How Zeltino works

Zeltino brings the visitor-facing chat widget and your team's follow-up work into one connected experience. A visitor may only see a conversation, but behind it are the knowledge, behaviour, automations, and project records that help the right answer or next step appear.

You do not need to memorize every product term before getting started. This guide gives you a practical map of the main parts of Zeltino, what each one is responsible for, and how information moves from a website conversation to your team.

In this article, you'll learn:

• How workspaces and projects organize your account
• How a visitor conversation reaches your team
• How knowledge and behaviour work together
• Which automations guide visitors toward a next step
• How conversations connect to leads, tickets, and bookings
• How availability and meeting types support booking
• How to use analytics without losing the conversation context

Workspaces and projects#

Think of a workspace as the home for a team, and a project as the customer experience configured for a particular website or use case. Your account can belong to more than one workspace, and your role is assigned separately in each one.

• A workspace contains members, roles, workspace notification policy, usage, and owner-level plan access.
• A project contains the assistant experience for a website: knowledge sources, widget settings and domains, behaviour, automations, conversations, leads, tickets, meeting types, bookings, and analytics.

The active workspace and project appear in the dashboard path and navigation. Check both before changing configuration or reviewing customer data. If a setting or record seems to be missing, the selected project is the first thing to check.

The visitor-to-team lifecycle#

Once the widget is installed, a conversation can connect several product areas without the visitor needing to understand any of them. A typical interaction moves through Zeltino in this order:

1. A visitor opens the chat widget on a verified, widget-enabled domain.
2. The assistant receives the visitor's message and the current page context.
3. It uses the project's available knowledge and behaviour settings to prepare a reply.
4. A suggested question, contact option, or chat flow can guide the visitor through an additional step.
5. The interaction appears in Conversations.
6. Depending on the configured path, Zeltino can also create or connect a lead, ticket, or booking.
7. Your team reviews the resulting records and uses analytics to identify what should improve.

Not every conversation needs to become a lead, ticket, or booking. Those records are created only when the visitor completes the relevant configured action. This lets ordinary questions remain simple while more valuable or complex conversations can continue into a structured team process.

Knowledge contains business facts#

Knowledge is where the assistant looks for approved business information. If a visitor asks about delivery, pricing, eligibility, or a policy, the answer should come from a current source rather than from a behaviour instruction or a flow label.

Source type	Use it for
Website source	Current public pages such as product information, help content, policies, or pricing explanations
Text source	Private business context, support scripts, or focused information that is not published as a webpage

A source hint tells the assistant when the source is relevant. Zeltino uses imported or saved source content; it does not search the public internet live for each visitor message.

Keep important facts in knowledge even when you also mention them in a flow message. This gives ordinary questions and automated journeys a consistent source of truth. When the fact changes, you know which source to update.

Behaviour controls communication#

Knowledge answers “what is true?” Behaviour answers “how should the assistant communicate it?” Keeping those responsibilities separate makes both easier to maintain.

• Voice style controls the overall conversational tone.
• Emoji use controls whether occasional emojis are allowed.
• Excluded subjects define topics the assistant should refuse or redirect away from.
• Behaviour layers add ordered, style-only instructions.

Behaviour layers do not add business facts, trigger forms or flows, or replace product safety controls. If a reply contains the wrong policy, update the knowledge. If the policy is correct but the answer sounds wrong for your brand, update the behaviour.

Automations guide the next step#

Automations are useful when a conversation should lead somewhere specific. They can offer a helpful starting question, collect details, route the conversation, or let a visitor book time with the team.

Feature	Purpose
Suggested question	Offers a short, clickable question before the visitor types or throughout the conversation
Contact option	Offers callback, meeting-scheduling, or enquiry paths when their required form or meeting type is ready
Lead form	Collects the contact details and consent fields you configure
Chat flow	Runs a triggered sequence of messages, questions, branches, lead capture, meeting booking, category assignment, or ticket creation
Smart greeting	Uses a proactive-chat trigger to start a page- and visitor-aware conversation

An automation can depend on another project asset. For example, a meeting action needs an eligible public meeting type, and ticket creation needs an email captured earlier on every route that reaches the action. These relationships are why it is important to review active automations whenever you pause, replace, or delete an asset.

Customer work appears in connected records#

A single visitor journey can produce several connected records, but each record has a different job for your team:

• A conversation contains the visitor message history, internal comments, context, category evidence, captured flow details, and linked outcomes.
• A lead contains contact details, submitted form values, relevant flow answers, qualification status, groups, and links back to the conversation or booking.
• A ticket tracks a customer request through open, waiting-for-customer, and closed work, with customer replies, internal notes, ownership, priority, category, and due date.
• A booking records a scheduled meeting and can connect to the relevant lead and conversation.

Changing one record's working state does not delete the others. For example, archiving a conversation is different from archiving a lead or closing a ticket. When investigating a customer journey, follow the available links rather than expecting every detail to live on one page.

Availability and meeting types control booking#

Booking has two layers. Availability answers when a host can be booked, including weekly windows and date overrides. A meeting type describes what the visitor is booking, including duration, buffers, notice, booking horizon, location, booking-page content, visibility, questions, and either one-on-one or group capacity.

Zeltino's current booking experience uses the availability configured in Zeltino. A public and active meeting type can be shared directly or offered through a chat flow or scheduling contact option when it meets the required eligibility rules.

Analytics helps you decide what to inspect#

Analytics summarizes widget engagement, AI reply delivery, chat-flow progress, leads, goals, and page activity over the selected time range. It is a useful starting point when you want to know where the customer journey may need attention.

A metric can show that an outcome changed, but it does not by itself explain why. Compare the relevant visitor journey, source content, and configuration before drawing a conclusion.

Continue with a focused setup#

Now that the product map is familiar, you can build one small journey without needing to configure everything at once. Follow Quick setup to configure one knowledge source, a saved assistant experience, a suggested question, and a website domain. Then use Test the complete visitor experience to verify the visitor and team results.