure appears in the reviewed sources.
Reken publishes no price. The reviewed pages carry no tier, rate or charging unit, and the route to the product runs through the early-access application. That is consistent with qualifying early adopters before publishing a commercial model, and the reviewed record does not show Reken's commercial terms.
One cost argument is already on the record. The platform page says local models require zero additional AI token costs, because inference happens on the device rather than against an outside model provider. Fortune reports Ghosemajumder's related claim that sending data to the cloud adds breach risk and lag for users.
The eventual unit will show what Reken believes buyers pay for. Charging per protected user or device would frame Northstar as workforce protection, while charging applications for Private Core access through the API would frame the platform as infrastructure. The reviewed record does not show which Reken will choose.
Delivery is software that runs on the user's own device. The Private Core executes local models on the machine, and the Northstar page says protection is immediately effective with no server integration required, so the deployment the vendor describes happens on the endpoint.
Reken presents the on-device design as an operational advantage. Data stays on the device, inference happens near real time without GPUs, and the platform page says the design minimizes Internet-exposed attack surface and avoids sending confidential data to third-party clouds or frontier models. Those are architecture claims a security review would test, and no independent operational evidence appears in the reviewed record.
Blocking raises the reliability bar. Northstar blocks threats inside email and chat rather than generating alerts for investigation, so a false positive would interrupt real communication. No error rate or tuning control appears in the reviewed record, so a buyer cannot size that risk before deploying.
Reken's answer to the third-party-risk question is architectural. The platform page says data is kept on the device and confidential content is not sent to third-party clouds or frontier models, which is the structural answer to what a security team would otherwise ask of a product that reads employee messages.
The probed surfaces carry no formal assurance. The trust, security and compliance paths on reken.ai return 404, the trust and security subdomains do not resolve, and a random control subdomain does not resolve either, so the missing subdomains are a real absence rather than a wildcard artifact. No SOC 2, ISO 27001 or FedRAMP reference appears on the home, platform, product, problem, about, signup or privacy pages.
A privacy policy is the one published commitment. It took effect on 2026-07-13, says Reken does not sell or rent personal information, and sets out access, deletion and correction rights for California, EU and UK residents. That is a stated commitment rather than an audited control, so a buyer evaluating a product that reads employee messages has nothing independent to review.
Reken now presents a platform with an application on top of it. The Private Core is the subsystem, the Reken API is the vendor's stated path for any application to connect to it, and Northstar is its first application, so the platform claim has named parts and a working example.
The ecosystem around it is undocumented beyond Reken. No third-party application, integration or partner appears in the reviewed sources, and Fortune reports that letting outside developers build on the Private Core is something Reken plans. What a buyer can observe is a vendor platform serving its own first application.
The Reken Network is the compounding claim, and no member appears in the record. The platform page says organizations using Reken join the network without exchanging confidential data or configuring anything between them, and Fortune describes suppliers and partners joining a widening protected circle as they adopt Northstar. No organization inside that network appears in the reviewed sources, so the network is a stated mechanism rather than a demonstrated one.
The team is the company's strongest verifiable asset, and the record is now both first-party and independent. Reken's about page says the co-founders led the creation of Shape Security products that used AI to defend large banks, airlines and federal agencies, and that Ghosemajumder founded Google's Trust and Safety product group and helped launch Gmail. SecurityWeek independently corroborates their senior Shape roles and Ghosemajumder's Google trust-and-safety leadership, and an InfoQ podcast published in December 2025 introduced him as one of three keynote speakers at QCon AI that month.
Fortune corrected one detail of that record in public. Its story states that F5 acquired Shape Security for a billion dollars in 2020, and its correction says Ghosemajumder was an early Shape employee rather than a co-founder. SecurityWeek reports that Rich Griffiths was VP of product at Shape.
Reken now names staff beyond its two founders. Its about page names Sébastien Soudan as lead architect and Suraj Venkata Raman as head of design, and a published Reken patent application lists Ghosemajumder, Griffiths, Soudan, Raman and Christian Glauch as inventors.
