The users of Aissist see 80%+ in resolution rate, leaving less than 20% of cases to be escalated to human teams. This resolution rate is at least 3X what is typically observed with conventional systems. The underlying force behind this accomplishment is Aissist.io’s proprietary technology based on Generative AI, known as the I.C.E (Instruction, Context, Example) engine. This innovative engine has been developed entirely from the ground up with Generative AI by the Aissist team.
No Flows, No Intents, No Rules
Over the last decade or so, the primary focus of conversational AI and chatbot developers has been on utilizing the “intent.” Identifying a user’s intent and then matching it with relevant flows, graphs or trees to drive the conversation. However, this method has shown its limitations in scalability and applicability, struggling to keep up with a wide spectrum of use cases. High-profile instances of such shortcomings can be seen in voice assistants like Siri, Alexa, and Google Assistant, which all rely on variations of intent-based frameworks. These technologies not only fall short in flexibility but also frequently fail when users deviate from their narrowly defined paths.
People have also been frequently frustrated by the chatbots which are designed with static rules and can not effectively use it to achieve the goals. Over 70% of chatbot users choose to click the “ask for human” button within 3 interactions if such a choice presents itself.
Introducing the I.C.E Engine
The I.C.E engine stands as a testament to Aissist.io’s forward-thinking approach to AI. It comprises three core components:
- Instruction: This component acts as a guide, outlining principles or procedures that enable the AI to perform complex business tasks. Notably, the setup process is quick and straightforward, typically taking about 10 minutes, in stark contrast to other systems that may require significantly longer timelines for implementation.
- Context: Unlike static flowcharts, the context component of the I.C.E engine provides dynamic information that empowers the AI to tackle a vast array of tasks in numerous expressive manners. This adds a level of flexibility and robustness that traditional systems lack.
- Example: Scalability remains a pivotal aspect of AI success, often hampered by obscure or exceptional scenarios. The I.C.E engine addresses this by learning from examples, an approach that proves more efficient than mere description, enabling the system to adeptly handle both common and rare cases.
The I.C.E engine enables the users to quickly build (easy) a robust AI solution within minutes to hours, compared with weeks even months with legacy system, and it is super easy to maintain, especially in complex cases. It is powerful – the accomplish rate across our users are 80%, with quite a number users reached 90%+. What about the rest? Aissist.io has auto-escalation technology which transfers to human team with notes and summary to empower a smooth hand-over. Moreover, the engine is equipped with capability to continuously learn with live examples to keep trimming the corners cases towards perfection.
Contact us (sales@aissist.io) for free consultation.
Aissist.io’s mission is to create the most powerful and accessible AI system for the rapidly expanding SMB market, with zero barriers to usage. For demonstrations and collaborative opportunities, interested parties are encouraged to reach out.
