AI voicebot handling customer needs and upselling additional services

Client NamePelicantravel.com

Client CountrySlovakia and the Czech Republic

  • Client typeSME
  • IndustryHospitality & Travel
  • Application areasMarketing, Sales & Customer Engagement
  • AI technologiesConversational AI (chatbots, voicebots), Generative AI, Large Language Models (LLMs), Multimodal AI, Speech Recognition & Synthesis
  • Business impactsCustomer Experience & Market Growth, Customer Product & Service Innovation
  • Data typesAudio Data, Documents / Semi-structured Data, Structured Tabular Data, Textual Data
  • Delivery modelsService / Subscription
  • DeploymentsCloud
  • Key capabilitiesConversational & Language Interaction, Recommendation & Personalization
  • Project stagesInitial Production Deployment
  • Solution formsAnalysis, Recommendation, or Report, API / Micro-service Interface, Conversational Interface

Solution Description

Problem description

Pelican Travel is the largest Slovak travel platform. Most reservations are handled online, but for changes, cancellations, or baggage inquiries, customers prefer to call. Phone support, however, is both staff- and cost-intensive. The situation escalated during the COVID-19 pandemic, when the customer support line collapsed under the flood of calls. Pelican faced overload, low efficiency in handling requests, and a lack of oversight over priority cases. They needed a solution to ease the burden on the team, improve workflow organization, and enhance support availability.

Solution

In cooperation with Feedyou, Pelican deployed AI voice assistants that gradually took over routine requests, streamlined call handling, and enabled proactive customer outreach. Thanks to smart IVR and improved caller verification, verification success increased from the original 65% to over 92%, dramatically reducing the number of escalations. Outbound campaigns opened new sales opportunities – for example, offering flight insurance achieved a conversion rate of 16.3%. This gradual and sustainable approach improved support availability, reduced the load on the contact center, and paved the way for further innovations.

Main Users of the Solution

Project owner Pelican Travel defined the needs and provided domain expertise. Customer support operators helped design scenarios and test the solution. End clients provided feedback. Feedyou handled solution development and integration, including language localization and system connections.

Technologies used

Microsoft Azure, JavaScript, Node.js, React, Cypress, LangChain

Additional services

  • AI strategy and roadmap
  • Audit / feasibility study
  • Identification and prioritization of suitable use-cases
  • Data collection and pre-processing
  • AI model selection and customisation

Implementation

Project Owner on the Client's Side

C-level management

Participation on the Client's Side

  • Business / Product Owner
  • Domain / Process experts

Form of Supplier Involvement

Full implementation

Operation and Maintenance

Operational Model

Supplier

Impact and Results

Qualitative Benefits

Improved availability of customer support (voicebot handles calls 24/7, including weekends and holidays). Higher quality of communication – customers don’t need to navigate IVR trees, they simply state their request. Clearer organization of contact center work (request prioritization, ticketing). Better customer experience – shorter wait times, clear and consistent information. Proactive communication opportunities (outbound campaigns with insurance, accommodation, transfers). Evolutionary deployment = less strain on internal teams and clients, smoother adoption.

Quantitative Results

Caller verification success increased from 65% to 92+%
Outbound voicebot: 16.3% conversion, up to 20% for insurance
Decrease in escalations to customer support
Thousands of calls per month handled without team expansion
Automatic routing of requests in the local language based on the market

Lessons Learned and Recommendations

Key Success Factors

Step-by-step evolutionary deployment (minimal burden for team and customers)
Strong collaboration between Pelican and Feedyou teams
Clear priority on key scenarios (caller verification, request routing, outbound campaigns)
High-quality language processing and localization
Client openness to innovation and willingness to test new approaches

Biggest Challenges

Overloaded support line during crises (e.g., COVID-19) → solved with automation and prioritization
Low caller verification success (originally 20–30%) → solved by switching to numeric codes (92+%)
Customer concerns about talking to a voicebot → resolved through natural language interface and combination with live agents
Integration with existing systems (CRM, helpdesk) → solved with phased introduction and small-scale testing

Recommendation for Others

Start evolutionarily, not revolutionarily – test solutions in smaller steps.
Pay attention to user experience – the voicebot must communicate naturally and simply.
Focus on measurable scenarios with clear added value (verification, upsell, automation).
Plan for maximum value by combining AI + live agents.
Do not neglect localization and language quality, especially in international environments.

Stay informed with CNAIP. Subscribe to our regular mediamonitor and never miss an update in the world of AI. We’ll deliver a digest of the most essential news straight to your inbox.

By subscribing, you agree to our Terms of Service.

© cnaip 2026

Want to become a part of Czech AI?

Share your story and showcase what you can achieve with artificial intelligence. Your involvement will inspire others and help us map out the Czech AI scene in its entirety.