Yes, the Kore.ai Bots Platform is purpose-built to be a secure yet flexible enterprise-ready, API-centric platform with an execution framework to orchestrate interaction services. It is equipped to accommodate standard and custom integrations to nearly all existing systems to fit your unique workflows and requirements. The Platform uses an API request chaining framework for bots to push or request data from your websites, systems, and web applications using Webhooks, RSS, or Web Services. The Platform includes an execution framework to orchestrate five different high-level types of interactions between users, bots, systems and even “things” like sensors. We call these “tasks.” The following task types are supported by the Kore.ai Bots Platform execution framework:
- Alerts – Alert tasks deliver timely, relevant, and personalized information from enterprise systems to your customers or employees by polling the relevant service in real-time or according to a user-defined schedule. These alerts contain detailed information, such as name, ID, notes, amount, and more, in a variety of formats for each channel. The framework supports scheduled and filtered alerts with multiple field types for alert filters including textbox, textarea, static and dynamic dropdown, checkbox and dynamic checkbox, and type ahead.
- Actions – Action tasks collect, modify, and post information in systems of record, which eliminate repetitive, time-consuming steps or form-based data entry that customers and employees commonly perform. Action tasks can take information from people either as “long-form” requests or through a guided action.
- Information – Information tasks look up data or pull reports and quickly return easy-to-consume results that are convenient for customers and employees. Users can identify the specific parameters or filters for the information delivered – such as quantity of results – allowing bots to deliver information in a way that’s simple for users and other recipients to scan, process, and reuse. The Platform supports multiple formats for the delivery of information including a table, fixed-column table, card layout with group-by, shopping cart layout with actions and more.
- Knowledge/FAQ – Knowledge tasks take user’s questions and query a pre-defined set of information to rapidly find the right answers to what they asked. Ontology mapping based on machine learning algorithms get applied to maximize answer accuracy. Knowledge tasks can be designed to find answers from a variety of knowledge sources including FAQ databases, predetermined websites, and documents in Word, PDF, and other formats. When a user queries a chatbot to complete a task, the Platform automatically estimates the probability of the correctness of the match it identifies, called confidence, using a range of 1 to 100. For knowledge tasks, developers can set the confidence threshold, in increments of 10%, to determine the minimum acceptable level. For knowledge tasks, developers can determine confidence for the following actions: not matching a task, matching a task or multiple matches for a task. In this case a list of choices is provided to the user.
- Dialog – Dialog tasks are advanced tasks that developers design with logic-driven business processes and pre-established business workflows. See the section above for more information.