interaction.website automations are interactions with visitors on third-party websites that use continuations to pause and resume a multi-step workflow.
On websites, this can be used for surveys, sign-up forms, contact forms, troubleshooters, customer service bots, and much more.
Website interactions are usually started on any website when a page element is interacted with – a link, button, image, etc.
A website interaction can also be started with a shared link.
At its conclusion, an interaction returns a dictionary and exit state to the caller, which is then responsible for acting on the results.
Inputs
An interaction automation dictionary starts with the following input values:
| Key | Type | Notes |
|---|---|---|
interaction |
string | The name of the interaction. |
interaction_params |
dictionary | Arbitrary interaction parameters. |
inputs |
dictionary | Custom input values from the caller. |
client_browser_name |
string | The client browser name (e.g. Safari). |
client_browser_platform |
string | The client browser platform (e.g. Macintosh). |
client_browser_version |
string | The client browser version. |
client_ip |
string | The client IP address. |
portal |
record | The community portal. |
Outputs
await:form:
When suspending in the await state, the interaction displays a web form with the desired elements. The form may prompt for user input, validate it, and set dictionary keys (placeholders) with the responses.
await:
form:
title: Your form title
elements:
# ...
title:
The title of this form to be displayed in the interaction popup. This is usually a summary of the current step.
elements:
Form elements are defined with a key in the format type/name:.
The name must be unique within the form. When an element prompts for user input, a placeholder with the same name will be created with their response. For instance, text/prompt_name: will create a placeholder of {{prompt_name}} with the value of that text element.
A form can be created with any combination of the following element types:
| Element | |
|---|---|
| fileUpload: | File upload prompt |
| say: | Block of text or Markdown |
| sheet: | Sheet with row selection |
| submit: | Continue to next step |
| text: | Text input with data types |
| textarea: | Multiple lines of text |
When the interaction suspends in the await state, a submit: element is automatically appended to the form if one doesn’t already exist.
start:
await/who:
form:
title: Introduction
elements:
text/prompt_name:
label: What is your name?
required@bool: yes
await/hello:
form:
title: Hello!
elements:
say/hello:
content: Hello, {{prompt_name}}!
return:
user:
name@key: prompt_name
await:interaction:
When suspending in the await:interaction: state, the interaction temporarily hands control to another delegate interaction. The interaction resumes at the current point when the delegate exits.
Delegates can be nested to any depth. For instance, a reusable delegate could handle email or SMS validation, and be shared by many other interactions.
This makes interactions much more modular and reusable.
uri:
The uri: parameter specifies the delegate automation. This must use the interaction.website trigger.
output:
An output: key specifies the placeholder that should receive the results from the delegate.
start:
while:
if@bool: yes
do:
await/menu:
form:
title: Menu
elements:
say:
message: How can we help?
sheet/prompt_menu:
required@bool: yes
data:
0:
key: map
label: Map
1:
key: echo
label: Echo
schema:
layout:
style: buttons
headings@bool: no
paging@bool: no
title_column: label
columns:
selection/key:
params:
mode: single
text/label:
submit:
continue@bool: no
reset@bool: no
await/do:
interaction:
output: results
uri@text:
cerb:automation:{{{
'map': 'wgm.interaction.locationByIP',
'echo': 'wgm.interaction.echo',
}[prompt_menu]}}
return:
When the interaction concludes in the return state, it returns any number of key/value pairs to the caller. Keys may be nested to return dictionaries.