Table of Contents
Bot Builder User Interface (UI)
Constructing the Conversation Flow
Customizing Candidate Interactions
Create a New Chatbot
To begin creating a chatbot, navigate to Automation > Chatbots section.
- Click Create New Bot + to begin setting up a new chatbot.
- When creating a Chatbot 2.0 Flow, start by selecting the use case. Each card represents a chatbot type based on its primary function:
- Data Enrichment: Engage with candidates from your database to update or reactivate their profiles.
- Pre-screening: Build a flow to evaluate candidates who have applied for a job.
- Sourcing: Create an embedded chatbot flow to attract candidates, typically hosted on your careers page or website.
- Application: Allow candidates to apply for specific job postings directly through a chatbot interface.
Note: If you're a new user or your agency hasn’t enabled all features, you may see only a limited set of chatbot options.
- Select the Entity Type (for Outbound Bots only). If you select Data Enrichment or Pre-screening, you’ll be prompted to choose an Entity Type. This defines what kind of records the chatbot will work with and is essential for mapping the bot to your database.
Common options include:- Candidate
- Internal Employee
Note: For Sourcing and Application bots, the entity type is fixed.
- Sourcing bots always target Candidates.
- Application bots are linked to the Submittal entity type, which tracks job applications.
- Click Next to launch the Bot Builder interface and start designing your flow.
Flow-Level Chatbot Settings
You can edit the chatbot name by clicking the pencil icon next to it.
The Chatbot Settings are specific to individual chatbot flows and allow for more granular control over each chatbot's behavior and integration.
- Chatbot Name: Specific to each flow, allowing unique identification.
- Brand Selection: The selected brand dictates the chatbot's appearance, as brand colors will be applied automatically.
- Description: A description can be added for each chatbot flow.
- UTM Parameters: A new feature released recently, allowing users to set up UTM parameters for inbound and application chatbots. This helps in tracking and analytics.
- Outbound Chatbot Configuration: For outbound use cases, the settings primarily involve the chatbot name, brand selection, and description. Outbound chatbots do not support as many channels as inbound chatbots currently.
Branding Considerations
Chatbot settings are closely tied to a brand's identity. When a brand is selected, the chatbot automatically adopts the specified brand colors, ensuring a consistent brand experience. The chatbot settings page includes configuration options for branding and customization.
Three-dots Menu
Below are the actions you can perform from the three-dots menu:
Duplicate
Creates an exact copy of the current flow, automatically redirecting you to the new workflow for editing.
Archive
Allows you to retire a flow from the active dashboard. Archiving typically preserves the data without deleting it, allowing you to declutter your workspace while maintaining a record of past versions.
Export to .png
The Export to .png feature triggers a browser download of the current workflow diagram, allowing for offline review or team distribution.
Manage Reminders
Allows you to configure automated nudges or follow-up notifications. You can refer to this article for more details.
Manage FAQ
The Manage FAQ tool allows you to integrate a knowledge base directly into a specific workflow to address candidate inquiries.
To use this, you must toggle Show FAQ and select a pre-existing FAQ set from your Library.
If a specific FAQ is missing, the Edit in Library option redirects you to the main conversation flows page to update your global FAQ database.
Key Notes:
- If you enable the FAQ button functionality, the candidates will be able to see the FAQ button with your top 3 default questions.
- If you disable the FAQ button functionality, the candidates will not be able to see the FAQ button, but will still be able to ask their questions. The system will automatically provide the answers (if added to the global library as mentioned in step A2). For Example: {FAQ Button Enabled}
- You can manage your default questions from the drop-down menu for the selected conversation flow.
Manage Live Chat
Live Chat facilitates a seamless transfer from the bot to an internal team member via the Messaging platform.
You can choose to Show Live Chat in Chatbot UI, which places a dedicated button at the top of the chat interface, or customize the Menu option label for the web chat's overflow menu.
For Live Chat to function, you must create a separate Live Chat workflow and link it within the configuration box, unless the bot’s sole purpose is the Live Chat Node. Once configured, specific routing rules define how the transfer to an agent occurs.
Bot Builder User Interface (UI)
The bot builder UI is where you configure the complete flow of your chatbot.
-
Default Nodes: Every new chatbot flow starts with two default, essential nodes:
- Greeting Node: This is the first node where you greet the candidate. You can configure the greeting message, add ATS variables (like <First Name>), and set the initial question and response type.
