> For the complete documentation index, see [llms.txt](https://docs.openagrinet.global/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.openagrinet.global/deliberations/deliberations/ai-as-orchestration-not-source-of-truth.md).

# AI as Orchestration, Not Source of Truth

## Context

A conversational AI layer sitting in front of agricultural advisory, scheme, and pricing data could either generate answers itself from what it has learned, or be constrained to only relay information it retrieves from verified systems. For a farmer-facing agricultural network, an unconstrained model risks presenting a hallucinated price, scheme eligibility rule, or pest treatment as if it were fact.

## Decision

The OAN AI module is designed so that AI never acts as a source of truth. It only decides which systems or data sources to invoke and how to present the resulting information — every factual response is:

* Grounded in a knowledge base sourced from trusted institutions (for advisory and schemes), or
* Backed by live API/tool calls (for weather, prices, status checks, and grievances), and
* Constrained by domain-specific guardrails.

## Why it matters

* Responses stay verifiable and auditable, since every factual claim traces back to a registered tool call or an institutional document rather than to the model's own generation.
* It structurally rules out a class of failure — hallucinated or speculative agricultural advice — that would otherwise be the highest-consequence failure mode for this domain.
* It keeps the AI layer swappable/LLM-agnostic (see [Reuse across Deployments](https://docs.openagrinet.global/implementations/implementations/reuse-across-deployments.md)): because the model's job is orchestration rather than knowledge, the underlying LLM/SLM can be changed without changing what the system is allowed to assert.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.openagrinet.global/deliberations/deliberations/ai-as-orchestration-not-source-of-truth.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
