User Guide
eva 3.4.1
eva 3.4.1
  • Home and Support
  • What's new
  • Changelog
  • Using eva
    • Overview
      • Virtual Agents
      • Main Concepts of eva
      • Glossary
    • Developing virtual agents
      • Quick Start with eva
      • Dialog Flows
      • Dialog Cells
        • Intent Cells
        • Entity Cells
        • Answer Cells
        • Service Cells
        • Input Cells
        • Jump Cells
        • End Cells
        • Code Cells
        • Rule Cells
      • Training task
      • Export and import agents
      • Create virtual agents from templates
      • Create and manage users
    • Channels
      • Integrating Existing Channels
    • Testing Virtual Agents
      • Test a virtual agent
      • Automated Test
    • Natural Language
      • eva NLP
      • Training with eva NLP
      • Training with eva Automated Learning
      • Using other NLP engines
    • Analytics
      • External analytics platforms
    • Experience
      • Context Management
    • Advanced Options
      • Parameters
      • Intent Navigator
      • PII Data masking
    • How Tos
      • Creating variable answers with Code and Rule cells
      • Videos
  • for technicians
    • eva server Installation guide
    • API Guidelines
      • Conversation API
      • Cloner API
      • Webhooks
    • Supported/verified third-party software
    • Appendices
    • Data Extraction
Powered by GitBook
On this page
  • A step-by-step guide to build your first virtual agent
  • Login into the eva platform using your account
  • Create your first virtual agent
  • Configure the NLP engine and the language
  • Create the main channel
  • Design and Create your dialogs flows
  • Training the agent language model
  • Test your virtual agent
  1. Using eva
  2. Developing virtual agents

Quick Start with eva

PreviousDeveloping virtual agentsNextDialog Flows

Last updated 2 years ago

Watch the video below and follow this step-by-step guide. You'll find out how easy and simple it is to create a virtual agent with eva!

A step-by-step guide to build your first virtual agent

Login into the eva platform using your account

The first thing to do is to set up an account. After the sign-up process, insert your email and password to log in.

Request access to an eva environment

Create your first virtual agent

Then, you'll see a new screen where you can create and manage your virtual agent. Click on the button Create bot to start.

From there, you'll be redirected to the screen Create new bot. Choose a name for your virtual agent on the field "Name". You can also upload an image to make your desktop more visually appealing (but it's not mandatory, you can do that anytime).

It is also possible to create a bot from an exported virtual agent. For more information on exporting and importing, please visit:

Configure the NLP engine and the language

Now, you will select an NLP engine of your choice and the language of your virtual agent. eva provides its own NLP engine and also allows the use of the main NLP engines on the market:

  • eva NLP. Internal NLP engine. More information about training eva NLP in this section:

  • IBM Watson.

  • Microsoft Luis

  • Google Dialogflow

  • Microsoft QnA Maker

After that, you have to select the language. Different NLPs support several languages. For more information:

Create the main channel

Channel is the connection between communication applications and the virtual agent. In other words, it'll be the platform where the interaction with the user will take place (Whatsapp, Web, Facebook, etc.). eva allows you to connect with 32 different options! Now select your main channel. You can add more channels anytime.

Take the following steps to choose a channel:

  1. When you click “Select a Channel”, a dropdown with the 32 channels available in eva will appear.

  2. After you select a channel, you'll be asked to add a name and a description for it (as seen below).

3. Every channel must have a name, but it is not mandatory to have a description (it's good to guide your team and helps new members to learn more about the virtual agent and the project).

4. After selecting the first Channel, you will be redirected to the virtual agent menu screen.

Important:

  • The Channel name is very important! Your virtual agent might malfunction if you rename the channel. And you cannot share a name either.

  • Avoid giving the channel the same name as the platform calling it, for example, Facebook Messanger, Web, Google Home, etc… It can confuse later when you create another virtual agent and try to use the same channel. Tip: give it a name linked to the project and its description.

Very Important:

After choosing the main channel, you will be able to go forward and start to create a dialogue. You can also integrate more channels for the same virtual agent, but you can do this whenever you want.

Design and Create your dialogs flows

Once the virtual agent has been created, we can start creating conversation flows. In the Workplace, there are different types of flows that we can create: Welcome, Not Expected, User Journey, and Jump flows. Two of them are mandatory, the Welcome and the Not Expected. The rest of the flows of our agent are the ones we create to cover our use cases.

