Skip to main content

Bot Flows

Section Description
Feature Area Architect / AI & Bots
Navigation Admin → Architect → select flow type from the flow list
Primary Function Build native AI-powered bots that automate customer conversations before routing to a live agent
Flow Types Dialog Engine Bot Flow (voice + digital) · Digital Bot Flow (digital only)

Genesys Cloud offers two native bot flow types built directly inside Architect. Both use Natural Language Understanding (NLU) to interpret customer input and guide conversations. The key distinction is the channel scope and PCI compliance status.


Study Notes

Topic Explanation
Dialog Engine Bot Flow Native bot for voice, chat, and message channels. PCI DSS-compliant — can be used in secure call flows
Digital Bot Flow Native bot for digital/messaging channels only (chat, messaging). Not PCI DSS-compliant — cannot be used in secure call flows
Intent A customer goal or request the bot is trained to recognize (e.g., "Check Balance", "Cancel Order")
Utterance A sample phrase the customer might say to express an intent — used to train the NLU model
Slot A piece of information the bot needs to extract from the conversation (e.g., account number, date)
Slot Type Defines the format/type of a slot: built-in (e.g., date, number), custom list, regex, dynamic list, or AI-powered
Confirmation A step where the bot confirms captured slot values with the customer before proceeding
Learning The bot reviews unrecognized utterances and suggests additions to improve NLU over time
Intent Health Dashboard that shows how well intents are performing and highlights training gaps
Optimization Dashboard Per-flow dashboard showing total interactions, average duration, average turns, and end states
NLU Natural Language Understanding — the AI model that maps customer input to intents and slots
Call Bot Flow action Architect action used in an Inbound Call, Chat, or Message Flow to invoke a Dialog Engine Bot Flow
Call Digital Bot Flow action Architect action used in a Message Flow to invoke a Digital Bot Flow
Virtual Agent Advanced AI bot powered by Genesys AI — generates intents and utterances from descriptions
Intent Miner Analyzes transcripts/recordings to discover real customer intents that can be imported into a bot
Knowledge Integration Bots can query a Knowledge Base to answer customer questions automatically
Rich Media (Digital) Digital Bot Flows support quick replies, cards, and carousels for structured customer choices

Bot Flow Type Comparison (Exam Critical)

Attribute Dialog Engine Bot Flow Digital Bot Flow
Channels Voice, Chat, Message Digital (Chat, Message) only
PCI DSS Compliant Yes — can be used in secure call flows No — must not be used in secure call flows
Used In (Architect) Inbound Call, Chat, or Message flows via Call Bot Flow action Message flows via Call Digital Bot Flow action
Pricing — Voice Per minute (15-second increments) N/A
Pricing — Digital Per session Per session
Rich Media Quick replies, cards, carousels (on messaging) Quick replies, cards, carousels
Knowledge Base Yes Yes
Virtual Agent Yes Yes
Intent Miner Yes Yes
DTMF Input Yes (voice) N/A

Permissions

Permission Purpose
Architect > UI > View Access Architect
Architect > Flow > Add Create bot flows
Architect > Flow > Edit Edit bot flows
Architect > Flow > Delete Delete bot flows
Language Understanding > All Required for NLU/intent management in bot flows

For Virtual Agent specifically: Architect > virtualAgentFlow > Edit


Navigation

Task Path
Open Architect Admin → Architect (opens in separate window)
Create a Dialog Engine Bot Flow Architect → flow type list → select Bot FlowAdd
Create a Digital Bot Flow Architect → flow type list → select Digital Bot FlowAdd
View Optimization Dashboard Architect → [selected Bot or Digital Bot Flow] → Insights & Optimizations → Optimization Dashboard
View Intent Health Architect → [selected flow] → NLU menu → Intent Health

Key Concepts in Detail

Intents

An intent represents a specific customer goal. Each intent is trained with a set of utterances that the NLU model learns to recognize.

Attribute Detail
Definition A categorized customer goal the bot recognizes
Training input Utterances (sample phrases)
Best practice Provide a diverse set of utterances per intent
Intent Health Tool to identify weak or conflicting intents

Slots

Slots are the specific data points the bot collects during a conversation.

Slot Type Description
Built-in Pre-built types: date, time, number, currency, etc.
Custom List Fixed list of values (e.g., product names)
Custom Dynamic List List fetched at runtime via a data action
Custom Regex Pattern-matched input (e.g., account number format)
AI-Powered Uses Genesys AI to extract free-form values — recommended over free-form text slots
Timeslot For appointment scheduling (e.g., available time picker)

AI-Powered slots are recommended by Genesys. Free-form slot capture should be used carefully — see Virtual Agent slot authoring recommendations.

