3 types of tasks suitable for Generative AI Applications

Let's dive into the three types of tasks that Large Language Models (LLMs) can perform: augmented, transactional, and autonomous.

1. Augmented Tasks

Definition: Augmented tasks involve using LLMs to enhance or augment human capabilities. These tasks require human input and the model acts as a powerful assistant, providing suggestions, generating content, or summarizing information to aid human decision-making.

Examples:

• Content Creation: An LLM can assist writers by generating text based on a given topic, creating outlines, or suggesting improvements to existing content. For instance, a content creator might ask an LLM to draft an introduction for a blog post on renewable energy.
• Code Assistance: Developers can use LLMs like GitHub Copilot to get code suggestions, debug code, or understand complex algorithms. For example, a developer might use an LLM to generate boilerplate code for a web application.
• Customer Support: Augmented by LLMs, customer service representatives can respond to inquiries more efficiently. The model can suggest responses based on previous interactions, reducing response time and improving consistency.

2. Transactional Tasks

Definition: Transactional tasks involve using LLMs to carry out specific, often repetitive, actions based on predefined rules or user inputs. These tasks typically involve structured interactions where the model processes requests and returns precise, often factual, responses.

Examples:

• Scheduling Meetings: An LLM can handle the scheduling of meetings by understanding user availability and preferences, then coordinating with others to set up the best time.
• Answering FAQs: An LLM can be trained to answer frequently asked questions on a website or app, providing users with quick and accurate information based on a predefined knowledge base.
• Data Entry and Retrieval: In a database application, an LLM can help users retrieve specific information or enter data correctly by understanding natural language queries and translating them into SQL queries. For instance, a user might ask, "Show me all sales from last quarter," and the LLM converts this to an appropriate SQL query to fetch the data.

3. Autonomous Tasks

Definition: Autonomous tasks involve using LLMs to perform complex operations with minimal human intervention. These tasks typically require the model to make decisions, adapt to new information, and execute actions independently based on its training and inputs.

Examples:

• Autonomous Agents: An LLM can be used to create autonomous virtual agents that perform tasks such as market analysis, trading, or managing smart home systems. For instance, a virtual financial advisor could analyze market trends and execute trades based on predefined strategies and real-time data.
• Workflow Automation: LLMs can automate entire workflows, such as processing and responding to emails, managing tasks, or even handling parts of the recruitment process by screening resumes and scheduling interviews without human intervention.
• Content Moderation: In social media platforms, LLMs can autonomously moderate content by identifying and removing inappropriate posts, comments, or spam based on community guidelines and evolving patterns of misuse.

In each of these categories, the role of the LLM ranges from assisting humans to fully taking over tasks, demonstrating the versatility and potential of these models in various applications.