# Finding and attaching sources
> **What you will learn**
> How to give a model your internal material, why "just paste everything" stops
> working, and why a citation on the answer is the decisive part.
## A model only knows what it was trained on
A [language model](/guide/ai-language-model) predicts the next word from what it
saw during training. So it structurally cannot answer this:
```
"What was our return rate in Q4 last year?"
→ our return rate is not in the training data
→ yet a plausible-looking number comes out more easily than "I don't know"
```
This is where [hallucination](/guide/ai-hallucination) most commonly appears.
**You asked about something it does not know.**
## The fix is one line
> **Before you ask, find the material that would justify the answer and pass it
> in alongside the question.**
The industry calls this **RAG** (retrieval-augmented generation). The name is
harder than the idea.
```mermaid
graph TD
A["Question: what was the Q4 return rate?"] --> B["1. Search internal material"]
B --> C["2. Pass 3-5 retrieved chunks + the question to the model"]
C --> D["3. Model answers from those chunks only"]
D --> E["Answer + source (which document, which page)"]
```
**You are not retraining the model's weights.** You are fetching material each
time you ask. That is why updating a document takes effect on the very next
question.
This is still learning — it just **accumulates in the material rather than in the
model.** [Which kind of learning?](/guide/ai-rag-vs-finetune) splits that four ways.
## "Can't we just paste everything?"
The most common question, and the answer is **yes, when there is little of it.**
For a single policy handbook, pasting the whole thing beats building search.
The problem is growth.
| Paste everything | Search, then paste |
|---|---|
| Beyond the [context window](/guide/ai-context-window) it simply doesn't fit | only the chunks needed |
| Every question costs the full volume | a few chunks' worth |
| Irrelevant content dilutes accuracy | only what is relevant |
| Breaks as documents grow | unchanged as documents grow |
The third row surprises people. **More material does not mean better answers.**
Bury the right paragraph among fifty irrelevant documents and the model misses
it exactly like a person would.
## How retrieval works
Two approaches, and in practice you want both.
| Approach | Basis | Good at | Bad at |
|---|---|---|---|
| Keyword search | matching words | product codes, IDs, proper nouns | "refund policy" won't find "return terms" |
| Semantic search | closeness in meaning | finds it despite different wording | weak on exact codes and figures |
**Semantic search** turns text into a list of numbers (an embedding) and groups
what lands close together. "Refund policy" and "return terms" share no words,
but as numbers they sit near each other.
> **Combining the two is the standard.** Pick only one and there is a class of
> question you will always miss.
## Chunking decides the outcome
Documents are not stored whole — they are split into **chunks**. And a chunk
torn out of its page loses its context.
```
Page 3: "...for grade B customers this limit is 3 million won..."
→ as a chunk: "for grade B customers this limit is 3 million won"
→ limit on what? "monthly credit limit" was in the paragraph above, now gone
```
**An experiment Anthropic published** shows the size of this problem. Prepending
one line of context — which document, which section — before storing each chunk
cut the retrieval failure rate from 5.7% to 3.7%, a **35% reduction**. Adding
keyword search brought it to 49%, and re-ranking the retrieved chunks brought it
to 67%.
```
1. Chunk with context attached → 35% fewer failures
2. Add keyword search → 49%
3. Re-rank what was retrieved → 67%
```
**"We added search but it doesn't find things" is usually step 1 missing.**
## What a citation buys you
This is the part that matters most in practice.
```
✗ "The Q4 return rate was 3.2%."
✓ "The Q4 return rate was 3.2%. (2024_Q4_quality_report.pdf, p.7)"
```
**The second one can be checked.** Someone opens page 7 and confirms it in five
minutes. The first one has to be re-derived from scratch to verify.
> **An answer without a source is unusable in an organisation.** When it's right
> you cannot show why; when it's wrong you cannot tell where. The real reason to
> add retrieval is **verifiability**, more than accuracy.
## In the product
| What it does | Where |
|---|---|
| Uploaded files become answer material | [Connect Drive](/guide/cn-drive) |
| PDFs and documents converted to searchable text | [turning documents into material](/guide/ht-documents) |
| Picking only the relevant tools when there are many | [when there are too many tools](/guide/mcp-too-many-tools) |
The last row is the interesting one. **Choosing tools runs on the same
principle** — tool descriptions are stored as numbers, and only the ones close
to the question are pulled in.
## Common misconceptions
### "Does retrieval eliminate hallucination?"
**It reduces it; it does not remove it.** Three cases survive: retrieval fetched
the wrong document, the fetched document is itself stale or wrong, and the
material has no answer so the model fills the gap with something nearby.
The last one is largely fixable by instruction. State explicitly: **"if the
provided material does not support an answer, say so."**
### "Should we load in all our documents?"
Decide **who is allowed to see what** first. Retrieval does not know about
permissions. Put HR files in the same store and HR files will surface in
anyone's question. Same argument as [tools are
permissions](/guide/mcp-security).
---
## Check yourself
**1. Why does the model invent a number when asked for last year's return rate?**
Answer
**You asked about something absent from its training data.** It predicts the
next word from what it learned, so in an unknown area a plausible number comes
out more naturally than "I don't know." Fetching the material and passing it in
is the fix.
**2. What changes when you prepend one line of context to each chunk?**
Answer
**Retrieval failures drop sharply.** In the published experiment, 5.7% → 3.7%, a
35% reduction — and 67% with keyword search and re-ranking added. Without it,
"this limit" in an isolated chunk no longer says a limit on what.
**3. Why does a citation matter more than accuracy?**
Answer
**Because it makes the answer checkable.** With a source, someone confirms it in
five minutes. Without one, you cannot show your basis when right or locate the
error when wrong. For organisational use, verifiability comes first.
---
Now spend thirty minutes doing it. The next chapter has you see the difference
for yourself → [Try attaching sources](/guide/ai-try-grounding)