Improving Warehouse Efficiency with Unified Data and AI-Driven Visibility - with Dan Keto of Easy Metrics (19 min)
- Release date: 2026-02-19
- Listen on Spotify: Open episode
- Episode description:
Today's guest is Dan Keto, President and Co-founder at Easy Metrics, where he focuses on helping warehouse and distribution teams turn fragmented transactional data into a unified "single pane of glass" that supports faster diagnosis of variance and more defensible decision-making. Dan joins Emerj's Matthew DeMello to explore what a solid data foundation looks like in warehouse networks — and why it matters before teams attempt to layer AI on top. He also shares practical takeaways on how enterprises can align stakeholders around a common data language, avoid costly "AI-first" missteps, and use repeatable investigations and alerts to surface real cost drivers. This episode is sponsored by Easy Metrics. If you've enjoyed or benefited from some of the insights of this episode, consider leaving us a five-star review on Apple Podcasts, and let us know what you learned, found helpful, or liked most about this show!
Summary
- 📊 Data Fragmentation Crisis: Warehouses generate unprecedented data from automation but lack visibility, turning simple variance questions into days-long delays and reactionary crisis management.
- 🔗 Unified Data Prerequisite: Aligning disparate transactional data into stakeholder-tailored models is essential before AI to avoid hallucinations, 1000x cost spikes, and rebuilds.
- 🤖 AI for Mathematical Workflows: Preprocessing warehouse math data with algorithms enables AI ‘investigations’ on overtime, fatigue, and utilization, harvesting insights proactively.
- 👥 Human-in-the-Loop Essential: AI acts as recommender flagging issues; supervisors add context, preventing errors in dynamic ops until full visual capture evolves.
- ⚙️ Holistic vs. Siloed Optimization: Treat distribution as manufacturing flow: integrate silos for true ROI, balancing service, costs, and spillover effects from tech like auto-stores.
Insights
Why do modern warehouses drown in data yet suffer from poor visibility?
Time: 3:36 – 4:51
Category: AI in Workforce DisruptionAnswer: Despite massive datafication from robotics, automation, and WMS systems, fragmented sources create silos that obscure true performance, delaying decisions on costs and efficiency amid rising customer velocity demands like Amazon’s 30-minute deliveries. (Start at 3:36)
What makes unified data the essential foundation before applying AI in warehouses?
Time: 5:23 – 6:38
Category: AI-Driven Innovation EconomyAnswer: A unified data model aligns transactional data from diverse systems into a single pane of glass, tailored to stakeholder views (ops, engineering, execs), preventing AI hallucinations, skyrocketing costs, and the need to rebuild models later. (Start at 5:23)
Why are LLMs ineffective for core warehouse operations without preprocessing?
Time: 6:54 – 7:27
Category: AI-Driven Innovation EconomyAnswer: Warehouse workflows generate mostly mathematical transactional data (e.g., robot activities), unsuitable for direct LLM use which leads to poor results; algorithms and conditioning models must preprocess data first, unlike admin tasks where LLMs shine on language. (Start at 6:54)
How does AI transform data grinding into proactive ‘harvesting’ of insights?
Time: 8:10 – 9:58
Category: AI-Driven Innovation EconomyAnswer: Post-unification, AI powers ‘investigations’ like overtime analysis factoring fatigue and safety, delivering alerts on issues (e.g., $100k utilization loss) with fixes, supervisors, and network views—replacing manual BI mining with automated prioritization. (Start at 8:10)
Why must AI remain a recommender, not decision-maker, in warehouse ops?
Time: 12:03 – 13:26
Category: AI in Workforce DisruptionAnswer: AI excels at surfacing issues but lacks floor-level context, human supervisors provide oversight like developers reviewing Copilot code; premature full automation risks errors from data gaps until visual AI captures everything. (Start at 12:03)
How can leaders avoid siloed AI optimizations that backfire in distribution?
Time: 14:03 – 15:44
Category: AI-Driven Innovation EconomyAnswer: Warehouses evolve like manufacturing, needing holistic flow optimization across processes (e.g., auto-store picking boosts but spills costs elsewhere); align industrial eng, finance, ops lenses to balance service levels and total costs. (Start at 14:03)