Understanding Gnatta

Quick links in this article:

Gnatta is flexible — which means it doesn’t always behave like other systems you’ve used before. This page explains how the main parts of the platform work together so you can get a clearer picture of what’s happening behind the scenes.

Conversations vs Interactions

Open

Gnatta groups conversations across channels into a single interaction, ready to assign to an agent

Conversations = message threads

• A conversation is a single thread on a single channel (e.g. an email chain, WhatsApp chat, or Facebook DM conversation)

• It includes all replies within that thread

Interactions = the bigger picture

• An interaction is what your agents work from — it may include multiple conversations across different channels

• Think of it like a folder that contains all messages about a specific issue, plus key info, notes, and data fields

What is a Workflow?

When something happens in Gnatta — like a message coming in, or a survey being completed — a Workflow tells the system what to do next.

But what is a workflow, really?

Workflow vs Flow

Gnatta treats these as two separate things:

• Workflow = the big picture — a set of flows triggered by an event

• Flow = a series of specific actions (like “Create Interaction” or “Queue this message”) that make something happen

The difference between a Workflow, and Flows

Workflow: A workflow is triggered by an event, and contains a series of flows. A workflow is the ‘bigger picture’ of connected flows that work together to react to the event that was triggered.

Open

Example Workflow for inbound messages to an email address

Flow: A flow is a micro-series of actions that can be visualised as a ‘map’ in the builder. These actions (like Create Interaction or Connect Conversation) are what you’ll use to make changes and decisions about how to handle the event at hand. More about actions here.

Open

A Junk Checker flow we’ve used on our email inbox

Think of a Workflow as your “recipe” — and each Flow as a step in that recipe.

One Workflow event might contain:

1. A Find or Create flow to check if this message is part of an existing interaction

2. A Detect and Update flow to pull out key data

3. A Queue Interaction flow to route it to the right team

This modular approach means you can reuse flows across multiple workflows — and scale easily without rebuilding from scratch.

Workflow Actions: Your Building Blocks

When you're building automations in Gnatta, everything starts with actions.

Open

Some Actions in a flow!

Each flow is made up of individual actions — like sending a message, checking a data field, or routing to a queue. These are your “Lego bricks” for automating customer journeys.

Some examples:

• Create Interaction → Start a case when a new message comes in

• Decision → Build simple if/then rules to split your flows

• Send Autoresponse → Instantly reply to customers with buttons or prompts

• Update Interaction → Move to a different queue, or save a piece of data

• Add Note → Leave internal notes or AI summaries for the agent

• Detect Intent → Use AI to figure out what the customer’s asking for

• Detect Sentiment → Spot frustrated customers and fast-track them

• Translate → Show agents an English version of the original message

• Suggest Response → Let AI preload a reply for the agent to edit and send

• Start Timer / End Timer → Automate follow-ups or delays (e.g. “Nudge in 24 hours”)

There are dozens of actions available — and you can stack them together however you need.

🔗 Explore all standard actions in Gnatta: Actions
🔗 Get to know AI Workmate - a power-pack of AI Actions: AI Workmate Actions

Queues and the Agent Radar

What is a queue?

A queue is a list of interactions waiting to be handled. You decide how many queues to use, how they’re organised, and what rules send interactions into each one.

Agents don’t pick work manually — instead, Gnatta automatically assigns them the right interaction from the right queue, based on your logic. This is what we call the Agent Radar: a real-time feed of work matched to that agent’s skillset or team.

Open

You can see this agent’s radar on the left - they’ve been assigned 5 interactions, automatically.

Routing is your decision

You choose how and why messages are routed - not just where they came from. While you can use simple routing (e.g. all email → Emails queue), that’s just the starting point. Most teams benefit from more intelligent logic.

🔗 Create A Routing Plan

Important: Not all queues should mix

To keep things efficient:

• Don’t put live and non-live channels in the same queue (e.g. chat + email)

• Be mindful of queue priority: everything in a single queue is treated equally, regardless of channel or urgency. To get urgent queries handled first, that queue should be assigned a higher priority.

🔗 Queue Setup

How data fits into Gnatta

Data is one of the most powerful tools in your Gnatta setup.

You can:

• Show agents relevant info (like customer ID or product) directly on the interaction

• Route queries based on data fields — like account type, Trustpilot rating, or subscription status

• Store values automatically using RegEx or AI, so your team doesn’t have to enter them manually

• Report on patterns like refund reasons, complaint types, or delivery delays

🧠 Where does this data come from?

Gnatta automatically detects a huge amount of contextual data — such as sender, channel, subject line, and message content — every time something happens.

You can also set up your own custom fields, or populate them dynamically through workflows, AI extraction, or integrations.

Optional follow-on links:

• 🔗 Contextual Data

• 🔗 Defining Custom Data

• 🔗 Workmate Demo: AI-Driven Reporting

• 🔗 Updating Data Fields

Want to see it in action?

Check out these quick examples in our Demo Video Library →