Prompt Engineering

Information Extraction Prompt — the Anatomy

The six sections a reliable extraction prompt needs: source guidance, field definitions, extraction rules, missing-data behavior, ambiguity policy, example.

Open in Extraction Prompt Generator

Overview

Most extraction prompts are one sentence and a hope. The ones that survive production have an anatomy: a SOURCE section that tells the model how to read this kind of text, FIELDS that define each piece of information, per-field EXTRACTION RULES (emails validated, dates ISO-formatted, identifiers untouched), a MISSING DATA contract, an AMBIGUITY POLICY, and an example of a valid extraction. This resource loads a job-posting extraction that exercises every section — including a salary_range field whose rule explicitly stops the model from collapsing a range into one number.

How to use this resource

Check each section's job

SOURCE teaches reading, FIELDS define meaning, RULES handle the traps, MISSING DATA and AMBIGUITY remove improvisation, the EXAMPLE shows the shape.
Watch the salary_range rule

"Extract the range as written — do not collapse it to one number." Per-field rules exist for exactly these traps.
Swap in your own fields

Replace the job-posting fields with yours; the engine derives new rules from the names you choose.

Why This Works

Each section closes a specific failure mode instead of adding generic words
Per-field rules catch the traps generic instructions miss — ranges, formats, identifiers
A strict ambiguity policy makes "I'm not sure" produce a blank, not a guess

Best for

Anyone whose extraction prompt is currently one sentence long
Prompts that work in testing and drift in production
Fields with traps — ranges, relative dates, lists inside prose

Not for

Quick one-off questions about a text — anatomy is overhead there
Output-format depth (types, strictness) — pair with the JSON Output Prompt Builder

Use cases

Learning the structure before writing your own extraction prompts
Auditing an existing extraction prompt against the six sections
Tracking job postings, listings, or announcements into a sheet

FAQ

How does the extraction prompt keep a salary range from being collapsed into one number?

A per-field EXTRACTION RULE handles that trap directly: salary_range must be extracted as written (e.g. "60,000–75,000 EUR") and explicitly not collapsed to a single number. Other fields get their own rules too; application_deadline uses ISO YYYY-MM-DD when the date can be determined. The extraction-prompt-generator derives these from your field names; you run the prompt in your assistant and check the JSON before trusting it.

How does the prompt return a field that isn't in the source text?

It includes the field with an empty string "" and never invents or guesses a value, per the MISSING DATA contract, backed by an AMBIGUITY POLICY that extracts only what's explicitly stated and treats implied-but-unstated values as missing. Output is a single JSON object using the exact field names, no markdown fences. The generator writes this; you still verify blanks are genuinely absent, not just missed.

Customize This Resource

Opens this setup in Extraction Prompt Generator. Generate to get the full extraction prompt — then adjust the fields, missing-data behavior, and ambiguity policy.

Open in Extraction Prompt Generator

Prompt Template

Copy it as-is, or use Open in Extraction Prompt Generator to load it pre-filled and customize it with your own context.

TASK
Extract role details from the job posting in the input for our market tracking sheet.

SOURCE
The input is unstructured free text.
Reading guidance: Read the entire text before extracting — relevant values may appear anywhere, in any order.

FIELDS TO EXTRACT
- position (required): Job title as posted
- company (required): Hiring company
- location (optional): Work location or remote policy
- salary_range (optional): Stated compensation range as written
- required_skills (required): Skills listed as requirements
- application_deadline (optional): Last day to apply

EXTRACTION RULES
- position: Exactly as written in the source.
- company: The organization's name as written, without surrounding commentary.
- location: The complete value as written, as a single string.
- salary_range: Extract the range as written (e.g. "60,000–75,000 EUR") — do not collapse it to one number.
- required_skills: Extract every distinct item as a separate list entry; deduplicate near-identical ones.
- application_deadline: Use ISO format (YYYY-MM-DD) when the date can be determined; otherwise extract it as written.
- Extract values from the source text only — do not use outside knowledge to fill fields.
- Do not summarize or interpret the document — return only the fields listed above.

MISSING DATA
- If a field is not found in the source, include it with an empty string "".
- Never invent or guess a value for a missing field.

AMBIGUITY POLICY
- Extract only information explicitly stated in the source text — do not infer values.
- If a value is implied but not stated, treat the field as missing.

OUTPUT FORMAT
Return the extracted data as a single valid JSON object using the field names below.

EXAMPLE OF A VALID EXTRACTION
{
  "position": "Senior Data Analyst",
  "company": "Acme Corp",
  "location": "Berlin, Germany",
  "salary_range": "60,000–75,000 EUR",
  "required_skills": [
    "Python",
    "SQL"
  ],
  "application_deadline": "2026-06-10"
}

OUTPUT RULES
- Return only the JSON object — no text before or after it.
- Do not wrap the output in markdown code fences.
- Use the field names exactly as defined — no renaming, no extra fields.

More resources from Extraction Prompt Generator

Resource

Extract Data From Text with AI

Free text in, named fields out. The extraction prompt pattern that turns any unstructured text into consistent, parseable records.

Prompt Engineering

Resource

Extract Invoice Data with AI

Invoice number, vendor, dates, total, currency — extracted into clean fields with strict no-inference rules, ready for accounts payable.

Operations

Resource

Missing Data in AI Extraction — Null, Unknown, or Skip

The most consequential setting in any extraction prompt: what the model does when the field isn't in the text. Four behaviors, and when each is right.

Prompt Engineering

Resources that pair well

Resource

Force JSON Output from AI

Stop getting 'Sure, here is the JSON…' — the output-contract pattern that forces models to return only parseable JSON: schema, example, and a strict rule block.

Prompt Engineering

Resource

Classify Support Tickets with AI

Billing, Technical, Account, How-To, Feature Request — ticket triage with definitions that decide the borderline cases for the model.

Support

Resource

Fix Invalid JSON from AI

The JSON won't parse and you can't see why. Deterministic cause-sniffing — trailing commas, single quotes, unclosed brackets — and the repair prompt that fixes it.

Engineering

Related tools

Tool

Extraction Prompt Generator

Build prompts that extract defined fields from unstructured text — emails, invoices, tickets, résumés.

Structured Output

Workflows that use this resource

Workflow

AI Classification Workflow

Build a text classification step you can automate on — pull out the unit to classify, assign a label from a fixed set, and validate the label is one you actually allow.

3 steps 25–45 minutes

Guides for this resource

Guide

Design an Extraction Schema Before You Extract

AI extraction returns clean JSON that quietly means something different each run. Design the extraction schema first — each field's meaning, type, allowed values, missing-value rule, and required evidence — and review it before a single record is pulled.

Structured Outputs & JSON

Tip: Save time by exploring related resources and tools that integrate with this resource.