Frequently Asked Questions

This section addresses common questions about OneByZero Neo, its capabilities, and how it fits into enterprise AI strategies.

What is OneByZero Neo?

OneByZero Neo is an enterprise-grade GenAI Factory that helps teams design, build, deploy, observe, and govern AI agents at scale. It provides a structured environment to assemble agents using models, tools, knowledge bases, guardrails, and workflows—while maintaining enterprise controls around security, cost, and reliability.

How is OneByZero Neo different from a chatbot framework?

Neo is not a chatbot builder. It is a factory and operating system for AI agents. While chatbots are a single interaction surface, Neo supports:

• Multi-agent systems - Coordinate multiple agents working together

• Tool-using and action-taking agents - Agents that can perform real actions through integrations

• Voice, chat, and backend agents - Multiple interaction modalities for different use cases

• Full lifecycle management - Build → Test → Deploy → Observe → Iterate

Does OneByZero Neo require AWS?

Neo is AWS-first by design and deeply integrates with services like Amazon Bedrock, IAM, S3, OpenSearch, CloudWatch, and EKS.

While the core platform is extensible, the default distribution is optimized for AWS environments to leverage native security, scalability, and governance capabilities.

Which LLMs are supported?

Neo supports multiple foundation models via pluggable providers. Out of the box, this includes:

• Amazon Bedrock models - Nova, Claude, Llama, and other models available through Bedrock

• Provider-specific fine-tuned models - Custom models from various providers

• Custom endpoints for enterprise-approved models - Self-hosted or privately deployed models

Model choice is abstracted from agent logic, allowing teams to swap or A/B test models without rebuilding agents.

How does Neo handle security and data isolation?

Neo is built with enterprise isolation by default:

• IAM-based access control - Fine-grained permissions using AWS IAM

• Workspace isolation - Logical separation between teams and projects

• VPC-native deployments - Network-level isolation within your AWS environment

• No training on customer data - Your data remains yours

• Fine-grained permissions - Granular control for models, tools, and knowledge bases

All data access is explicit, auditable, and controlled by the customer’s cloud account.

Can we bring our own tools, APIs, and knowledge sources?

Yes. Neo is designed for tool-centric agents and supports integration with:

• REST / SOAP / internal APIs - Connect to any HTTP-based service

• Databases and data warehouses - Query enterprise data sources

• Vector knowledge bases - Both in-house and Bedrock-managed knowledge bases

• Event-driven systems - Integration with message queues and event streams

Tools are registered once and reused across agents, with schema validation, retries, and observability built in.

How are agents tested and evaluated?

Neo supports pre-deployment and continuous evaluation, including:

• Prompt and behavior testing - Validate agent responses before deployment

• RAG quality checks - Ensure knowledge retrieval accuracy

• Regression testing across model changes - Verify behavior consistency when models are updated

• Human-in-the-loop feedback - Incorporate human review into the evaluation process

This allows teams to ship agents with confidence and safely evolve them over time.

Is OneByZero Neo suitable for production workloads?

Yes—Neo is explicitly designed for production-grade, mission-critical use cases, including telcos, banks, and large enterprises. It supports:

• High-availability deployments - Architected for enterprise uptime requirements

• Cost and usage tracking - Detailed analytics for budget management

• Versioning and rollback - Safe deployment practices with the ability to revert

• Centralized monitoring and alerts - Comprehensive observability across all agents

Neo is not a demo platform—it is a system for running AI agents as real software.