---
title: "Discussion on \"Agentic AI Beyond the Hype: Why the Boundaries Are Key in Deterministic and Non-Deterministic Behavior for Agents\" | Hashnode"
description: "Discussion on \"Agentic AI Beyond the Hype: Why the Boundaries Are Key in Deterministic and Non-Deterministic Behavior for Agents\". 90% of LLM systems fail in production due to rigid controls. Discover how to architect deterministic boundaries and handle uncertainty in your agents."
url: "https://hashnode.com/posts/agentic-ai-beyond-the-hype-why-the-boundaries-are-key-in-deterministic-and-non-deterministic-behavior-for-agents/6aa9f0845ae9caa091ac0ba4"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Agentic AI Beyond the Hype: Why the Boundaries Are Key in Deterministic and Non-Deterministic Behavior for Agents](https://designednotmagic.hashnode.dev/agentic-ai-beyond-the-hype-why-the-boundaries-are-key-in-deterministic-and-non-deterministic-behavior-for-agents?utm_source=hashnode&utm_medium=feed)

90% of LLM systems fail in production due to rigid controls. Discover how to architect deterministic boundaries and handle uncertainty in your agents.

Sep 16 · 4 min read · designednotmagic.hashnode.dev  
Read the article

-   [#agenticai](https://hashnode.com/tag/agenticai)
-   [#aiagents](https://hashnode.com/tag/aiagents)
-   [#softwarearchitecture](https://hashnode.com/tag/softwarearchitecture)
-   [#systemdesign](https://hashnode.com/tag/systemdesign)
-   [#aiinfrastructure](https://hashnode.com/tag/aiinfrastructure)
-   [#distributedsystems](https://hashnode.com/tag/distributedsystems)
-   [#softwareengineering](https://hashnode.com/tag/softwareengineering)
-   [#machinelearning](https://hashnode.com/tag/machinelearning)
-   [#llms](https://hashnode.com/tag/llms)
-   [#artificialintelligence](https://hashnode.com/tag/artificialintelligence)
-   [#generativeai](https://hashnode.com/tag/generativeai)
-   [#designednotmagic](https://hashnode.com/tag/designednotmagic)

## Written by

[Joselo Martinez](https://hashnode.com/@joselozn)

@joselozn

  
[View profile](https://hashnode.com/@joselozn)

Design-driven AI. Real tools. No tricks.

## Responses 2

Markdown supported · ⌘↵ to post

[Joselo Martinez](https://hashnode.com/@joselozn)

[Sep 18](https://hashnode.com/posts/agentic-ai-beyond-the-hype-why-the-boundaries-are-key-in-deterministic-and-non-deterministic-behavior-for-agents/6aa9f0845ae9caa091ac0ba4/comment/6aacb20eb3da689d7e712d65)

Thank you so much! Coming from a team that develops a production-grade SaaS like BotSailor, this means a lot to me. Managing omnichannel state (WhatsApp, Telegram, etc.) must make balancing the line between deterministic and probabilistic behavior incredibly complex. If a sales bot "hallucinates" a workflow, the impact on the business is immediate. I’m really looking forward to reading your series on "Agent AI in Production" to see how you handle state persistence. Cheers!

[BotSailor](https://hashnode.com/@botsailor)

[Sep 16](https://hashnode.com/posts/agentic-ai-beyond-the-hype-why-the-boundaries-are-key-in-deterministic-and-non-deterministic-behavior-for-agents/6aa9f0845ae9caa091ac0ba4/comment/6aaa6c0bc28e4c50fac5ce12)

Really enjoyed this perspective. The idea that uncertainty should be treated as an explicit part of the system rather than simply another error to patch really stood out to me. It’s tempting to keep adding guardrails whenever an agent behaves unexpectedly, but that can eventually make the system overly rigid. I especially liked the comparison with exception handling in distributed systems it makes the concept much easier to understand. Defining where deterministic control ends and where the agent is allowed to reason and adapt seems like an important architectural decision. Looking forward to seeing how the next parts explore reliability and maintainability in practice.
