Myma.ai's Conversation Flows feature empowers businesses to build intelligent, guided experiences for guests across multiple channels. Whether your goal is to route inquiries, collect information, or automate booking-related tasks, Conversation Flows offer a structured way to guide users through predefined or dynamic interactions.

This guide introduces the two types of flows:

This guide is mostly focused on Dynamic Conversation Flows as they are used the most, almost in every scenario.

Access this feature by navigating to Workflows→ Conversation Flows in your Myma.ai Dashboard.

Under Conversation Flows, you’ll see two buttons at the top: Add a Conversation Flow and Import from Templates.

When you select this option, you can choose the flow type (Dynamic or Static), enter a description, and click Save. From there, you’ll be able to add questions and answers that the chatbot can reference.

Selecting this option opens a page where you can choose from existing Dynamic Conversation Flow templates to use as a starting point.

Static conversation flows are predefined, rule-based, button-driven interaction paths used when you need controlled, guided filtering—such as narrowing down room types when multiple options exist or helping users select the right product without relying on dynamic AI interpretation.

These features are helpful when there is typically a predefined set of responses to user queries, or when users can select from a series of questions. Once a user’s query is resolved, if a redirect to a specific page is necessary, a button is provided for this purpose. This streamlined process enables users to easily navigate the conversation and access relevant information without any difficulty.

You can configure a button click to do one of the following:

Use Case: Guiding users through a product selection process.

A conversation flow in relation to Gen AI is the structured sequence of prompts, user inputs, system instructions, and model responses that guides how a generative AI interaction progresses from start to outcome.

Dynamic flows allow you to create personalized, adaptive guest experiences by combining natural language instructions, training data, and optional API integrations. Unlike static flows, dynamic flows adapt in real time based on guest input, making conversations more flexible, contextual, and intelligent.

You will define a conversation flow based on the channel as behavior for chat responses, voice conversation and email drafting will be different. Each needs its own styling, tone and restructions. You will usually include following in the conversation flows:

To improve usability and safety, while minimizing the risk of dynamic flows causing unintended issues, this screen is organized into structured sections:

Assistant Instructions define how the AI behaves, communicates, and operates within the flow. Instructions are separated them into four distinct sections to make it easier to add, remove, or troubleshoot specific instructions and to create a more standardized and streamlined configuration process.

The four sections are outlined below.

Pre-built blocks are reusable conversation components that combine prompts, logic, data collection, and external API triggers—with built-in instructions on when and how to invoke them—to quickly build structured Gen AI workflows.

Example: Get Rates & Availability.

Each Pre-Built Block includes:

When you select Browse Library, you will see a collection of Pre-Built Blocks to select from.

When a Pre-Built Block (such as “Room Rates & Availability”) is added, it automatically includes:

Following Pre-Built Tools are available:

Lastly, at the bottom of the Dynamic Conversation Flow box, you will see a Change Log. Clicking this allows you to view a history of updates made to the conversation flow, including when changes were made and what the previous version contained. This is helpful for tracking revisions or reviewing earlier configurations.

It is important to understand the structure and limitations of conversation flows to ensure a successful configuration that aligns with client needs.

Dynamic conversation flows create a more natural, conversational experience. However, responses may vary each time a question is asked. While conversation flows establish strong parameters, answers are not always delivered verbatim due to multiple influencing factors and the inherent variability of generative AI.

When structuring conversation flow, keep in mind that users come to chatbots for quick, direct answers. Most users will ask something short and clear, like: “How deep is the pool?”, “Can I book a tee time in advance?”, "I want a copy of my reservation confirmation email", "Do you have availability from 1st April to 3rd for 2 adults?"

In these examples answers will be provided based on the FAQ's and other information you added to Knowledge Hub. Only add instructions where you want conversation to go further from just answering the question.

Here is an example of chat conversation flow used for website chatbot, and other messaging channels like WhatsApp, Facebook Messenger and Instagram:

Establishing conversation flow structure helps guide the chatbot to deliver responses in the intended sequence. However, as scenarios become more complex, there may be limitations in how consistently the chatbot can follow that structure. While the chatbot generally adheres to defined conversation flows, it may not follow them exactly in every instance.

Only add what is necessary.

Use FAQs for standalone, factual question-and-answer responses, and use conversation flows when you need guided, multi-step interactions that collect information or drive a specific outcome.

Besides conversation flows it is important that Knowledge Hub is maintained properly as most of the contextual information is sourced from Knowledge Hub; however, there are situations where the data may be incomplete, require updates, or lack sufficient detail. In these cases, the following hierarchy should be followed:

The preferred first step is always to update information directly within the SynXis CRS, as it serves as the primary source of truth. Within SynXis, content is organized by categories (e.g., Wellness Facilities). When adding or updating information, it is recommended to place it within the appropriate category to maintain consistency and organization.

If the information does not clearly fit within an existing SynXis category, it is best to add it as a manual FAQ. This ensures the information is still accessible without disrupting the overall data structure.

As an additional option, hotels can update their website to include missing information, which can then be pulled in as an external source. However, this should be considered a last resort compared to updating SynXis or adding an FAQ.

Sharing URLs in response to a question is a common requirement. Add the following instructions in Conversation Handling to ensure consistent link usage across all responses:

Some features are only available to Myma Super Admins, which include the following:

Model Type refers to the AI model being used for the assistant. This setting should not be changed, as it is managed directly by the Myma team.

Include Knowledge, however, is important to enable. When turned on, it ensures that all relevant information within the Dynamic Conversation Flow is incorporated into the assistant’s responses, allowing it to provide accurate and contextually informed answers.

Custom Actions allow you to complete non-call-related actions through external APIs, such as pulling up a form or triggering a third-party integration. These provide extended flexibility beyond standard pre-built actions and enable deeper system connectivity where required.

Conversation Flows are a powerful way to guide guest interactions with purpose, structure, and clarity. Whether you’re simplifying straightforward tasks or designing complex, multi-step experiences, Myma.ai provides the flexibility to tailor flows to virtually any scenario. Use this guide as your foundation for building AI-driven experiences that engage guests, deliver meaningful information, and drive conversions.

This updated structure creates a clear separation between conversational logic and operational actions, making dynamic flows easier to configure, scale, and manage over time.