CueIn AI Experience - Tifany Pan

The Context

CueIn AI's Mission

CueIn AI built conversational AI solutions for enterprise customer success teams, helping companies like Airbnb and Uber understand and categorize customer support interactions at scale.

My Role

I worked on the operational foundations that made AI models work - from data labeling and taxonomy design to automating sales processes and supporting business development.

Learning While Contributing

This was my introduction to hands-on AI work. I got to see how AI models actually get built and deployed in the real world - it's a lot more about good data and smart operations than just the algorithms.

What I Actually Did

Data Labeling & Taxonomy Design

Designed classification systems for customer support conversations
Labeled 179K conversation samples for Airbnb's intent prediction model
Created taxonomies for root cause analysis across different industries
Achieved 0.87 F1 score on model training data (apparently that's really good!)

Go-to-Market Operations

Automated lead generation using LinkedIn Sales Navigator and Python scripts
Used Google Trends to identify high-value prospects based on search volume
Improved lead pipeline efficiency by replacing manual outreach processes
Discovered insights about pricing models and product-market fit challenges

Cross-functional Support

Collaborated with 4 data scientists and 3 engineers on model development
Introduced roadmaps and milestone tracking to weekly planning
Supported business development across 5 clients in different industries
Learned to code through scripting and automation projects

Client Work

Worked with enterprise clients including Airbnb and Uber
Supported smaller brands like Crocs and Urban Outfitters
Delivered production-ready models within tight 3-week timelines
Gained experience across healthcare, finance, travel, and retail industries

What I Learned & Contributed

0.87

F1 Score achieved on Airbnb model training data

Technical Skills Developed

Hands-on experience with model training data preparation
Understanding of how enterprise AI actually gets built
Data analysis and taxonomy design skills
Basic Python scripting and automation

Business Insights Gained

Learned how AI startups approach enterprise sales
Understood the gap between AI capabilities and business value
Experienced the operational challenges of AI implementation
Discovered the importance of product-market fit in AI

Key Realization

Good AI isn't just about algorithms - it's about having the right data, the right processes, and understanding the real business problems you're trying to solve. The operational side is just as important as the technical side.

What This Experience Taught Me

Data quality is everything in AI

The best models are only as good as their training data. Careful labeling and taxonomy design directly impact model performance - it's detailed work that really matters.

Enterprise AI is about operations, not just algorithms

Most of the work happens before and after the model training - understanding client needs, preparing data, and integrating into existing workflows.

Product-market fit challenges are real

I learned firsthand how difficult it is to find the right balance between serving high-value customers who want to build their own solutions and smaller customers who may not provide enough revenue.

Automation can have immediate business impact

Even simple automation projects can significantly improve business operations. My lead generation work showed me how technical skills can directly solve business problems.