| Id | Source | Tier | Accessed |
|---|---|---|---|
| f1 | Reken Early Access Program signup: Northstar and the Reken Platform offered to applying organizations | official | 2026-08-28 |
| f2 | SecurityWeek: Reken raises a 10 million dollar seed round | press | 2026-08-28 |
| f3 | Reken privacy policy: the company's contact address | official | 2026-08-28 |
| f4 | Goodwin: Reken Closes $10 Million Seed Round | press | 2026-06-21 |
| f5 | The Last Watchdog: Reken raises $10M from Greycroft | press | 2026-06-21 |
| f6 | Northstar product page: active defense against social engineering in existing channels | official | 2026-08-28 |
| Id | Source | Tier | Accessed |
|---|---|---|---|
| s1 | Reken homepage: the Reken Platform, its first product, and the Early Access Program “The Reken Platform A new AI security architecture to build a trust layer for the Internet” | official | 2026-08-28 |
| s2 | Reken Platform page: the Reken Private Core with Model Zero, the Trust Sensor and the Reken API, and the Reken Network “We spent two years of R&D building a groundbreaking on-device security subsystem that protects bidirectional communications to and from devices and accounts, with three main components:” | official | 2026-08-28 |
| s3 | Northstar product page: active defense inside email and messaging, the Jim Routh endorsement, and the Reken Network improvement claim “Our first application, Northstar: revolutionary, active AI defense against social engineering and AI-enabled cybercrime; intuitively integrated into existing channels, such as email, direct messages, and group chat.” | official | 2026-08-28 |
| s4 | Reken problem page: the vendor framing of AI-enabled cybercrime, detection limits, deepfake detection and phishing training “Cybercrime losses reported to the FBI up 26% since last year Source: FBI Internet Crime Report 2025” | official | 2026-08-28 |
| s5 | Reken About page: co-founders, the Shape Security and Google lineage, and named staff appointments “Our co-founders Shuman Ghosemajumder and Rich Griffiths led the creation of products at Shape Security that used AI to defend the world's largest banks, airlines, and federal agencies against sophisticated bots and cybercrime.” | official | 2026-08-28 |
| s6 | Reken Early Access Program signup: the application form fields “Apply to join our Early Access Program and be among the first organizations to be protected by Northstar and the Reken Platform.” | official | 2026-08-28 |
| s7 | Fortune: Reken emerges from stealth with an on-device AI shield against AI-powered scams “Google and Apple are both reported to be working on AI-powered phishing detection software that would function on device, but neither has rolled out a product yet.” | press | 2026-08-28 |
| s8 | Biometric Update: Reken launches an on-device AI platform to detect phishing and impersonation scams “The company has not yet published performance benchmarks, customer case studies, or independent evaluations showing how accurately the system detects attacks in real-world conditions.” | press | 2026-08-28 |
| s9 | PR Newswire: the Reken launch announcement issued by Reken Corporation “/PRNewswire/ -- Reken, an AI cybersecurity company, today announced its emergence from stealth, introducing a new on-device AI security software platform and the first product built on that platform, to address the worldwide problem of safety and trust in online communications.” | press | 2026-08-28 |
| s10 | SecurityWeek: Reken raises a 10 million dollar seed round “Reken, an AI-defense cybersecurity firm founded on January 1, 2024, announced an oversubscribed $10 million seed round on January 31, 2024 – but without a publicly demonstrable product. The product will be unveiled later in the year.” | press | 2026-08-28 |
| s11 | InfoQ podcast: GenAI security, deepfakes and automated social engineering with Shuman Ghosemajumder “In this episode, QCon AI New York 2025 Chair Wes Reisz speaks with Reken CEO and Google Trust & Safety founder Shuman Ghosemajumder about the erosion of digital trust.” | press | 2026-08-28 |
| s12 | MIT CSAIL event page: Grant Ho on the efficacy of phishing training in practice “Specifically, our work conducted and analyzed an 8-month randomized controlled experiment involving ten simulated phishing campaigns sent to over 19,500 employees at a large healthcare organization.” | research | 2026-08-28 |
| s13 | Google Patents: US20250342245A1, Detecting Fraudulent Electronic Communications, assigned to Reken Corp “Publication number US20250342245A1” | regulatory | 2026-08-28 |
| s15 | Reken privacy policy: data collection, sharing and California, EU and UK data rights “Reken Corporation (“Reken”) respects the privacy of visitors accessing our website and applications, and takes seriously its responsibilities regarding the privacy and security of visitor information and data.” | official | 2026-08-28 |
| s14 | Trust-path probe 2026-08-28: reken.ai/trust, /security and /compliance each return 404, no trust or security page served | official | 2026-08-28 |
| s16 | Trust-subdomain probe 2026-08-28: trust.reken.ai and security.reken.ai do not resolve, and a random control subdomain does not resolve either | official | 2026-08-28 |
| Id | Source | Tier | Accessed |
|---|---|---|---|
| s1 | Reken homepage: the Reken Platform, its first product, and the Early Access Program “The Reken Platform A new AI security architecture to build a trust layer for the Internet” | official | 2026-08-28 |
| s2 | Reken Platform page: the Reken Private Core with Model Zero, the Trust Sensor and the Reken API, and the Reken Network “We spent two years of R&D building a groundbreaking on-device security subsystem that protects bidirectional communications to and from devices and accounts, with three main components:” | official | 2026-08-28 |
| s3 | Northstar product page: active defense inside email and messaging, the Jim Routh endorsement, and the Reken Network improvement claim “Our first application, Northstar: revolutionary, active AI defense against social engineering and AI-enabled cybercrime; intuitively integrated into existing channels, such as email, direct messages, and group chat.” | official | 2026-08-28 |
