DataCove AI

What is DataCove AI?

DataCove AI is a Agent-as-a-Service platform provided by the Canadian company Datacove Inc for dealing with fraud, risk, and security operations; it focuses on converting high-risk inputs into structured decisions that can be processed directly by the system.

The product that is currently available officially is the mobile anti-phishing tool DefenceNet. The enterprise platform also offers customized workflows, risk scoring, compliance markings, recommendations, and audit records; however, some of the documents related to deepfake detection are still in the preview stage.

Current product and deployment status

The FAQ still contains the descriptions from the earlier versions of the universal documentation tool, but on the current platform page, document fraud detection and advanced authenticity verification are listed as features that will be available in the near future. When selecting a product, it is necessary to rely on the current range of functions demonstrated and the list of features specified in the contract.

Main functions of DefenceNet

• Real-time link checking: It determines whether a URL appears to be part of a phishing attempt, a scam, or a malicious redirect before the user opens it, and issues a warning in such cases.
• SMS and message protection: Provides risk detection for suspicious links found in SMS messages, WhatsApp, emails, and other applications.
• Browser protection: The iOS version includes a Safari extension that reads the current URL string on the device itself in order to classify it.
• QR code security: First, scan the link contained in the QR code to assess the risks, and then it is up to the user to decide whether to open it or not.
• New fraud detection: It relies not only on known blacklist entries, but also analyzes URL structure, brand imitation, and fraudulent patterns.
• Android screen URL detection: The Accessibility Service is used to identify the links displayed on the screen; the store states that it does not record keystrokes, does not capture screen content, and does not read or save messages.
• Device-side model: The detection process takes place on the device itself, as indicated on the product page and on the dedicated privacy page; no full URLs, messages, or browsing history are sent to the server.

Main functions of the enterprise platform

• Data connection: It allows for the uploading of documents, as well as integration with enterprise systems via REST APIs or real-time data streams; local deployment can also be arranged to meet specific data storage requirements.
• Structured risk score: It returns a calibrable confidence score along with an explanation, rather than just providing a piece of free-text.
• Compliance markers: Identify violations, anomalies, and regulatory triggers in accordance with GDPR, PIPEDA, or company-specific rules.
• Structured recommendations: Output handling suggestions in the form of typed fields, so that product or operational systems can read them directly.
• Audit trail: It records the reasoning, inputs, and outputs, offers configurable retention options, and facilitates the review of decisions as well as the preparation of audits.
• Pre-demasking of the model: Personal identification information is removed before the model accesses the data, thereby reducing the extent to which sensitive data is exposed to the model layer.
• Custom workflows: By integrating generative AI, traditional machine learning, and industry-specific tools, processes are created to address particular fraud, risk, or security challenges.

Input and output

DefenceNet usage process

1. Install DefenceNet from the official app store corresponding to the device, and check the current subscription price, trial options, and permission details.
2. Enable Safari extensions on iOS, or explicitly grant the Accessibility permission for URL detection on Android.
3. When browsing web pages, checking messages, or scanning QR codes, let the application first check the link and read any risk warnings.
4. If you encounter suspicious results, do not continue to log in, make payments, or enter verification codes; instead, verify the organization and domain name through separate channels.
5. Regularly update applications and mobile systems, and revoke high-level access when it is no longer necessary.

Process for integrating enterprise workflows

1. Select clear decisions related to fraud, risk, or security; organize the input fields, desired outputs, the cost of false positives, and the rules for manual review.
2. Schedule a demonstration to review the current workflows in use, data retention, model supply chain, throughput, latency, and contract prices.
3. Test data can be introduced through document upload, REST APIs, or real-time data streams, in order to verify the effectiveness of PII anonymization.
4. Map risk scores, compliance indicators, and recommendations to the business system, and establish manual fallback procedures for cases of low confidence or anomalies.
5. Review audit logs, false positives, and missed detections; gradually increase traffic volume, and avoid using pending features in production environments ahead of time.

Price and subscription

The enterprise version of the DataCove platform does not have a fixed, publicly announced price; mobile subscriptions to DefenceNet are billed through the app stores. Region, taxes, trials, and the distribution channel affect the final amount to be paid.

