Building bots

Build scenario-based bots in Ebbot using scenarios, triggers, entities, and the training center.

This section covers scenario-based bots in Ebbot.

You build a bot by chaining scenarios. Each scenario has a trigger, training phrases, and one or more responses.

If you’re building a generative AI agent, start in EbbotGPT.

The building blocks

• Scenarios: the “units” of your bot. Each one covers a user intent.

• Triggers: how a scenario is entered (NLP, keyword, catch-all, etc.).

• Entities: extract structured values from user messages (dates, names, IDs, …).

• Training center: review unmatched phrases and improve intent matching over time.

Context (memory)

Context lets the bot handle follow-ups without the user repeating details. Ebbot auto-generates context for scenarios to keep conversations coherent.

Context is not limited to the chat transcript. You can enrich it with signals like:

• What page the visitor is on (and has visited)

• Device and browser

• Location (if you collect it)

• Customer data from other systems (CRM, ticketing, order history, etc.)

If you want details on the NLP/intent model options, see Intent detection (NLP).