Client
HAUFE
Deliverables
AI UX Conversation Design Enterprise SaaS Research Rapid Prototyping
Year
2025
Role
Senior Product Designer
Designed the end-to-end user experience for enterprise AI agents supporting customer service and internal operations. The project combined conversational AI, agent management tools, and human-centered workflows to create transparent, trustworthy, and scalable AI experiences for enterprise environments.
Outcome
The project established a scalable UX foundation for enterprise AI agents by combining conversational experiences with intuitive management tools.
The final solution enabled organizations to deploy AI confidently while maintaining transparency, operational control and seamless collaboration between autonomous systems and human teams. It introduced consistent conversation patterns, structured agent behavior and intuitive administration tools that supported continuous improvement over time.
2
Enterprise AI experiences designed for customer support and internal automation.
1
Conversation design framework governing AI behavior across products
1
Admin platform enabling non-technical teams to monitor and manage AI agents
35% Tickets Reduction
Reduction in support tickets escalated to human agents
50% Lower Onboarding Time
Reduction in onboarding time for operations teams
85% Higher Adoption
Internal adoption among support leadership after rollout
Final Impact
The result was more than a redesigned interface. It was a strategic transformation of fleet operations that reduced product fragmentation, improved operational visibility, simplified complex enterprise workflows and created a scalable foundation for future innovation.
Challenges
Designing AI products required rethinking traditional UX principles.
Instead of guiding users through predefined journeys, the experience had to support dynamic conversations, uncertainty, and autonomous decision-making while remaining transparent and predictable.
Key challenges included:
Designing conversations instead of linear user flows
Building trust in autonomous AI systems
Creating intuitive administration tools for non-technical users
Designing graceful fallback experiences and seamless human handover
Balancing automation with user control and operational visibility
Objectives
The project focused on creating a scalable AI experience that could be adopted confidently across enterprise environments.
Key objectives included:
Define a structured conversation framework
Design transparent AI interactions
Enable non-technical teams to manage AI agents
Improve trust through explainable interactions
Create reliable human escalation paths
Support continuous improvement through operational insights
Process
RESEARCH
Understanding how people interact with AI required combining traditional UX research with conversation design.
The research phase included stakeholder workshops, support team shadowing, workflow analysis, customer support ticket reviews and competitive benchmarking. Together these activities revealed how users communicated, where AI assistance created value and which situations required human intervention.
These insights formed the foundation for the conversation taxonomy, interaction principles and overall AI behavior.
STRATEGY
Before designing interfaces, we first defined how the AI should think, communicate and collaborate with users.
Research findings were translated into a structured conversation framework that aligned product, design and engineering around a shared behavioral model. Conversation taxonomy, escalation rules, confidence indicators and fallback patterns became reusable building blocks for every AI interaction across the platform.
IDEATION
The design process combined rapid prototyping with AI-assisted exploration.
Conversation frameworks were translated into interactive prototypes that allowed both customer-facing experiences and internal management interfaces to evolve in parallel. AI-assisted ideation accelerated early exploration, while iterative refinement in Figma ensured consistency, accessibility and alignment with enterprise design standards.
VALIDATION
Prototypes were evaluated with both end users and internal operational teams through moderated usability testing.
Testing focused on conversation clarity, user confidence, navigation, task completion and administrator workflows. Feedback informed multiple design iterations, particularly around transparency, confidence indicators, escalation behavior, and management interfaces that improved trust without increasing complexity.
Final thoughts
Designing AI products extends far beyond creating interfaces.
This project demonstrated that successful AI experiences depend on clear communication, transparent system behavior and giving people confidence that they remain in control. By combining conversation design, enterprise UX and operational tooling, the project created a scalable foundation for trustworthy AI adoption within complex enterprise environments.