LILT Workflows: Designing Configurable Translation Processes
Role
Principal Designer
Team
1 Designer (me), 1 PM, 3-4 Engineers, 1 QA
Collaborated with
TL;DR
Designed smart workflow templates reducing enterprise setup time by 75% through 4 predefined patterns covering 80%+ of use cases—unlocking enterprise deals without overwhelming complexity.
Challenge & solution
LILT's simple "Translate → Review → Done" worked for most customers, but enterprises needed AI quality checks, file preprocessing, and additional review stages. Add complexity without breaking simplicity.
Competitor workflow builders were overly complex. Predefined templates could serve 80%+ of enterprise needs while maintaining ease of use.
- Smart template system with Translate, Review, AI Review, and file processing steps
- 75% reduction in setup time, improved enterprise satisfaction, enabled new market segments
- Process: cross-role research, competitive analysis, template architecture, customer validation
Design process & key decisions
Research & pattern discovery
25+ interviews revealed predictable workflow patterns—most complexity came from specific steps (AI review, file processing), not endless customization. Competitive analysis showed overwhelming workflow builders that confused 80%+ of users, so I focused on the real 'why': serving 80%+ of enterprise needs without overwhelming complexity.
Template architecture
Created 4-5 predefined workflows covering 80%+ of use cases: Simple Translation, Translation + Review, Translation + AI Review + Human Review. Added file processing for complex formats.
AI integration
Designed AI Review steps with seamless handoffs between automated and human stages. Enabled hybrid workflows combining AI preprocessing, human translation, AI review, and final approval.
Enterprise validation
Template workshops with key customers validated the approach. Ensured compliance requirements and audit trails for regulated industries.
Cross-role alignment
Gathered input from linguists, PMs, and enterprise customers to ensure workflows met real-world needs—the collective wisdom of diverse stakeholders shaped the template approach.
Results & user feedback
- Template-first success: eliminated decision paralysis, served 80%+ of users without customization
- AI positioning: adding AI review positioned LILT ahead in quality assurance
- Enterprise unlock: complex file format support through templates enabled major deals
Some edge cases still needed more customization than templates provided, requiring careful balance between simplicity and flexibility.
Broader impact
The template-first approach to workflows created a design philosophy applied across LILT's enterprise features:
80/20 rule in practice
Competitive analysis showed that 80%+ of users needed the same 4-5 patterns—serving the majority well beats serving everyone poorly.
Smart defaults over complexity
Well-designed templates eliminate decision paralysis while still accommodating edge cases through progressive disclosure.
AI as workflow step
Integrating AI review as a workflow step demonstrated that enterprise AI features need thoughtful UX to build trust—not just capabilities.
Complexity can be simple
Enterprise features succeed when they feel powerful but not overwhelming—template-based system patterns now apply across LILT's tooling.
Key learnings
Templates beat builders
Predefined templates covering 80%+ of use cases provide better UX than endless customization options.
Competitive analysis shows what not to build
Competitor workflow builders proved that complexity doesn't equal value for user experience.
AI needs thoughtful UX
Adding AI as a workflow step requires careful design to maintain user trust and efficiency.
Customer validation prevents over-engineering
Direct workshops revealed simple templates solved enterprise needs better than complex systems.
Enterprise features succeed when they feel powerful but not overwhelming—smart defaults often serve users better than endless options.
Building enterprise-grade process management?
Get in touch