Cancellation and refund

• For DefenceNet iOS, Apple is responsible for handling billing, refunds, and disputes; automatic renewal should be turned off at least 24 hours before the renewal date.
• DefenceNet Android can be managed or canceled on Google Play; the specific prices, refund policies, and renewal terms are subject to the store account settings.
• Previous version logs mentioned a one-month free trial, but to find out whether this is available in certain regions or for specific accounts, one should check the in-app purchase page.
• The old FAQ stated that enterprise subscriptions could be canceled at any time, with the cancellation taking effect at the end of the billing cycle; however, the general refund page and the details of current enterprise contracts could not be verified.

API, SDK, and open-source status

DataCove indicates that enterprise platforms use REST APIs as the primary means of connection, but the official documentation does not currently provide information on interface endpoints, authentication methods, request structures, rate limiting, error codes, or versioning strategies.

No official SDKs, GitHub repositories, open-source licenses, or example projects were identified this time. The API-first approach does not mean that the platform, DefenceNet models, or proprietary algorithms are made available under an open-source license.

Platforms and applications

Privacy and security considerations

• DefenceNet’s dedicated privacy page and Apple’s labels state that the app does not collect, store, or transmit personal data, and that the URL is deleted immediately after being analyzed in the device’s memory.
• The Google Play Data Security label indicates that app activities and device identifiers may be collected, and it states that such data cannot be deleted; this is not entirely consistent with the concept of \"zero data collection,\" and Android users should check the current version and permission settings.
• The Android Accessibility permission allows access to the URLs displayed on the screen; it is a highly sensitive permission. Even though app stores claim that they do not record keystrokes, messages, or screen content, users should still install applications only from official stores and exercise control over such permissions.
• The enterprise platform handles accounts, bills, uploaded documents, metadata, IPs, devices, and usage information; its data processing approach is different from that of DefenceNet, which is a purely device-based system.
• The platform’s privacy policy states that uploaded documents are automatically deleted within 180 days after processing, unless the user chooses to keep them; once an account is deleted, the associated data is removed within 90 days, and reports can be deleted by the user at any time.
• DataCove states that customer data is not used for training, fine-tuning, or improving any models, and PII is removed from such data before it is accessed by the models.
• The platform lists AES-256, restricted cloud storage, role-based permissions, as well as compliance with GDPR and PIPEDA; SOC 2 Type II is still in the certification process, so it cannot be considered to have a relevant report available.

Capacity limits and applicable scenarios

• DefenceNet is suitable for individuals who wish to add an extra layer of security warning before clicking on links or scanning QR codes; however, it does not guarantee 100% detection, nor does it replace antivirus software, bank anti-fraud measures, or corporate security systems.
• Enterprise platforms are suitable for banks, fintech companies, digital platforms, and security operations teams that require structured risk signals, compliance markers, and audit trails.
• Document fraud and deep fake verification are still in the experimental phase; it is not possible to plan their immediate implementation based on the promises made on future pages.
• Automatic risk scores may produce false positives or false negatives; high-risk payments, identity, account, and compliance decisions should be subject to manual review.
• API authentication, quotas, SLAs, available regions, deployment methods, and prices are not disclosed; companies need to obtain written technical and contractual documents in order to conduct an assessment.

Frequently Asked Questions

What products are currently available from DataCove AI?

DefenceNet and the customized fraud, risk, and security workflows are already available. Document fraud and deep fake verification are still marked as upcoming.

Is DefenceNet a free application?

It can be downloaded for free, but advanced protection requires automatic renewal. The iOS store in the United States lists the annual Premium version at $9.99; prices in other regions and for the Android version are indicated on the app’s internal page.

Does DefenceNet read the content of text messages?

The store’s privacy policy states that it only scans URLs, and does not record keystrokes, save message texts, or keep a history of browsing activities; given the sensitive nature of Android permissions, it is still necessary to check the data security features of the current version.

Does the enterprise platform provide APIs?

A REST API interface is provided; however, the official documentation does not reveal the endpoints, authentication methods, request structure, rate limits, or SDK – these details must be obtained as part of the project discussions.

Has DataCove AI achieved SOC 2 Type II certification?

It has not been confirmed that the process is complete; the page clearly states that authentication is still in progress.

Is DataCove AI an open-source platform?

No official GitHub account, SDK, or open-source license was identified in this case; therefore, DefenceNet’s patented models and enterprise platform cannot be considered open source.