- End Node: This is the last node, marking the end of the conversation flow. An optional "Feedback" can be enabled or disabled within the end node to ask the candidate or the person communicating with the bot how the conversation was. You can edit the feedback question and preview the rating scale.
Note: These default nodes cannot be removed, but feedback can be disabled.
Conversation Builder
The Conversation Builder is a specialized flowcharting tool designed to automate candidate interactions and guide them through various stages of the recruitment funnel. The tool is highly customizable, allowing users to automate frequent conversations by creating a structured sequence of messages and questions.
Constructing the Conversation Flow
The architecture of a chatbot conversation is built through a series of nodes.
• Initiation: Conversations generally begin with a greeting to set expectations for the candidate.
• Adding Elements: Users can click the ‘+’ button to insert new nodes, which include questions, messages, or instructions to jump to a different branch.
• Management: Existing nodes can be edited or removed by selecting the three dots in the upper-right corner of the specific box.
• Logic and Branching: Conversations can be structured to diverge into different branches based on candidate responses. For example, a "Yes" response can trigger a different path than a "No" response. To navigate candidates between these branches, the Jump-to option is utilized.
Customizing Candidate Interactions
To ensure data is captured correctly, the Conversation Builder allows for specific response types for every question:
• Yes/No answers.
• Single-select or multi-select buttons.
• Free-form answers for open-ended input.
Finalizing and Wrapping Up
A conversation must be logically concluded to provide a professional experience.
• Closing Messages: Every branch must end in a closing message. This can include a thank-you note, a request for the best time to connect with a recruiter, or a meeting node to schedule an appointment on a recruiter’s calendar.
• The "Sad-Path": In pre-screening scenarios where a candidate may not qualify, it is recommended to include a consolation message at the end of that specific branch.
• Technical Constraint: System logic dictates that the last node in a sequence cannot be a question; a warning statement will appear in the first node to remind users of this requirement.
Review and Deployment
Before the chatbot is live, users must perform a quality check for typos, grammar, and logical flow.
- Preview: You can preview your chatbot via different channels, such as web and mobile, before building.
- Build: Once the configured flow is verified, click the Build button in the top right-hand corner to initiate the back-end creation of the bot.
- Reminder ON/OFF: Indicates if the reminders are on or off.
- Status Check: Users should monitor the "status" in the main conversations tab.
- Activation: When the status changes to "ready for use," the chatbot can be attached to a journey/workflow for outbound or integrated on a website for inbound to begin interacting with candidates.
The following table provides a comprehensive quality checklist for designing chatbot conversations, ensuring they are professional, clear, and logically sound before you hit the "build" button.
| Category | Quality Requirement | Best Practices & Recommended Lengths |
| Content Quality | Grammar & Spelling | Ensure the conversation is free of spelling, grammatical, and orthographic errors. Use a style guide (Microsoft or AP) for consistency. |
| Hyphenation | Always hyphenate compound adjectives such as "full-time job," "temp-to-hire position," and "contract-to-hire role.” | |
| Tone & Voice | Maintain a consistent voice aligned with the audience and goals. Avoid recruiting jargon that might confuse candidates. | |
| Conversation Length | Data Enrichment | SMS: 3–5 questions ideal. <br> Webchat: 5–7 questions ideal. |
| Pre-screening | SMS: Aim for a maximum of 7–10 questions. <br> Webchat: No maximum, but be mindful of user bandwidth. | |
| Node Structure | Clarity of Copy | Use only one question per node unless they are highly dependent. |
| Requirement Phrasing | For filter questions, clearly indicate whether a response is a requirement (e.g., "Being available to work overtime is required for this role") or a preference. | |
| Logic & Flow | Jump-to Logic | Verify that every "jump-to" directs the candidate to the correct intended location. |
| Sequencing | Place authorizations and legal questions (e.g., drug screens, work authorization) at the very end of the conversation. |
Key Execution Tips
• Drafting: Draft your conversation in a word processor first to utilize advanced spelling and grammar checkers before inputting copy into the designer.
• Button Review: Double-check all buttons, select options, and error messages for hyphenation or unnecessary capitalization.
• The "Build" Rule: Always review this checklist before finalizing, as any changes after hitting "build" may require creating an entirely new bot.