| s4 | Reken problem page: the vendor framing of AI-enabled cybercrime, detection limits, deepfake detection and phishing training “Cybercrime losses reported to the FBI up 26% since last year Source: FBI Internet Crime Report 2025” | official | 2026-08-28 |
| s5 | Reken About page: co-founders, the Shape Security and Google lineage, and named staff appointments “Our co-founders Shuman Ghosemajumder and Rich Griffiths led the creation of products at Shape Security that used AI to defend the world's largest banks, airlines, and federal agencies against sophisticated bots and cybercrime.” | official | 2026-08-28 |
| s6 | Reken Early Access Program signup: the application form fields “Apply to join our Early Access Program and be among the first organizations to be protected by Northstar and the Reken Platform.” | official | 2026-08-28 |
| s7 | Fortune: Reken emerges from stealth with an on-device AI shield against AI-powered scams “Google and Apple are both reported to be working on AI-powered phishing detection software that would function on device, but neither has rolled out a product yet.” | press | 2026-08-28 |
| s8 | Biometric Update: Reken launches an on-device AI platform to detect phishing and impersonation scams “The company has not yet published performance benchmarks, customer case studies, or independent evaluations showing how accurately the system detects attacks in real-world conditions.” | press | 2026-08-28 |
| s9 | PR Newswire: the Reken launch announcement issued by Reken Corporation “/PRNewswire/ -- Reken, an AI cybersecurity company, today announced its emergence from stealth, introducing a new on-device AI security software platform and the first product built on that platform, to address the worldwide problem of safety and trust in online communications.” | press | 2026-08-28 |
| s10 | SecurityWeek: Reken raises a 10 million dollar seed round “Reken, an AI-defense cybersecurity firm founded on January 1, 2024, announced an oversubscribed $10 million seed round on January 31, 2024 – but without a publicly demonstrable product. The product will be unveiled later in the year.” | press | 2026-08-28 |
| s11 | InfoQ podcast: GenAI security, deepfakes and automated social engineering with Shuman Ghosemajumder “In this episode, QCon AI New York 2025 Chair Wes Reisz speaks with Reken CEO and Google Trust & Safety founder Shuman Ghosemajumder about the erosion of digital trust.” | press | 2026-08-28 |
| s12 | MIT CSAIL event page: Grant Ho on the efficacy of phishing training in practice “Specifically, our work conducted and analyzed an 8-month randomized controlled experiment involving ten simulated phishing campaigns sent to over 19,500 employees at a large healthcare organization.” | research | 2026-08-28 |
| s13 | Google Patents: US20250342245A1, Detecting Fraudulent Electronic Communications, assigned to Reken Corp “Publication number US20250342245A1” | regulatory | 2026-08-28 |
| s15 | Reken privacy policy: data collection, sharing and California, EU and UK data rights “Reken Corporation (“Reken”) respects the privacy of visitors accessing our website and applications, and takes seriously its responsibilities regarding the privacy and security of visitor information and data.” | official | 2026-08-28 |
| s14 | Trust-path probe 2026-08-28: reken.ai/trust, /security and /compliance each return 404, no trust or security page served | official | 2026-08-28 |
| s16 | Trust-subdomain probe 2026-08-28: trust.reken.ai and security.reken.ai do not resolve, and a random control subdomain does not resolve either | official | 2026-08-28 |
This site is an experimental research aid created by Zeltser Security Corp. All its data gathering and analysis was performed autonomously without human review, and it can contain errors of fact, interpretation, and judgment that a human reviewer might catch.
The analyses are statements of opinion, not statements of fact. Machine analysis produced the scores, summaries, and matrix placements by weighing the public sources each page cites, and reasonable people can weigh the same sources differently. Where a page states a fact, it cites the public source and the date it was checked, and the statement is only as accurate as that source. Unless a profile expressly says otherwise, the analysis involves no hands-on testing and no independent validation of any company's products or services.
Nothing here is professional, security, legal, financial, investment, or purchasing advice, and nothing here is a recommendation to invest in, do business with, or avoid any company. Inclusion of a company is not an endorsement, and absence of a company is not a judgment about it. Reading this site creates no advisory or client relationship. Verify any detail you plan to act on against the vendor's current materials.
The content is provided "as is" and "as available," with all warranties disclaimed, express or implied, including merchantability, fitness for a particular purpose, accuracy, and non-infringement. No entry is warranted to be complete, current, or correct. Companies change, vendors update their claims, sources can be wrong, and automated analysis can misread them.
To the fullest extent permitted by law, the operator, Zeltser Security Corp, is not liable for any damages that arise from using this site or relying on its content, including direct, indirect, incidental, special, and consequential damages and lost profits, even if advised that such damages were possible. If you are dissatisfied with the site or disagree with these terms, your remedy is to stop using it.
Entries link to vendor pages, press coverage, and other external sites that Zeltser Security Corp does not control and is not responsible for. A link is not an affiliation with the destination or an endorsement of it. Product and company names and trademarks are the property of their owners, used here nominatively to identify the companies described. Short quotations from cited sources appear for identification and commentary.
Use, quotation, automated retrieval, and redistribution of the content are governed by the Terms of Use at cybercompanyprofiles.com/terms, which permit personal and internal business use with attribution and prohibit republication and resale.