Learn more about dialogs in the following section:

The Welcome Flow

If you have started with Welcome, you'll see a tab on the right corner. Type Welcome on the field "Name", register the tag Welcome, and click on the "Save and View Flow" button.

Add cell

On the left side, you'll see the plus icon + to "Add Cell", as seen on the screen below, and then click on the "Welcome" button.

Write an Answer (and add a menu if you want)

A tab on the right corner where you can write your message will appear.

If you use the "Option" field, you can add a menu on your virtual agent. Depending on the channel of your choice, it may come in the form of text (such as WhatsApp) or buttons (such as web and Facebook).

On the "Value" field, you can register an Intent example, a Synonym Entity value or an expression predicted by a pattern entity. And then, press enter to register your buttons. After writing your text and your possible buttons, just click "Save".

User Journey

Now, it's time to improve your flow. To do that, go back to the Flow tab, and click on "Create Flow" button. When you select "User Journey", the tab "Create user journey flow" opens on the right. Name your flow (Remember: use _ or - instead of spaces).

Create intents

After that, you will be redirected to the Workspace's canvas. Press the + button to see the table, "You can create an intent or choose one from the list". Click "New" and then, a Create Intent tab will appear on the right side.

Name your intent

Name your Intent. Never use spaces! Instead, you must use _ or - to connect 2 (two) words. For example, Check_Balance.

In the "Add example / Utterance" lines, you should register phrases that your user would use to express an Intent. For example, in the Check Balance, it could be expressions such as "Check Balance", "I need to know how much I have in my account", "I'd like to know my balance", etc. Just press enter to register an Utterance.

More about Intent Cells in the following section:

Tip: By registering the utterances, it is also worth predicting possible typos or grammatical mistakes of the user

Next cell

After clicking on the "Save" button, you will be redirected again to the Workspace. By hovering the mouse on top of the newly created cell, you will see 3 (three) icons: + (to create), pen (to edit), and a trash can (to delete).

Name your answer

A second pop-up will emerge saying that "You can create an answer or choose one from the list". Click on the button "New" and you will see a tab on the right side. Name your Answer and click "Next".

Write your answer

Write your answer and click "Save". As explained above, you can add buttons in the answer cells.

More about Answer Cells in the following section:

Tip: It's a good practice to use the exact same name for Intents and Answers

Continue your flow

Important: Building flows with more than 100 (one hundred) cells can slow down the system. Avoid it.

Training the agent language model

On the Training tab, click on the "Train" button and try it out on eva's icon (in the bottom right corner of the page).

More information about training you agent using NLP:

And training using external NLP engines:

Very Important: If you are using eva NLP (NTT DATA proprietary Natural Language Processing engine), you must register at least 5 (five) Intents, each one with at least 5 (five) examples/utterances to start off your virtual agent.

Test your virtual agent

Finally, it's always a good practice to test your virtual agent from time to time. Just click the eva icon in the bottom right corner of the page. And then, you'll get to your virtual agent simulator.

More information about testing you agent here:

In case you do not have access to an eva environment and would like to test the platform, please contact the presales team using the web form at the following .

eva platform landing page
The main page of eva with the list of all agents to which you have access
New agent creation page

IBM Watson

Microsoft Luis

Google Dialogflow

Name your Channel screen

Tip: If this is your first time using eva, you may find it easier to start with Welcome. It's good to point out this is not a mandatory flow and will depend on the project strategy

Tip: You can use emojis in eva

To do that, copy and paste emojis from

By clicking on the + icon, you'll see a pop-up with the question, "Which is the next cell?" It could be , , or using more advanced ones, such as , , , , and.

Tip: If this is your first time in eva, it would be easier if you predict an Answer cell to dialog with the newly registered Intent

You can continue your flow, adding cells such as , , or using more advanced ones, such as , , , and .

😉
😍
😉
link
Export and import agents
Training with eva NLP
eva NLP
Here
Here
Here
Dialog Flows
Get Emoji
Intent Cells
Entities
Answers
Jump
Service
Input
Rule
Code
Answer Cells
Entities
Answers
Jump
Service
Input
Rule
Code
Training with eva NLP
Using other NLP engines
Testing Virtual Agents
😉
Basic cells: Intents, Entities and Answers
Advanced cells: Jump, Service, Input, Rule and Code
Click on the eva figure
And then, you'll get to the virtual agent simulator