Utterances

Sample phrases used to train the bot's NLU model. More diverse, realistic utterances improve recognition accuracy.

Confirmations

Optional step that reads back captured slot values and asks the customer to confirm before the bot proceeds.

Learning

Reviews utterances the bot did not recognize and suggests adding them to improve the model over time.


Optimization Dashboard Metrics

Metric Description
Total Bot Interactions Total number of customers who interacted with the bot
Average Duration Average length of time customers spent in the bot
Average Turns Average number of steps a customer went through
Contained Interactions fully resolved within the bot (no agent handoff)
Transferred Interactions handed off to ACD
Agent Escalation Customer explicitly requested a human agent
Abandoned Customer disconnected before completing or transferring
Recognition Failure Bot could not match customer input to a known intent
Error System/expression errors during the interaction

Data retention: Utterance history and the Bot Conversation Library are available for the last 10 days.


Rich Media (Digital Bot Flows)

Type Description
Quick Replies Pre-defined response buttons the customer taps to reply — structured, fast responses
Cards Bot message with image, title, body text, and action buttons
Carousels Multiple cards displayed in a scrollable horizontal layout
List Pickers Structured lists for guided selection (e.g., appointment time slots, Apple Messages for Business)

Architect Actions (Used in Parent Flows)

Action Used In Purpose
Call Bot Flow Inbound Call / Chat / Message flows Invokes a Dialog Engine Bot Flow
Call Digital Bot Flow Message flows Invokes a Digital Bot Flow
Call Dialog Engine Bot Voice / Chat / Message flows Legacy action for Dialog Engine bots (in-flow reference)

How Bot Flows Integrate with Inbound Flows

Inbound Call / Message Flow
        ↓
Call Bot Flow action (or Call Digital Bot Flow)
        ↓
Bot Flow Runs (NLU processes customer input)
        ↓
Bot resolves intent → Collects slots → Confirms
        ↓
Exit Reason: Contained / Transfer to ACD / Agent Escalation
        ↓
Parent flow continues based on exit reason

Third-Party Bot Options (For Reference)

If a native Genesys bot is not used, the following third-party options are available:

Integration Channel
Amazon Lex V2 Voice (Inbound Call flows)
Google Dialogflow CX Voice / Message flows
Google Dialogflow ES Voice / Message flows
Nuance Mix Bot Voice flows
Genesys Bot Connector Message flows (up to 5 third-party bots)
Genesys Digital Bot Connector Message flows (up to 5 third-party bots)

Third-party bots are configured under Admin → Integrations.


PCI DSS Compliance Note (Exam Critical)

Flow Type PCI Compliant Can Use in Secure Call Flow
Dialog Engine Bot Flow Yes Yes
Digital Bot Flow No No — must not be used in Architect secure call flows

Pricing Overview

Channel Dialog Engine Bot Flow Digital Bot Flow
Voice Charged per minute, billed in 15-second increments N/A
Digital (chat/messaging) Charged per session Charged per session

Contact your Customer Success Manager or Genesys Sales for volume discounts.


Best Practices

Practice Reason
Use AI-Powered slot types where possible More flexible than free-form; Genesys-recommended
Provide varied, realistic utterances per intent Improves NLU accuracy and reduces recognition failures
Use Intent Health regularly Identifies weak intents before they impact customers
Handle all exit reasons in the parent flow Ensures graceful routing regardless of bot outcome
Use Dialog Engine Bot Flow for voice/PCI contexts Only compliant option for secure call flows
Use Intent Miner on existing transcripts Discovers real customer intents faster than manual authoring
Monitor the Optimization Dashboard Track contained vs. transferred rates to measure bot effectiveness

Key Takeaways

Topic Summary
Two native bot types Dialog Engine Bot Flow (voice + digital, PCI-compliant) · Digital Bot Flow (digital only, not PCI-compliant)
Built in Architect Both types are authored directly in Architect — no separate tool needed
Core NLU concepts Intents → trained with utterances; Slots → extract data; Confirmations → verify data
Parent flow connection Use Call Bot Flow or Call Digital Bot Flow action in parent Inbound flows
Optimization Dashboard Tracks interactions, duration, turns, and exit states including contained vs. transferred
PCI distinction Dialog Engine = PCI DSS compliant; Digital Bot = not compliant
Pricing Voice: per minute (15s increments); Digital: per session
AI enhancement Virtual Agent, Intent Miner, Knowledge Base, and AI-Powered Slots available in both