"Across the show floor, AI was in production. AI chatbot usage has increased, with customers who engage converting at higher rates. But as retailers deploy AI at scale, they're discovering new operational challenges: hallucinations in customer service responses, variable token costs that spike unexpectedly, and unpredictable latency from retrieval-augmented generation (RAG) adding 800ms or more to response times. This aligns with our research showing that 50% of retailers cite AI as their primary driver for observability adoption."
"As AI agents proliferate and interact with each other, the governance and risk management require real-time visibility into the entire AI stack. New Relic AI Monitoring provides visibility into the entire AI stack, whether it's LLMs, vector databases, or agentic workflows: Model performance tracks throughput, latency, and error rates, and model inventory auto-discovers models in production detecting "shadow AI" usage Token cost analysis automatically calculates cost per request, while token efficiency ratios reveal successful outcomes versus tokens consumed, help"
"AI dominated conversations at NRF 2026-but not in the way it has at previous shows. This year, the focus shifted from demos and pilots to operational deployment and risk management. Christian Beckner, NRF's Vice President of Retail Technology, summarized the shift: "Agentic AI continues to transform all aspects of retail, including supply chains, warehouse operations, marketing and customer engagement. Agentic AI also creates new risks that retailers need to understand and manage...companies will need to develop strong governance to address these challenges.""
Consolidating the retail stack onto a single observability platform reduces MTTR by eliminating tool sprawl and unifying data. AI at NRF 2026 moved from pilots to operational deployment, highlighting production use, increased chatbot-driven conversions, and new operational issues such as hallucinations, spiking token costs, and RAG-induced latency. Half of retailers cite AI as the primary driver for observability adoption. As agentic AI agents proliferate and interact, real-time visibility across models, vector stores, and agent workflows becomes essential. New Relic AI Monitoring tracks model throughput, latency, error rates, discovers shadow models, and analyzes token costs and efficiency.
Read at New Relic
Unable to calculate read time
Collection
[
|
...
]