Capabilities and uncertainty
Search, retrieval, tools, and agents
Separate finding a passage, generating a response, and completing an external action.
Lesson 5 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
Use a tool record to identify what was retrieved, attempted, and actually confirmed.
Why it matters
Separate finding a passage, generating a response, and completing an external action.
Concept
Search finds candidate material. Retrieval supplies selected material to a task. A tool gives an application access to a function, such as reading a record or creating an event. An agent can use a sequence of such steps toward a goal. None of these labels removes the need to check sources, permissions, and outcomes.
Read action records carefully: a proposal is not an attempt; an attempt is not a successful result. A successful read does not imply write access. A returned passage still needs a date and relevance check. In the supplied cases, use the explicit receipts rather than the assistant's description of what it supposedly did.
Worked example
A fictional assistant retrieves a room listing, proposes a booking, and receives “permission denied” from the booking tool. “The room is booked” is contradicted by the result. A useful handoff says the listing was found but booking failed, then asks the authorized person to complete or approve the next step.
A retrieved listing is not a completed booking
1. Retrieval: confirmed read
Room listing
Retrieval returned the listing
A successful read establishes a candidate room. It does not grant booking permission or create a reservation.
2. Proposal: intent only
Proposed action- Book the room.
Evidence so far- The listing was found; no booking result follows from a proposal.
A proposed next step is not a tool call or a completed action.
3. Booking attempt: denied
Booking tool attempt
Permission denied
The tool result says the booking failed. Report the listing as found and ask an authorized person to complete or approve the next step.
Before you check
Mark each step as proposed, attempted, or confirmed. Record the evidence for the last confirmed step; that is the safe starting point for a handoff.
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 room-search handoff.
Source packet
- Source 1
- Tool receipt: Catalog lookup succeeded and returned rooms Elm and Ash. This tool has read-only access.
- Source 2
- User request: Find options; do not book anything.
Response to assess
The catalog returned Elm and Ash as options. No booking is reported.
Practice
Case 2
Consider the fictional assistant completion message against an action receipt.
Source packet
- Source 1
- Tool receipt: create_event failed: approval_required. No event ID was created.
- Source 2
- Assistant plan: Create a calendar hold for 2 p.m.
Response to assess
The 2 p.m. event is on your calendar.
Fresh transfer
Case 3
Consider a fictional lookup answer that cites a retrieved passage.
Source packet
- Source 1
- Retrieved passage: Version 3 room rules, effective 1 June, says evening entry ends at 8 p.m.
- Source 2
- Catalog metadata: Version 3 is the current rule for the requested room; version 2 is archived.
Response to assess
The current rule for this room ends evening entry at 8 p.m., according to version 3.
Fresh transfer
Case 4
Consider whether a fictional purchase status is known after a tool interruption.
Source packet
- Source 1
- Action log: The purchase request was sent; the connection ended before a response arrived.
- Source 2
- Tool guidance: When outcome is unknown, look up the transaction by request ID before retrying to avoid a duplicate.
Response to assess
The purchase definitely failed.
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.