Customer support · Production

TourRadar saves 780+ hours monthly and resolves 75%+ of customer inquiries with boost.ai generative AI

The problem

TourRadar's support team managed a high volume of diverse inquiries for a global customer base across multiple time zones and wanted to move beyond rules-based chatbots that provided rigid, templated responses, requiring instead a scalable solution capable of dynamic, contextually relevant answers without constant manual content updates.

First attempt

A rigid, keyword-based chatbot previously in use provided templated responses that could not handle the inherent complexity of travel booking inquiries.

Workflow diagram · grounded in source

Customer submits support inquiry

trigger

“The focus was on building an AI Agent capable of providing helpful, human-like support across essential areas such as booking and tour details, account support, cancellations, and payment issues”

Generative Action reads tour content

ai_action

“Generative Action, a feature that allows the AI Agent to generate dynamic, contextually relevant responses from approved knowledge sources. One of the most valuable implementations of this feature is on Tour Detail Pages, where the AI Ag…”

Personalized tour recommendation

ai_action

“the AI Agent can instantly suggest the top three tours based on a traveler's preferences and needs”

CRM integration for agent drafting

integration

“integrates directly into TourRadar's CRM, assisting human agents with message drafting in the Booking Conversation Page, where they engage with both customers and tour operators”

Escalation to human agent

human_review

“only 20% requiring escalation to a human agent, freeing up the human team to focus on more complex and high-value customer interactions”

Reported outcome

The AI Agent handles 82% of all customer inquiries with 98% accuracy, live chat deflection improved from 50% to 80%, and the team saves 780+ hours per month—580 on live chat and 200 on written replies—with only 20% of conversations requiring escalation to a human agent.

Reported metrics

customer inquiries handled by AI (results at a glance)75%

customer inquiries handled by AI (since launch)82%

Live chat deflection rate (current)80%

Live chat deflection rate (previous)50%

Show all 12 reported metrics

customer inquiries handled by AI (results at a glance)75%

customer inquiries handled by AI (since launch)82%

live chat deflection rate (current)80%

live chat deflection rate (previous)50%

live chat deflection gain30-point gain

response accuracy rate98%

total hours saved per month780+ hours

hours saved on live chat per monthapproximately 580 hours

hours saved on written replies per monthapproximately 200 hours

CSAT score for AI-handled chats60%

conversations requiring human escalation20%

Generative Action live topics31

Reported stack

boost.aiGenerative ActionCRM

Source

https://www.boost.ai/case-studies/how-tourradar-created-a-more-seamless-customer-experience-and-saved-780-hours-monthly-with-generative-ai

Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

The AI Agent handles 82% of all customer inquiries with 98% accuracy, live chat deflection improved from 50% to 80%, and the team saves 780+ hours per month—580 on live chat and 200 on written replies—with only 20% of…

What tools did this team use?

boost.ai, Generative Action, CRM.

What results were reported?

customer inquiries handled by AI (results at a glance): 75%; customer inquiries handled by AI (since launch): 82%; Live chat deflection rate (current): 80%; Live chat deflection rate (previous): 50% (source-reported, not independently verified).

What failed first in this deployment?