69% of firms deploy AI at scale, only 19% redesign work

Sixty-nine percent of supply chain-intensive enterprises have pushed AI beyond isolated pilots, yet only 19% have significantly or extensively redesigned their workflows to let people and machines actually work together. That gap — adoption outrunning operational redesign — is the central finding of an Everest Group research report released by Lean Solutions Group, titled "Designing the Human + AI Enterprise: A Transformation Agenda for Supply Chain-Intensive Industries."

The numbers get thinner from there. Just 15% of organizations report comparable redesign of workforce roles. The survey base: more than 200 U.S.-focused senior decisionmakers across transportation, logistics and other supply chain-intensive industries.

"The challenge is no longer simply getting AI into the business. It is changing how work gets done," said Jack Freker, CEO of Lean Solutions Group. "Technology can analyze, recommend, and automate at scale, but people remain accountable for the outcome. The organizations that get the most from AI will combine experts in the loop, intelligent technology, and redesigned processes in a way that improves performance for customers."

Freker drew a hard line on autonomy, too. "The goal is not maximum AI autonomy," he said. "It is a faster, more reliable path from insight to accountable action. AI should handle work within clearly defined boundaries, while experts provide the judgment and ownership that complex supply chain operations require."

Where do shippers see the payoff?

Asked to name the leading opportunities for Human + AI transformation over the next two to three years, respondents pointed first at transportation planning and execution, cited by 53%. Workforce scheduling and labor management followed at 49%, with supply chain planning and coordination close behind at 48%. In other words, the money is expected in the operational core — moves, labor, plans — not in peripheral functions.

How big is the readiness gap?

The study flags several mismatches between what firms are deploying and what their operations can absorb:

• Human judgment still runs the shop. Decision-making is primarily human-led at 72% of organizations; exception handling remains human-led at 60%.
• Investment skews toward technology, not people. Some 67% of firms prioritize AI platforms and 66% prioritize data and integration, but only 33% prioritize workforce upskilling. Governance and change management get attention from just 17% each.
• Accountability structures lag badly. Only 26% of organizations have established formal structures for AI accountability.
• Talent pressure bites hardest in logistics. Fully 58% of logistics and transportation organizations report talent shortages, against 35% in other surveyed industries.
• Returns are inconsistent. While 84% report or expect meaningful AI value within two years, 30% say returns have come in below expectations. Only 24% report returns above expectations.

What does this mean commercially?

For carriers, 3PLs and shippers, the survey reads as a caution against equating deployment with value. Firms pouring capital into platforms and data plumbing while leaving roles, governance and training untouched are the most likely to land in that 30% cohort watching returns fall short. The 58% talent-shortage figure in transport and logistics suggests the constraint on AI value in this sector is people, not compute — and that competitors who solve the upskilling problem first will convert the same tools into better service and cost performance.

The accountability deficit matters most where freight decisions carry real money. If only a quarter of organizations can formally assign responsibility for AI-driven outcomes, exception handling will stay human-led by default — which limits how much of the 53%-cited transportation planning opportunity can actually be captured.

Expect the squeeze to show up in budgets. With 84% of firms banking on meaningful value inside two years and nearly a third already disappointed, the next spending cycle is likely to shift from AI platforms toward process redesign, governance and the experts-in-the-loop models Freker describes.