<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/plan-before-you-build-deterministic-planning-patterns-for-ai-agents-by-dan-dobrin-spring-i-o-26-1xjdll3cn" -->

---
title: Plan Before You Build: Deterministic Planning Patterns...
description: A conference talk presents deterministic and non-deterministic planning patterns for AI agents, moving from a naive multi-model deliberation demo (an 'LLM...
canonical: https://daily.dev/posts/plan-before-you-build-deterministic-planning-patterns-for-ai-agents-by-dan-dobrin-spring-i-o-26-1xjdll3cn
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: Plan Before You Build: Deterministic Planning Patterns for AI Agents by Dan Dobrin @ Spring I/O 26 | daily.dev
og:description: A conference talk presents deterministic and non-deterministic planning patterns for AI agents, moving from a naive multi-model deliberation demo (an 'LLM...
og:url: https://daily.dev/posts/plan-before-you-build-deterministic-planning-patterns-for-ai-agents-by-dan-dobrin-spring-i-o-26-1xjdll3cn
og:image: https://api.daily.dev/og/posts/1xJDLl3Cn.png
og:image:alt: Plan Before You Build: Deterministic Planning Patterns for AI Agents by Dan Dobrin @ Spring I/O 26
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Plan Before You Build: Deterministic Planning Patterns for AI Agents by Dan Dobrin @ Spring I/O 26

**[Spring I/O](https://daily.dev/sources/springio)** · 46 min read · 0 upvotes · 0 comments

## Summary

A conference talk presents deterministic and non-deterministic planning patterns for AI agents, moving from a naive multi-model deliberation demo (an 'LLM council' inspired by an Andrej Karpathy exercise) toward a spectrum of architectures: compile-time deterministic workflows, LLM-driven supervisor agents, hierarchical task networks, utility AI (borrowed from game design), and goal-oriented action planning (GOAP) using the A* algorithm. Each pattern is demoed live using a Java 25 / Spring Boot 4 / Spring AI 2.0 milestone 4 / Vaadin Hilla application querying multiple Gemini and Claude models, with trade-offs discussed around cost, latency, determinism, parallelism, and observability via Google Cloud Trace. The speaker argues no single pattern wins and encourages picking the right approach per problem rather than the trendiest one.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://www.youtube.com/watch?v=0fH-tWLvDC4>

## Questions this post answers

### What is the LLM council pattern for multi-model AI agent deliberation?

It is a planning pattern where multiple LLMs independently answer the same query anonymously, then peer-review and rank each other's anonymized responses before a chairman model synthesizes a final consensus answer. The idea, based on an exercise by Andrej Karpathy, assumes no single model is always right and a council reduces hallucination and single points of failure, at the cost of higher API spend and latency.

_Teams weighing multi-model consensus approaches for agent reliability can follow pattern breakdowns like this on daily.dev._

### What is goal-oriented action planning (GOAP) and why is it useful for AI agents?

GOAP is a deterministic planning pattern borrowed from game AI that uses the A* search algorithm to find a cost-optimal sequence of actions from a current world state to a goal state, replanning after each step if conditions change. It has zero LLM planning cost, is fully reproducible for the same prompt, and discovers parallelism automatically, though it relies on a fixed action set and keyword-based goal selection.

_Developers picking between deterministic and LLM-driven agent orchestration can track pattern comparisons like this on daily.dev._

### What tech stack was used to build a multi-model AI agent deliberation demo?

The demo application was built with Java 25, Spring Boot 4, Spring AI 2.0 milestone 4, and Vaadin Hilla 25, deployable as a serverless app on Google Cloud Run. It uses Google Cloud Trace for observability of agent reasoning, token counts, and latency, and can run as a GraalVM native image to start in a few hundred milliseconds instead of several seconds.

_Java developers assembling an agent-ready stack can follow releases like Spring AI and Spring Boot updates on daily.dev._

## Similar posts on daily.dev

- [The Roadmap to Mastering Agentic AI Design Patterns](https://daily.dev/posts/the-roadmap-to-mastering-agentic-ai-design-patterns-rnti5xf8i) · Machine Learning Mastery · 4 upvotes · 0 comments
- [The AI Agent Engineer's Guide: 60 Patterns for Building Autonomous Systems \[Full Book\]](https://daily.dev/posts/the-ai-agent-engineer-s-guide-60-patterns-for-building-autonomous-systems-full-book--50d5hfpw5) · freeCodeCamp · 3 upvotes · 0 comments
- [How to Build Reliable AI Agents: 5 Engineering Patterns](https://daily.dev/posts/how-to-build-reliable-ai-agents-5-engineering-patterns-gpdk9dujm) · Salesforce Engineering · 10 upvotes · 0 comments
- [Agentic AI Patterns: one size does not fit all](https://daily.dev/posts/agentic-ai-patterns-one-size-does-not-fit-all-e2gaon1al) · Quarkus · 1 upvotes · 0 comments

