How it works
KONTXT uses a combination of machine learning, business rules (policies), and human analysis to deliver comprehensive mobile messaging spam/fraud/classification detection. It works with short-code A2P, 10-digit long code A2P, and traditional P2P senders. It’s also syntax-agnostic, and as such can manage SMS, MMS and RCS messages – any message traffic that traverses a network.
KONTXT inspects messages crossing your network and examines markers in common messages types like P2P, A2P, spam, direct REST/gRPC API, even when metadata is incomplete or misrepresented, or in cases where messages are not vetted by service provider or carrier. The sheer volume and complexity of messages requires an advanced machine learning approach to reliably identify message types.
KONTXT uses an artificial intelligence technique called guided machine learning, combining advanced machine learning algorithms with just the right amount of human curation to build classification models far more accurate than either machines or human beings alone. KONTXT crawls, analyzes and processes billions of messaging data points using NLP (natural language processing) and deep learning techniques to teach its systems to understand the true nature and origin of messages.
KONTXT can assign messages with shared patterns to categories, or classes which are meaningful to your business, particularly for A2P senders with different price sensitivity and delivery expectations. Examples include two-factor authentication, appointment notifications, shipping confirmations, and bulk promotional, to name just a few.
Antispam and Fraud Protection
Spam and fraudulent messages have historically always been a challenge, but the threat is growing with the proliferation of A2P messages and the ever-increasing complexity of enterprises and inter-carrier relationships. World-class anti-spam capabilities are included as part of Kontxt’s next generation approach to identifying and classifying messages. Kontxt can understand phishing, for example, as simply one additional instance of the classification approach that it applies to all messages. Best of all, Kontxt’s self-learning algorithms allow its threat assessment techniques to evolve with changes in message traffic, again with just the right amount of human curation.
Kontxt offers detailed reporting on traffic trends to answer questions like: Of the messages coming across my network: how many are valid or spam? What are the types of messages and use cases being sent? What senders account for the most messages? What routes are used? Many of these insights can only be generated using the advanced classification technology in Kontxt.
Policy controls are critical to managing traffic and maximizing revenue. Kontxt lets you apply business rules to control delivery, blocking, throughput, and pricing. Vetted senders or carrier partners might be permitted faster delivery or increased message throughput, for example. Grey route traffic can be blocked, throttled, or queued for manual inspection. High-value senders can be offered fast-track delivery or delivery confirmation services as part of premium product tiers.