The agentic AI conversation in retail has been dominated by the consumer-facing stuff: ChatGPT's instant checkout, Amazon's Buy for Me, Shopify's AI storefronts. But the most consequential deployment of AI agents may be happening behind the scenes, in the unglamorous plumbing of supply chain logistics. On April 9, project44 — the supply chain visibility platform used by thousands of shippers and carriers globally — acquired LunaPath.ai in an all-cash transaction. LunaPath brings more than 50 purpose-built AI agents that automate the repetitive operational work that currently consumes freight teams: carrier check calls, proof-of-delivery retrieval, claims initiation, appointment confirmations, and the endless email and phone workflows that keep goods moving.
It's project44's second strategic AI acquisition, following the ClearMetal deal in 2021, and it signals a fundamental shift in what supply chain technology is supposed to do — from showing you what's happening to actually doing something about it.
From Visibility to Execution
The promise of supply chain visibility platforms has always had a gap: they're excellent at telling you a shipment is late, but the actual work of fixing the problem — calling carriers, chasing paperwork, rebooking loads — still falls to humans. LunaPath's agents are designed to close that gap.
"LunaPath brings execution into that graph, turning intelligence into real-time action across the supply chain," said Jonathan Scherr, project44's Chief Strategy Officer, in the announcement.
LunaPath founder Abhishek Porwal framed the acquisition as solving AI's biggest logistics limitation: "project44's supply chain data graph gives our agents the context they were missing." Without real-time operational data, AI agents are fast but uninformed. Combined with project44's data layer — which the company describes as the world's largest logistics data graph — the agents can understand downstream impact before they act.
Why This Matters for Retailers
The timing is not accidental. Retail supply chains are under more stress than at any point since the pandemic: the Strait of Hormuz remains effectively closed, Cape of Good Hope rerouting has added weeks to shipping lanes, and the Iran war's fuel cost impact has pushed Amazon to impose surcharges and USPS to hike parcel rates. In this environment, the manual work of managing freight exceptions doesn't just consume time — it directly impacts whether goods arrive at stores and distribution centers on schedule.
Project44 reportedly tested eight different AI agent providers in live operations over a sixteen-month period before selecting LunaPath. CEO Jett McCandless cited "better cash flow, greater efficiency, and lower costs" as the observed results — the kind of operational language that CFOs actually respond to.
The deal also reflects a broader strategic bet: rather than relying on a single monolithic AI provider, project44 is pursuing a multi-vendor AI orchestration approach — unified governance and analytics across specialized agents. It's a model that mirrors how large retailers themselves are approaching AI: deploy best-in-class tools for specific tasks rather than betting everything on one platform.
The Agentic Supply Chain Is Here
We've covered agentic commerce extensively at Endcap Brief — from Shopify's AI storefronts to Amazon's Buy for Me to ChatGPT's checkout pivot. But project44's acquisition of LunaPath suggests the most immediate ROI for agentic AI in retail won't come from replacing shopping assistants or building autonomous checkout. It will come from automating the thousands of small, repetitive, high-friction tasks that keep supply chains functional.
Axios reported that project44 is in "deal mode" with an IPO potentially on the horizon, suggesting the company sees this AI-plus-logistics positioning as its path to public markets. If that happens, it would be one of the first pure-play supply chain AI companies to go public — and a signal that Wall Street is finally pricing the unsexy side of agentic AI.
For retailers watching the agentic hype cycle with justified skepticism, here's the reframe: the AI agents that matter most might not be the ones talking to your customers. They might be the ones calling your carriers at 2 a.m.