Class 01

Conversational Analytics with Looker

From semantic layer to trustworthy data agent.

1. Why conversational analytics changes BI

Natural-language interfaces, business questions, and why removing SQL increases the need for governed meaning.

2. LookML as the semantic contract

Measures, dimensions, descriptions, synonyms, joins, grain, and business definitions.

3. Designing Explores for CA

Field quality, model hygiene, scope, discoverability, and question patterns.

4. Building data agents

Agent instructions, business context, multiple Explores, and verified query patterns.

5. Prompt vs semantic modeling

What belongs in LookML, what belongs in instructions, and what should never depend on prompts alone.

6. Evaluation and trust

Golden questions, expected answers, query inspection, failure taxonomy, and accuracy tracking.

7. Conversational Analytics APIs

Agents, conversations, messages, chat APIs, streaming, persistence, and custom application integration.

8. Enterprise production architecture

Auth, allowlists, audit logs, observability, governance, security boundaries, and deployment patterns.