CX Today Why CX Leaders Are Tired of AI That Doesn’t Work
6 snips
Dec 15, 2025 Simon Adnett, VP of Sales for the UK & Ireland at Enghouse Interactive, dives into the real challenges of AI in customer experience. He discusses the dangers of automating broken processes and emphasizes the necessity of understanding before implementing AI. Simon reveals how automated evaluation empowers quality assurance teams and addresses failure demand, which can be tackled before it impacts customers. He also shares insights on leveraging real-time agent assist tools to enhance onboarding while keeping agents engaged, showcasing AI's potential to drive meaningful results.
AI Snips
Chapters
Transcript
Episode notes
Measure Before You Automate
- Many AI failures stem from automating broken processes instead of understanding them first.
- Simon Adnett stresses 'measure twice, cut once' before deploying AI to avoid multiplying bad processes.
Start With Automated Evaluation
- Automate evaluation first to understand every interaction before designing front-end automation.
- Use that analysis to prioritise agent assist or full automation where ROI and agent pain points align.
Sentiment Reveals Root Causes
- Spotting sentiment shifts across all contacts reveals hotspots and systemic issues in the business.
- Simon explains this enables fixing root causes rather than adding front-end automation to mask failure demand.