CX Today Is Your Approach to Dirty Data Killing Your AI Implementation?
6 snips
Dec 1, 2025 Brion Johnson, Director of Presales at TechSee, dives deep into the pitfalls of dirty data in AI implementations. He uses a motorcycle trip analogy to illustrate the risks of poor data quality. The discussion reveals how high-quality, timely, and accurate data can significantly enhance AI-driven customer experiences. Brion shares action steps to prioritize data cleanliness, tackle automation challenges, and harness multi-modal inputs for better resolutions. Moreover, he highlights the power of visual support in reducing resolution times and minimizing unnecessary field visits.
AI Snips
Chapters
Transcript
Episode notes
Motorcycle Trip As Data Metaphor
- Brion compares planning a motorcycle trip to organizing enterprise data, listing weather, roads, health, time and money as critical inputs.
- He uses this to show data sources must be timely, accurate and trusted to avoid dangerous outcomes.
Trusted Sources Beat Old Maps
- Brion says identify a trusted source for each data type and prefer recent, authoritative inputs over stale ones.
- He highlights timeliness, accuracy and trust as the three pillars of usable data.
Fix The Biggest Problems First
- Tackle the biggest 70–80% problems first by automating high-volume customer pain points like warranty claims.
- Start with the end in mind, map required verified inputs, then automate with human checks where needed.