APR 20, 2026|7 min read
APR 30, 2026|7 min read
From AI Pilots to AI Operations: How Leaders Across Retail and Manufacturing Industries Make Agentic AI Real
The experimentation phase is over. At Accelerate 2026, the signal from industry leaders was loud and clear: the future of AI is no longer a theoretical exercise — it’s an operational necessity, and leading companies are already making it a core part of their baseline functions. In the retail sector alone, 58% of companies are actively deploying AI. Consumers are increasingly shopping through agents that have never visited a brand’s website. And in manufacturing, some companies are deploying AI-powered agents as live, queryable knowledgebases to address the loss of institutional knowledge as millions of skilled trade jobs are projected to be unfilled by 2030.
For both retail and manufacturing leaders, the choice is clear: AI is a critical operational priority. The organizations that showed up at Accelerate 2026 weren’t just talking about pilot projects, they were demonstrating production-grade, agentic workflows that are changing their business models in real time.
The big shift: from search to agents
The Snowflake 2026 Data Trends reports are unambiguous: we have entered the Era of Agentic AI. For retail, the traditional model of digital commerce is being disrupted as consumers increasingly embrace conversation over search. As Shanthi Rajagopalan, Global Head of Strategy for Retail & Consumer Goods at Microsoft, noted during our executive panel with Microsoft at NRF 2026, “In the traditional retail world, we always thought about stores and real estate as location, location, location. And as we moved into digital and ecommerce, it became all about search, search, search.” However, as Rajagopalan added, “What we’re really seeing now is a disruption again, with consumers turning more and more to conversational interfaces for product discovery and purchases.”
This new reality, this fundamental change to the processes and norms of retail, requires a change that’s just as significant in how retailers handle data. “We’re now in a position where shoppers are telling us so much more information about who they are and what they want,” Rajagopalan said. “As an industry, we really need to figure out how to translate those natural language sentiments into usable metadata.”
In manufacturing, the shift has been equally as profound, a move away from static spreadsheets and siloed data toward AI in embedded systems. AI isn’t just a tool here, but rather the only viable succession plan available for a workforce experiencing a silver tsunami of departures. The research predicts what Accelerate 2026 confirmed: the leaders in those rooms weren’t running pilots — they were carefully considering how to become smarter, faster and more efficient organizations with focused AI use cases.
Retail: The rise of agentic commerce and the race to remain discoverable
Agentic commerce is the defining shift of the year. In a recent interview Michele Fisher, Global Director of Business Strategy for Retail and Consumer Goods at Microsoft, shared how retailers must consider how consumers reach them to remain discoverable amid the shift from an omnichannel approach to agentic commerce. Brands that haven’t designed their semantic space are finding themselves invisible to agents shopping on behalf of buyers.
Shalion, a market research leader, demonstrated the power of preparing for this important shift in consumer behavior via their conversational AI layer they’ve named Maestro. As Alejo Buxeres Solder, CDO at Shalion, explained, the real unlock isn’t the LLM itself — which he feels have become commodified — but the governed Snowflake semantic layer underneath. This architecture allows brand managers to ask complex questions, such as, “Why did my market share drop in Spain last week?” and receive immediate, data-backed answers.
Retailers are now focused on Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO), replacing traditional keyword SEO to ensure greater visibility to shopping agents. Those retailers are attempting to see themselves through the eyes of the agent, so to speak, and optimizing for key questions, rather than key words, that surface their products more effectively to these agents. And as Kelly Thompson, former Chief Merchant at Walmart noted, every AI use case must be quantifiable — generating revenue, reducing cost, or improving productivity — before committing any resources to them. As in manufacturing, the shift from running pilots to business-focused, transformational applications of AI is a key focus for retailers.
Manufacturing: Transforming processes through agentic operations
Traditional quality analysis examines defects in isolation, considering siloed variables independently, missing the network effects where root causes involve multiple interacting factors. By the time a manufacturer traces a defect to a specific supplier or process step, days have passed.
Snowflake's Manufacturing Field CTO, Tripp Smith, demonstrated a Graph Neural Network (GNN) Process Traceability solution, an approach that models the entire manufacturing process — suppliers, materials, stations and defects — as an interconnected network. Now, root-cause analysis that once took days takes minutes, with the ability to predict defect risk before production begins.
In manufacturing, the stakes are measured in uptime and precision. Smith showcased the power of GraphSage, a neural network running on NVIDIA GPUs that treats the entire supply chain — suppliers, stations, materials and defects — as an interconnected graph.
By unifying IT and OT data into a single namespace, companies like WolfSpeed and Lindt Chocolate have moved beyond independent variable analysis. WolfSpeed successfully deployed 12 Snowflake Intelligence agents, transforming week-long analyst tasks into instantaneous root-cause investigations. This ability to trace a defect back to a specific supplier or process adjustment in seconds is emblematic of the power of autonomous manufacturing.
The same shift is playing out in industrial services. United Rentals — managing equipment fleets across 1,600+ branches — deployed a Business Intelligence Agent on Snowflake Intelligence that lets branch managers and regional teams ask natural-language questions of their operational and financial data in real time, eliminating the manual analysis bottleneck that once slowed decisions in the field.
Agentic advantage proven in every industry
Retail and Manufacturing aren’t the only industries experiencing fundamental transformations through AI. The experience has become universal:
Financial Services: Asset manager agent swarms are reducing portfolio management and risk analytics workflows from weeks to moments.
Healthcare and Life Sciences: Innovalon is reducing 200-page medical chart reviews from weeks to minutes using agentic prior authorization.
Advertising, Media and Entertainment: Warner Music Group is using Cortex Code to empower analyst teams to build dashboards and semantic views without engineering support.
Public Sector: Agencies like the Township of King are surfacing operational wins, such as streetlight maintenance compliance, which were previously buried in hundreds of KPIs.
What industry leaders do differently
They Identify Outcomes First: They identify use cases against the three buckets that Kelly Thompson discussed — generating revenue, reducing cost, and improving productivity — before selecting any technology.
Data as a Prerequisite: They treat data readiness (unified namespaces and governed semantic layers) as a prerequisite for agent deployment, not a follow-on project.
Specific Workflow Focus: They start with high-value, specific workflows (a product catalog, a shift comment archive) rather than general AI strategies.
Operational Measurement: They measure agentic AI against a quantifiable business outcome, not a technology KPI.
As Luv Kothari, General Manager for Forward Deployed AI Engineering at Snowflake, noted, successful organizations achieve earnings-level impact in three to six months by moving from tactical queries to strategic, agent-led operations.
The Path Forward
The experimentation phase is over. The era of execution is here.
To see Agentic Commerce working in production, watch the Retail Accelerate session on-demand to see Maestro in action. Watch the Manufacturing session to see how Tripp Smith traced a defect using a live GraphSage demo. Then, download the Data Trends 2026 Manufacturing and Retail & Consumer Goods ebooks to identify the most crucial things to know about the future of AI in these industries.
The question is no longer whether your organization will adopt Agentic AI. It’s how fast you can operationalize it to be at the front of the pack.
Download the Data Trends 2026 ebook for your industry:
Author
Just For You
