β Glean-ia-acs Β· Day 4 of 56
Reading Results and Relevance
Interpreting what Glean returns
You'll be able to
- Interpret a result's title, snippet, and ranking position
- Tell an intent mismatch apart from a ranking mismatch
Read first
Day 4 Lab
- Run a query you care about.
- For the top result, name what the title, the snippet, and the position each tell you.
- Decide: is the best answer missing (intent) or just ranked low (ranking)?
- Adjust the query for intent, or add a filter for ranking β then re-run.
Working Example: Anatomy of a Result
Every result carries three readable signals. Verified against retrieval_basics.
[Benefits 2026 β Parental Leave] β TITLE: what this is
drive Β· updated 3 days ago Β· @hr β METADATA: source, recency, owner
"...eligible employees accrue..." β SNIPPET: why it matched your query
position #1 of 240 β RANK: how confident Glean is
Diagnose a miss:
right type, wrong doc at #1 β RANKING (add a filter, e.g. updated:)
wrong type entirely β INTENT (reframe the query)Check for understanding
Your top result is the right type but the wrong document. Intent or ranking problem?
Check yourself
1. What does Glean's Enterprise Graph primarily connect?
The graph models high-value entities and the network of signals β docs, tickets, messages β around each.
2. When is a natural-language question better than keywords?
Questions let Glean assemble and cite an answer; keywords are faster when you just need to open a known doc.
3. Your top result is the wrong TYPE of thing entirely. That is a...
Wrong type means the query expressed the wrong intent; fix it by reframing, not by tuning ranking.
4. What does a result's snippet tell you?
The snippet shows the matched text β why this result surfaced. Position signals rank confidence; metadata shows owner.
5. What does `type:document onboarding` do?
type: narrows by document kind; here, only documents (not messages or tickets) about onboarding.
6. `app:slack quarterly roadmap` returns...
app: scopes the query to one connected system β here, Slack.
7. In one sentence, what does `from:me` do?
from: finds documents associated with a person; from:me is the special case for yourself.
In 10 seconds
βRead a Glean result like a label: the title says what it is, the snippet says why it matched, the position says how confident the rank is. Wrong type, fix intent; wrong order, fix signals.β
The type: Filter