Pair your devices with a code and playback position follows you: pause on this device, hit resume on the other. Position is saved to the site every minute and on pause.
Open this panel on your other device and enter the same code.
18 questions, exam-styled. One best answer each, no partial credit. Answers with explanations at the end.
1. A viewer of a shared space reports they can open every uploaded file but get "access denied" on one linked dashboard. What's happening?
2. Which source's content can be ingested into the Quick Index via a knowledge base?
3. Space uploads across the org suddenly fail. The most likely cause is:
4. Your compliance team asks for the transcript of what employees asked Quick chat last month. Where does it live?
CHAT_LOGS CloudWatch vended log type — if it was enabled before the conversations happenedaws quicksight list-conversations5. A custom chat agent must always apply a 3-page refund policy, and must be able to search 4,000 historical support tickets. Correct placement?
6. A developer wants to call Quick chat from a backend service. The supported answer is:
aws quicksight start-automation-job with a chat payload7. What's the maximum number of datasets in a (new, multi-dataset) Topic, and where does semantic metadata live in that model?
8. A legacy topic answers "revenue by region vs target" wrongly, using only the revenue dataset. Why?
9. Your Terraform pipeline manages topics via create-topic. To provision the new
multi-dataset kind you must:
--multi-dataset to create-topiccreate-topic-v2 family — a different object shape with DataSetRelations10. Fastest way to give readers natural-language questions on one existing dashboard:
11. Which task belongs in Quick Automate rather than Quick Flows?
12. A flow step lists "all open tickets" from a connected app. Documented behavior:
13. What is the Quick Automate UI agent?
14. An automation edit is committed but production still runs old logic. Why?
15. Automate's custom agent in Custom mode (BYO Bedrock model, e.g. Claude 3.5 Sonnet) — where does inference cost land?
16. Ten Professional users; the team runs 100 Research reports this month at ~30 min each. Roughly what overage bills?
17. Which IAM action gates the built-in usage analytics dashboard?
quicksight:DescribeUsagequicksight:QuickSuiteUsageMetricsqbusiness:GetMetricscloudwatch:GetDashboard18. The user guide says a space holds 10 GB; the Service Quotas table says 1 GB (non-adjustable). For capacity planning you:
1. B — The documented rule: resources in a space respect their own permissions, but "files uploaded to a space are always available to everyone with access to the space." A (share) isn't implicated — files opening proves the share works. C would give a data error, not access denied. D — viewers can view.
2. C — SharePoint Online is one of the six KB-capable sources (with S3, Confluence Cloud, Google Drive, OneDrive, Web Crawler). A, B, D are Actions-only in the supported-integrations matrix: queryable live, never indexed.
3. B — "Files uploaded to spaces will be rejected if your index data storage capacity is full." A is per-conversation, not spaces; C would affect one space, not the org; D is unrelated.
4. B — Chat content is in CloudWatch vended logs, not CloudTrail (which has the API actions), and vended logs only capture from when they were enabled. D doesn't exist.
5. C — Reference documents stay permanently in context (right for mandatory policy; limits 10 docs/50 MB/100K chars). Spaces are retrieval (right for a large searchable corpus). A dies on the 100K-character wall; D risks the policy not being retrieved for some questions.
6. B — No conversation API exists in the quicksight namespace; the announced programmatic route is embedded chat. C starts automations. D — Quick's docs never document chat as a Q Business proxy.
7. A — "Add up to 12 datasets to a single Topic"; metadata is enriched per-dataset before adding. B describes the legacy metadata location; C wrong number; D confuses index with semantics.
8. B — Legacy engine: "ML-based fuzzy search model selects one dataset, then queries only that dataset." Cross-dataset needs a new topic with defined relationships. A is an undocumented number; C, D unrelated.
9. B — V2 is a parallel operation family (create-topic-v2 etc.) whose body is
DataSets + DataSetRelations + CustomInstructions. No flag (A) or account toggle (C) converts the
old API.
10. C — Dashboard Q&A is exactly this: one click, no topic, enablement fee noted in docs. A works but is days of modeling; B is what-if analysis; D grounds chat, not the dashboard's Q&A button.
11. C — Credentialed unattended runs, approvals, error handling = Automate's stated scope ("independent of an individual user"). A is the flows browser extension; B is a personal flow; D is literally a documented flow-creation path.
12. B — Verbatim: "Quick retrieves only the first batch to keep your flow fast and responsive." Complete processing belongs elsewhere.
13. B — Docs: "autonomously navigate websites, click, type, read data, and produce structured outputs." C is a different feature (flows in the extension); the extension runs flows, it doesn't autonomously drive pages.
14. B — "Only committed versions can be deployed" — commit and deploy are separate steps. Nothing auto-deploys (A). C affects HITL tasks, not versioning.
15. C — "Model inference is billed separately to the account associated with that Bedrock integration." It bypasses agent hours (A) and the flat fee (B) entirely — the classic hidden cost.
16. B — Pool: 10 users × 2 Research hours = 20. Usage: 100 × 0.5 h = 50 h → 30 over × $3 = $90. A false; C forgets the included pool; D confuses meters.
17. B — Documented verbatim on the incident-response/logging page. The others don't exist or gate different things.
18. C — Course rule for contradicting official pages: plan against the conservative documented number and re-verify. B is invented precision; D fails — the quota is marked non-adjustable.
| Score | Reading |
|---|---|
| 16–18 | Expert-track. Do the lab if you haven't, then teach this module to someone. |
| 12–15 | Solid. Re-read the lessons behind your misses — most gaps here are one fact wide. |
| 8–11 | You know the shapes but not the numbers. Re-listen on your next drive, then retake. |
| <8 | Restart at lesson 1; do the lab before retaking. The quiz will still be here. |