What is Prompt Engineering?

AIIntermediate#AI Safety#Interaction#Validation

Prompt Engineering

Definition

The practice of designing model instructions, examples, context, and output constraints for a bounded task. It can make a workflow easier to evaluate; it does not make model output deterministic, correct, secure, or suitable for a financial decision.

Source reviewed

NIST AI RMF: Generative AI Profile opens in a new tab
Supports: NIST describes confabulation as confidently generated false or inconsistent content and notes the heightened risk when people act on consequential outputs.
OWASP: LLM01 Prompt Injection opens in a new tab
Supports: OWASP documents that untrusted prompt content can alter model behavior and recommends handling application functions in code rather than delegating direct authority to model output.
OWASP: LLM06 Excessive Agency opens in a new tab
Supports: OWASP explains that broad tool authority lets unexpected, ambiguous, or manipulated model output cause damaging actions.

Concept Diagram

Key Takeaways

Prompt engineering defines a bounded task, evidence, constraints, output format, and response to missing or conflicting information.

Prompts and retrieved context are data, not authority: untrusted content must not change application policy or tool permissions.

Schema validation improves the handling of structured output; it does not make a claim true or authorize an action.

Keep transfer, order, and access-control decisions outside the model behind independent policy checks and explicit approval.

Practical Example

A review prompt asks the model to summarize only the supplied official notice, list each claim with a source excerpt and timestamp, identify missing facts, and return a schema-validated draft. The application rejects any output that asks for a transaction, changes a policy, or lacks a cited record.

Finished learning?

Related Terms

Agenticity

An informal way to describe how much an AI system can observe, plan, call tools, and act. It is not a standardized maturity score, a measure of trustworthiness, or permission to give an agent broad authority.

Perception Layer

The part of an AI workflow that collects and labels inputs such as blockchain records, market data, and public communications. It supplies context for review; it does not establish truth, identify people, predict prices, or justify an action.

LLM Reasoning

Using a language model to summarize, classify, or compare supplied information. Its output is probabilistic text, not verified market analysis, financial advice, or authority to execute a trade.

RAG (Retrieval-Augmented Generation)

A system pattern that retrieves selected external records and supplies them with a query to a language model. Retrieval can improve traceability and recency when sources are cited; it does not prove that a record is true, complete, current, or safe to act on.

Master the hub of AI

Explore all our strategic guides about AI to take your operations to the next level.

View all articles