PricingDocs

Architecture
overview

Alan AI is pioneering a new era in conversational AI. We provide cutting-edge technology and an array of  innovative components designed to manage natural language conversations, intuitively understand user requests and achieve goals.

Alan AI's approach transcends existing boundaries in conversational design, providing unparalleled flexibility and the ability to build enterprise-grade AI assistants with ease.

Alan AI Platform®

Dive into each component to see how Alan AI works.
Action Transformer
SemanticDB
LLM++
Controlled
Language Generation
Alan AI Studio Scripts

Action Transformer

The heart of the Alan AI deployment, the Action Transformer is a compact language model specifically designed to communicate with users, understand their intent and complete tasks.

Action Transformer is used for:

  • Natural Language Understanding: Uses state-of-the-art NLP and ML algorithms to comprehends user intent behind natural language requests
  • Dialog state tracking: Monitors and manages the flow of conversation, ensures coherent and contextually relevant interactions
  • Decision making based on user's request: Analyzes user requests to decide on the appropriate actions to take

The Action Transformer is equipped to perform various tasks, simplifying interactions and execution in various business domains:

  • Text generation: Communicates with external LLMs to summarize text and offer human-like text responses
  • Semantic search: Calls on the external memory to fetch precise and relevant information
  • API invocation: Capable of invoking external company APIs to achieve goals set by users
  • Code generation: Proficient in generating and executing code
  • Interactive communication: Engages in interactive dialogues, prompting users for missing information when needed

SemanticDB

Actionable AI redefines data processing with SemanticDB, a superior alternative to traditional vector databases. We have built our SemanticDB to work specifically with natural language queries, providing a more context-aware understanding and rich semantic comprehension.

SemanticDB is used for:

  • Natural language queries: Encourages users to query the database using everyday language
  • Alternative to vector databases: As a replacement for vector databases, SemanticDB offers a tailored solution that embraces natural language interaction, making it more adaptable to various applications
  • Profound contextual understanding: Focuses on semantic comprehension, offering a more nuanced understanding of context
  • Hybrid approach: Combines the best of both worlds, semantic and SQL-based search, to enable more complicated and intricate queries
  • Integration with Action Transformer: Designed to work seamlessly with the Action Transformer, aligning with the natural language focus

LLM++

Actionable AI comes with its own LLM optimized for handling semi-structured data formats, such as JSON.

These LLM are used for:

  • Data analysis: Diligently analyzes database entries and raw text input, extracting essential insights and identifying trends
  • Report generation: Crafts insightful and customized reports that support decision-making processes
  • Recommendation engine: Utilizes the analyzed data to offer context-aware recommendations, providing solutions that align with company objectives
  • Integration with different systems: Designed to work in conjunction with various platforms, enhancing data processing across different domains
  • Advanced data interpretation: Transforms complex data structures into understandable formats, making it accessible to employees and customers
  • Dynamic adaptation to business needs: Ensures relevance and applicability across diverse scenarios

Controlled Language
Generation

At Alan AI, we utilize controlled language models, paving the way for more intuitive and adaptive interactions. Controlled language models enable Alan AI to understand, respond, and engage in human-like conversations with unprecedented accuracy and flexibility.

Alan AI’s controlled language models allow for:

  • Seamless integration: Can be effortlessly embedded into your current workflow, enhancing functionality without the need for disruptive changes
  • Robust compliance and control: Adhere to industry regulations and guidelines, providing trustworthy interactions that align with legal and ethical standards
  • Personalized interactions: Tailor conversations according to individual user preferences, creating a unique and engaging user experience
  • Scalable solutions: Allow businesses of all sizes to implement conversational AI effortlessly, adapting to various domains
  • Enhanced efficiency: Significantly reduce response times and boost overall customer satisfaction

Alan AI Studio Scripts

Alan AI Studio is a robust web-based IDE enabling businesses to easily design and personalize conversational experiences to meet your specific business needs

  • Action Transformer configuration: Offers comprehensive tools to tailor the Action Transformer to specific use cases, aligning with organizational goals
  • Dataset management: Provides an easy way to define private datasets and trusted data sources, ensuring seamless integration and quality control
  • Business logic definition: Enables script developers to define complex business logic in a flexible JavaScript language, making adjustments easy, flexible, and efficient
  • ML Ops: Offers ML Ops capabilities that streamline the deployment, monitoring, and maintenance of conversational experiences
  • Customization options: Comes packed with numerous customization options, allowing businesses to design AI assistants that resonate with their brand and objectives.
  • Advanced analytics: Provides advanced analytics tools, offering insights into user interactions and areas of improvement