---

Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#spring-ai](https://daily.dev/tags/spring-ai), [#llm-observability](https://daily.dev/tags/llm-observability)

[View this post on daily.dev](https://daily.dev/posts/plan-before-you-build-deterministic-planning-patterns-for-ai-agents-by-dan-dobrin-spring-i-o-26-1xjdll3cn)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"Plan Before You Build: Deterministic Planning Patterns for AI Agents by Dan Dobrin @ Spring I/O 26","url":"https://daily.dev/posts/plan-before-you-build-deterministic-planning-patterns-for-ai-agents-by-dan-dobrin-spring-i-o-26-1xjdll3cn","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/plan-before-you-build-deterministic-planning-patterns-for-ai-agents-by-dan-dobrin-spring-i-o-26-1xjdll3cn"},"datePublished":"2026-09-27T07:26:35.989Z","dateModified":"2026-09-27T07:32:20.612Z","description":"A conference talk presents deterministic and non-deterministic planning patterns for AI agents, moving from a naive multi-model deliberation demo (an 'LLM...","image":"https://i.ytimg.com/vi/0fH-tWLvDC4/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/0fH-tWLvDC4/sddefault.jpg","isAccessibleForFree":true,"articleSection":"Spring I/O","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Spring I/O","logo":"https://media.daily.dev/image/upload/s--e-TcwM9B--/f_auto,q_auto/v1790493985/logos/springio?_a=BAMAMicg0","url":"https://daily.dev/sources/springio"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/plan-before-you-build-deterministic-planning-patterns-for-ai-agents-by-dan-dobrin-spring-i-o-26-1xjdll3cn","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,spring-ai,llm-observability","timeRequired":"PT46M","video":{"@type":"VideoObject","name":"Plan Before You Build: Deterministic Planning Patterns for AI Agents by Dan Dobrin @ Spring I/O 26","description":"A conference talk presents deterministic and non-deterministic planning patterns for AI agents, moving from a naive multi-model deliberation demo (an 'LLM...","thumbnailUrl":"https://i.ytimg.com/vi/0fH-tWLvDC4/sddefault.jpg","uploadDate":"2026-09-27T07:26:35.989Z","duration":"PT46M","url":"https://api.daily.dev/r/1xJDLl3Cn","embedUrl":"https://www.youtube.com/embed/0fH-tWLvDC4"}}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Spring I/O","item":"https://daily.dev/sources/springio"},{"@type":"ListItem","position":3,"name":"Plan Before You Build: Deterministic Planning Patterns for AI Agents by Dan Dobrin @ Spring I/O 26"}]}
{"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/plan-before-you-build-deterministic-planning-patterns-for-ai-agents-by-dan-dobrin-spring-i-o-26-1xjdll3cn#faq","mainEntity":[{"@type":"Question","name":"What is the LLM council pattern for multi-model AI agent deliberation?","acceptedAnswer":{"@type":"Answer","text":"It is a planning pattern where multiple LLMs independently answer the same query anonymously, then peer-review and rank each other's anonymized responses before a chairman model synthesizes a final consensus answer. The idea, based on an exercise by Andrej Karpathy, assumes no single model is always right and a council reduces hallucination and single points of failure, at the cost of higher API spend and latency. Teams weighing multi-model consensus approaches for agent reliability can follow pattern breakdowns like this on daily.dev."}},{"@type":"Question","name":"What is goal-oriented action planning (GOAP) and why is it useful for AI agents?","acceptedAnswer":{"@type":"Answer","text":"GOAP is a deterministic planning pattern borrowed from game AI that uses the A* search algorithm to find a cost-optimal sequence of actions from a current world state to a goal state, replanning after each step if conditions change. It has zero LLM planning cost, is fully reproducible for the same prompt, and discovers parallelism automatically, though it relies on a fixed action set and keyword-based goal selection. Developers picking between deterministic and LLM-driven agent orchestration can track pattern comparisons like this on daily.dev."}},{"@type":"Question","name":"What tech stack was used to build a multi-model AI agent deliberation demo?","acceptedAnswer":{"@type":"Answer","text":"The demo application was built with Java 25, Spring Boot 4, Spring AI 2.0 milestone 4, and Vaadin Hilla 25, deployable as a serverless app on Google Cloud Run. It uses Google Cloud Trace for observability of agent reasoning, token counts, and latency, and can run as a GraalVM native image to start in a few hundred milliseconds instead of several seconds. Java developers assembling an agent-ready stack can follow releases like Spring AI and Spring Boot updates on daily.dev."}}]}
```

