Structured Outputs: JSON Mode and Reliable Data Extraction

Because eval() is not a valid parsing strategy.

The Unstructured Problem

You ask the LLM: List the fruits.
It replies: Sure! Here they are: 1. Apple, 2. Banana.
Your code crashes because it expected [Apple, Banana].

The Solution: Enforced Structure

1. JSON Mode

Most modern APIs (OpenAI, Anthropic) support a response_format={type: json_object} flag.
You MUST include JSON in the system prompt for this to work reliably.

2. Function Calling / Tool Use

This is even better than JSON mode. You define a fake function schema, and the model attempts to call it with arguments.

• Schema: submit_fruits(fruits: List[str])
• Model Output: submit_fruits(fruits=[Apple, Banana])
• Your Code: Parses the arguments cleanly.

3. Instructor / Pydantic

We use libraries like Instructor to patch the OpenAI client so it returns Pydantic objects directly.

class UserDetail(BaseModel):
    name: str
    age: int

user = client.chat.completions.create(..., response_model=UserDetail)
print(user.age) # Its an integer!

Retry Logic

Even with JSON mode, the model sometimes generates invalid JSON (e.g., trailing commas).
Pattern:

1. Try to parse JSON.
2. except JSONDecodeError: Send the error message back to the LLM. You generated invalid JSON. Fix it.
3. The model usually fixes it on the second try.

Conclusion

If your application relies on Regex to parse LLM output, you are doing it wrong. Use Structured Outputs.

---

*Next: [LLM Evaluation](/blog/llm-evaluation-how-to-know-working)*

How This Article Was Made

This article is a live example of the AI-enabled content workflow we build for clients.

Stage	Who	What
Research	Claude Opus 4.5	Analyzed current industry data, studies, and expert sources
Curation	Tom Hundley	Directed focus, validated relevance, ensured strategic alignment
Drafting	Claude Opus 4.5	Synthesized research into structured narrative
Fact-Check	Human + AI	All statistics linked to original sources below
Editorial	Tom Hundley	Final review for accuracy, tone, and value

The result: Research-backed content in a fraction of the time, with full transparency and human accountability.

Why We Work This Way

Were an AI enablement company. It would be strange if we didnt use AI to create content. But more importantly, we believe the future of professional content isnt AI vs. Human—its AI amplifying human expertise.

Every article we publish demonstrates the same workflow we help clients implement: AI handles the heavy lifting of research and drafting, humans provide direction, judgment, and accountability.

Want to build this capability for your team? Lets talk about AI enablement →

---