AI Assistant "Elmo".
Case Study Notice:
To respect confidentiality, the designs shown in this case study have been recreated and do not reflect the original visual identity. Company-sensitive details have been removed. However, the strategic foundation, structure, and user experience approach remain true to the original project. All data and results are based on real outcomes.
About
In today’s digital world, people expect fast, helpful support. At the same time, businesses want to reduce support costs and uncover new revenue streams. I designed an AI assistant that eases the pressure on customer service while subtly boosting sales through smart, contextual product suggestions.
The goal was clear: help users solve problems instantly and discover relevant products – without disrupting the support flow.
Lead UX Designer
Business Project
2024
IT / E-Commerce
5,000+ employees
Challenge
As the platform expanded, so did the volume of support tickets – especially repetitive questions about common issues. The support team was overwhelmed, and many users struggled to find the help they needed quickly.
At the same time, key features were underused, and opportunities for cross- and upselling were being missed.
The challenge: design an AI assistant that can
handle simple issues in real time,
reduce pressure on the support team,
and introduce helpful product suggestions –without being pushy or distracting.
Results
The results showed real impact.
10% of support tickets are now resolved automatically, saving thousands of manual replies every month.
60% of users who interact with the assistant no longer need to contact support again.
Users now get quick answers, and the support team can focus on more complex issues. At the same time, contextual product suggestions gently guide users toward upgrades and add-ons – helping increase revenue while staying user-focused.
10%
Tickets resolved by AI
60%
Users don’t return to support
UX Strategy
The assistant was built with a “support first” mindset. Sales came second – woven into the experience in a way that feels natural and helpful. Key elements of the UX strategy included:
Instant help, less friction: Users get immediate answers to common problems – no forms, no waiting.
Always there, never intrusive: A floating bubble gives users access anytime, without interrupting their experience.
Conversational and clear: The assistant uses friendly, human language to guide users and build trust.
Relevant product nudges: Suggestions are tailored to user context and framed as helpful tips, not sales pushes.
Learning and improving: The assistant collects feedback and adapts, improving over time based on real user needs.
Solution
Conclusion
This assistant did more than reduce ticket volume – it reshaped the support experience. It helped users solve problems faster, eased the load on support teams, and quietly created new sales opportunities.
One of my biggest learnings? Automation doesn’t need to feel cold or transactional. When it’s user-first, with thoughtful design and real value, it can be warm, helpful – and even a little delightful.
Looking back, I’d explore more early user testing to better fine-tune the tone and intent behind messages. Closer collaboration with the support team from day one could also make the handover from bot to human even smoother.
This project deepened my belief that great UX lives at the intersection of empathy, clarity, and smart strategy.
