Seamless Handoff Protocol
Innovative Conversation Orchestration
Chatbot – CEC collaboration, rather than a point-to-point integration.
Innovation and differentiation opportunities for Chatbot vendors:
Customer is engaged during “wait for pairing” phase.
Refined handling for high friction situations as long waiting time.
Advanced flows like Human in the Loop.
Real-time adjustment of handoff parameters:
Sentiment and urgency updates
Intent refinement
Focus on core business, rather than playing the “we need more integrations” catch-up game.
Unified interface to any CEC:
Escalation
Business hours
Availability verification
Sticky assignment
Protocol diagram
Conversation control
The chatbot is in control of the conversation until an agent is paired.
Imagine using a switchboard to establish which party is connected with the customer.
In a typical handover, the chatbot ceases control of the conversation once an escalation is started, and control is taken over by the CEC. Chatbot resumes control when the CEC signals that the escalation is finished. The customer input is blocked at this moment and the feedback received depends on the CEC capabilities to send relevant updates about queue position or wait time.
With Seamless Handoff Protocol, the chatbot retains control of the conversation until SentioCX signals that an agent is paired. This enables the chatbot to continue engaging the customer using smart and innovative means. This improves the customer experience and provides an opportunity to determine sentiment, urgency or intent changes. SentioCX is continuously updated with these changes in real-time, and uses the new data for re-adjusting the handoff rank, leading to better timing and more precise pairing.
In a nutshell:
The conversation is in the chatbot realm until paired.
Chatbot continues to engage the customer and to determine changes for sentiment, urgency or intent.
The changes are pushed to SentioCX to update pairing requirements in realtime.
During this time, SentioCX provides regular wait time predictions.
Convenience orchestration features
Business Hours
The protocol has built-in support for business hours validation. It is not necessary to perform this validation in the chatbot dialog design. This is very convenient since it works out of the box for any CEC.
Availability
The protocol has built-in support for situations when no agents are online. It is not necessary to perform this validation in the chatbot dialog design (Check Queue, Check Agent etc.). This is very convenient since it works out of the box for any CEC.
Sticky pairing
The protocol has built-in support for situations when customers return to continue the conversation on a specific topic (case, ticket, order, etc.). SentioCX provides a unified approach to automatically escalate the customer to the optimal agent, either the original agent (sticky) or an equally adequate agent based on the original escalation intent.
Flexible orchestration
Fallback configuration
The protocol allows to configure the desired behavior for situations when agent pairing is not possible. The situation may occur when:
No adequate agents are online.
SLAs are exceeded.
Outside business hours.
The default behavior is to create a "paper trail" in the CEC. This may be an interaction/ticket/case/conversation, based on the CEC implementation.
The protocol provides an instrument to override the default behavior and allows the chatbot to handle the fallback procedure based on the conversation dialog design.
Cancel a handoff
Given that the conversation control is still with the chatbot, it is possible for this one to determine that the handoff is no longer necessary (enquiry is resolved, customer changed his mind, etc.).
The protocol provides an easy way to cancel the handoff and optionally to create a paper trail ticket/case/interaction in the CEC.
Easy re-escalation
Control of the conversation is resumed by the chatbot after the agent leaves the conversation.
It may happen that the customer needs/requires to reconnect with the agent in the same chatbot session. The protocol enables this scenario, maintaining the context and preventing duplication of transcripts.