Prompt construction
Goals, context, constraints, and output shape
Build a request whose result can be checked against supplied facts and explicit requirements.
Lesson 4 of 6 in the recommended order · About 25 min (estimate)
On this page
Practical AI glossary — terms and common confusions
- Model
- A learned component used to produce a result.
A drafting model is one part of a letter app; it is not the whole app.
- Application
- The software experience around components and services.
A letter app adds accounts, storage, and a Send button.
- Prompt
- Instructions and input supplied for a task.
“Summarize this notice in two bullets” specifies a task and shape, not a truth guarantee.
- Token
- A unit a model processes; it need not be a whole word.
A tokenizer can split a name into pieces; count using the actual system.
- Context
- Information available for the current request.
An earlier attachment may be absent even when its filename is visible.
- Training
- Adjusting a model using examples.
A training log differs from a conversation correction.
- Inference
- Using a trained model to produce a result.
Drafting another sentence does not by itself prove retraining.
- Retrieval
- Selecting material to supply to a task.
A retrieved notice still needs a relevance and date check.
- Embedding
- A numerical representation used to compare or organize items.
A similar search match is a candidate, not proof of the same meaning.
- Tool
- A function an application can call.
A calendar lookup can be read-only; finding an event is different from creating one.
- Agent
- Software that can choose and carry out steps toward a goal.
An agent still needs bounded permissions and checked action receipts.
- Hallucination
- Plausible generated content that is unsupported or false.
An invented opening time needs evidence or removal, even when phrased confidently.
- Evaluation
- Checking results against stated criteria.
Twenty checked examples establish a bounded result, not universal reliability.
- Multimodal
- Working with more than one kind of input or output.
A photo plus text does not prove every small label in the photo was read correctly.
Outcome
Evaluate whether a prompt and its output preserve the task’s facts, limits, and requested structure.
Why it matters
Build a request whose result can be checked against supplied facts and explicit requirements.
Concept
A checkable prompt states the goal, supplies relevant context, identifies constraints, and describes the output shape. A goal says what the work should accomplish. Context provides the facts needed. Constraints define boundaries such as source limits or length. Output shape tells the reader how to use the result, such as a short table with named columns.
A format instruction is not a truth guarantee. A tidy table can still contain invented values. Add a rule for unknowns, such as “write not provided,” and check every populated field against the source. If requirements conflict, resolve the conflict instead of silently ignoring one.
Worked example
For a fictional roster note, ask: “Using only the two supplied shifts, make a table with Person and Start columns. Do not infer missing times; write not provided.” A row inventing 9 a.m. for a missing start is wrong even if the table is beautifully formatted. A correct “not provided” cell should be retained.
Before you check
Draft your own four-part prompt for a supplied packet, then write one observable acceptance check for each constraint. You can complete this on paper without calling an AI service.
Practice and fresh transfer
The packets below are fictional and contain the facts needed for these cases. External references are optional background. Judge each response independently: it may be supported, contradicted, or unresolved. Select the passages needed to justify your judgment and write why the distinction matters before revealing feedback.
Assess the response
Use Accept when all material claims are supported. Use Revise when a supplied fact or requirement is contradicted. Use Evidence is insufficient when a key fact cannot be established either way. If a response contains both an unknown and a direct contradiction, choose Revise and explain both problems. Conflicting claims with no established authority remain insufficient; a claim does not become a governing fact merely because a source asserts it.
Some responses are fully supported. Others need correction or more evidence. Judge each on its sources; do not edit a correct answer just to change it.
Your written notes stay in this page and disappear when you leave. Only a self-reviewed completion can be saved to your learning account. These practice checks do not establish independent proficiency.
Practice
Case 1
Consider a fictional prompt designed from a short briefing request.
Source packet
- Source 1
- Request: From the supplied notice, produce three bullets for volunteers: date, place, and what to bring. Mark absent facts “not provided.”
- Source 2
- Notice: Volunteer briefing is 8 June in Hall B. No equipment instruction is included.
Response to assess
Use only the notice. Return three bullets labeled Date, Place, and Bring; use “not provided” for any absent fact.
Practice
Case 2
Consider a fictional formatted output, not merely its appearance.
Source packet
- Source 1
- Prompt: Use only the notice; return Date and Cost. Write “not provided” for missing values.
- Source 2
- Notice: The open day is 9 July. No cost is stated.
Response to assess
Date: 9 July. Cost: Free.
Fresh transfer
Case 3
Consider whether a fictional request is ready for a single compliant output.
Source packet
- Source 1
- Requirement A: The final response must contain exactly two sentences.
- Source 2
- Requirement B: The same final response must contain exactly five sentences. Both are current; neither has priority.
Response to assess
A single response can satisfy both sentence counts exactly without clarification.
Fresh transfer
Case 4
Consider a fictional requested comparison row.
Source packet
- Source 1
- Prompt: Compare only stated access methods; output Service, Access, Unknowns. Do not guess prices.
- Source 2
- Service note: Birch offers phone booking. Prices are not listed.
Response to assess
Service: Birch; Access: phone booking; Unknowns: price not listed.
Save your self-review
Completion records that you reviewed the cases. Your explanation and transfer performance need a facilitator to establish independent learning.
Sign in with your learning-center account to save completion.
Review the explanation for every case before saving.
Summary and next step
Apply the checklist to a new task. Preserve supported content, explain any change with evidence, and name what remains unresolved. Saving records self-review, not independently demonstrated proficiency. A facilitator must assess the explanation and fresh transfer for human learning evidence.