Overview

Pre-load your reference data — doctor directories, clinic locations, service catalogs — into a Dynamic Table. When users ask the agent to look up an entry, vector search finds the right match even with typos: searching "dr. smuth" still returns "Dr. Smith", and "naw york" still returns "New York". This tolerance for near-matches significantly reduces lookup failures in real conversations.

After an appointment is booked, the agent schedules a reminder automatically. For example, a clinic can send a notification to each patient one day before their scheduled check-up — no human intervention needed.

After a service interaction, the agent guides users through a satisfaction survey conversationally. Responses are captured in structured format for later review or export.

Many SOPs must align with external regulations — health ministry guidelines, financial regulatory requirements, environmental standards. With Public Documents, employees can ask questions like "What does the latest regulation say about data retention periods?" or "Which OJK rule applies to this customer complaint process?" and get answers sourced directly from the official public documents.

Some procedures require checking real-time data before the next step — confirming a supplier is approved, checking stock levels before a purchase order, or verifying an employee's active status. SQL Database Reader lets the agent pull this data directly from the system of record.

When a procedure requires a sign-off checklist or incident report, the agent walks users through each required field conversationally. Submitted data is captured in consistent structured format for audit or compliance review.

Business users can ask "What were our top five products last month by revenue?" or "How many new customers signed up this quarter?" and receive answers drawn directly from the database — no SQL knowledge required.

Some analytics require reference data that changes over time but is not in the main database — sales target tables, regional classification mappings, or promotion calendars. Dynamic Table lets you maintain these datasets within the agent.

When an analysis requires user-defined parameters — date range, product category, region, or comparison baseline — a Chat Form guides users through the inputs step by step, ensuring the analysis is scoped correctly before any queries run.

Pair internal business data with publicly available references — government statistics (BPS), industry benchmark reports, or regulatory publications. When a user asks how their numbers compare to the national average or what a regulation says about a specific metric, the agent can answer from the public documents alongside the internal database query.