Frequently Asked Questions (FAQ)

​

General

• Future AGI is an AI lifecycle platform designed to support enterprises throughout their AI journey. It combines rapid prototyping, rigorous evaluation, continuous observability, and reliable deployment to help build, monitor, optimize, and secure generative AI applications.

• You can get started by following our Quickstart Guide.

​

Evaluation

• Future AGI supports a wide range of evaluations including Hallucination, Guardrails, RAG, etc. See the Evaluation Overview for more details.

• Refer to the guide on Evaluating RAG Applications.

​

Knowledge Base

• You can add documents via the UI or using the SDK. See Create KB using UI and Create KB using SDK.

• PDF, DOCX, TXT, RTF, CSV, JSON, and more. Refer to the Knowledge Base Concepts.

​

Dataset

• Data can be added manually, via file upload, SDK, or imported from Hugging Face. See the Adding Dataset section.

• Dynamic columns allow you to generate new data based on existing columns using prompts, API calls, code execution, etc. Learn more in Create Dynamic Column.

​

Experimentation

• Experiments allow you to systematically compare different prompts, models, or parameters. See the Experimentation Overview.

• You can set up experiments to run different configurations against your datasets and compare results using evaluations. See How To Experiment.

​

Prototype

• Prototyping helps you iterate quickly on different versions of your AI application or prompt. See the Prototype Overview.

• You can compare prototypes based on evaluations and select the best performing one. See Choose Winner.

​

Observe

• Observe helps monitor key metrics like latency, cost, token usage, and evaluation results over time. See the Observe Overview.

• Alerts can be configured to notify you about anomalies or issues based on defined thresholds. See Alerts and Monitors.

​

Tracing (Observability)

• Tracing provides detailed visibility into the execution flow of your AI applications, helping debug issues and understand performance. See the Tracing Overview.

• We support auto-instrumentation for frameworks like OpenAI, Langchain, LlamaIndex, and more. See Auto Instrumentation.

​

Optimization

• Optimization provides a structured, iterative approach to refining AI-generated outputs by systematically improving prompts. Unlike experimentation, which focuses on testing multiple prompt variations, optimization enhances prompt by adjusting its structure based on evaluation-driven feedback. See the Optimization Overview.

​

Prompt Workbench

• The workbench provides tools to write, generate, and improve prompts. See the Prompt Workbench Overview.

​

Protect

• Protect helps enforce safety rules like toxicity, PII detection, prompt injection, etc., in real-time. See the Protect Overview.