Leading Travel Organization
AI-guided next best action for customer support
Summary
Proactive guidance for every interaction, reducing training time, improving first-contact resolution, and delivering €2M+ estimated ROI.
Support agents at this travel organization, especially new employees, faced inconsistent outcomes due to unclear next steps when navigating complex procedures, leading to repeat contacts and unnecessary escalations. DEUS designed an AI experience that surfaces timely, context-aware next steps during live calls, detecting meaningful signals to retrieve relevant rules and work instructions. The system provides step-by-step adaptive guidance while keeping the agent in full control.
Services
Industry
The Challenge
New agents navigating complex procedures without guidance, causing inconsistent outcomes, repeat contacts, and unnecessary escalations.
The contact center handles a high volume of diverse customer inquiries, many involving complex, multi-step procedures. For newer or temporary agents, the absence of clear guidance on the right next action led to inconsistent outcomes, longer ramp-up periods, avoidable escalations, and repeat contacts from customers who didn't receive the right resolution the first time.
Problem 01
Long ramp-up for new and temporary agents
- ▪High training overhead for agents handling complex procedures
- ▪Knowledge concentrated in experienced staff
- ▪New agents lacking confidence to resolve without escalation
- ▪Seasonal hiring amplified the inconsistency challenge
Problem 02
Inconsistent decisions driving repeat contacts
- ▪No standardized guidance on next best action per situation
- ▪Agents making subjective calls on complex procedure flows
- ▪Unnecessary escalations adding to handling time and cost
- ▪Customers repeating contact to correct first-call outcomes
The Solution
DEUS designed an AI-guided next best action experience that works in real time during live calls, detecting meaningful signals in the conversation and surfacing the right procedure, rule, or work instruction at exactly the right moment.
01
Signal detection and intent recognition
The system listens for meaningful signals within the live conversation, recognizing customer journey stage and intent to determine which rules, decision criteria, and work instructions are relevant.
02
Rules engine and adaptive guidance
Based on detected signals, a rules engine retrieves the appropriate work instructions and presents step-by-step adaptive guidance directly to the agent, keeping them in control while reducing cognitive load.
03
Confidence-building for every agent tier
The system is designed to work for both new and experienced agents, accelerating ramp-up for new hires while giving experienced agents a reliable decision support layer for edge cases and complex procedures.
Impact
Estimated €2M+ ROI by reducing training overhead and call time while enabling confident, consistent decisions at every agent level.
Training Time
Estimated value of reduced training time for new and temporary staff, accelerating time-to-competence across the team.
Call Time
Estimated value of average call time reduction at industry benchmark, driven by faster, more confident decisions in real time.
Resolution Quality
Conservative to optimistic improvement scenario, tracked across 5 first-order KPIs including resolution rate, decision efficiency, and relevance.
Methods
A human-centered, end-to-end design and delivery approach, grounding every decision in agent reality and organizational process.
User interview
Shadowing
Focus group
Personas
Journey map
Service blueprint
Stakeholders map
Ideation workshop
Concept book
Design sprint
UI / UX design
ROI modeling
Solution architecture
Engineering
Project management