Softgoods Business & MarketsTKT-2321
Agentic AI Cuts Textile Recycler Costs Without Hardware CapEx
Applied Industrials CEO Oliver Ardery told the SMART conference in Montréal that agentic AI delivers recycler ROI in weeks with zero hardware CapEx, citing a 2 percent FedEx billing error catch.
- Cut
- November 17, 2026
- Sew
- 4 min
- Ticket
- TKT-2321

Construction
SMART held its conference in Montréal, Québec, last month; Applied Industrials CEO Oliver Ardery delivered the keynote.
One AI deployment caught 2 percent of 2025 FedEx billing incorrectly charging a client.
An AI workflow flagged a 7 percent weight variance between ordered and post-shredding material.
EU Digital Product Passport and California SB 707 compliance framed as data-orchestration challenges.
Market intelligence tools screen used-clothing exporters (HS 6309) buyers down to city level in Chile, Uganda and Central America.
Agentic AI workflows pay back in weeks, not the multi-year horizons of optical sorting lines, according to Oliver Ardery, CEO of Applied Industrials, who keynoted the Secondary Materials and Recycled Textiles (SMART) association's conference in Montréal, Québec, last month.
Ardery's address, titled "SMART AI Adoption: Front- and Back-Office Automation," drew a sharp line between physical AI — robotic sorting arms and near-infrared scanners — and agentic AI: autonomous software agents that automate front- and back-office workflows. Applied Industrials builds the second kind for industrial, manufacturing and distribution companies.
"Their immediate operational bottlenecks are actually sitting in their email inboxes, spreadsheets and ERP systems," Ardery told Sourcing Journal, arguing that recyclers mistakenly picture "massive, multi-million-dollar hardware overhauls on the sorting floor" when they hear AI, which creates sticker shock and analysis paralysis.
Where are the real adoption barriers?
For regional operators, the blockers are fragmented data and workflow mapping, not capital equipment. Most recyclers run lean teams where institutional knowledge lives in people's heads or across disconnected PDFs, bills of lading and lab reports. Ardery said Applied Industrials starts every deployment with an AI use policy and workflow mapping, because without mapped workflows and clear guardrails, companies cannot move beyond basic chatbots to true operational automation.
What does AI do for quality documentation?
Reaching high-value secondary markets — chemical recyclers or premium fiber-to-fiber manufacturers — requires proving material composition and yield consistency on paper with 100 percent accuracy, Ardery said. In one textile-to-textile recycling workflow, AI agents automatically ingest lab test reports, supplier quotes and packing lists from multiple feedstock suppliers. The system:
- Cross-checks material specifications
- Flags missing safety or technical data sheets (SDS/TDS)
- Tracks exact weight and yield changes, catching a 7 percent weight variance between what was ordered and what arrived after shredding
That automated verification trail lets operators certify batches for high-value buyers without burying quality teams in manual paperwork.
What is the ROI timeline?
Applied Industrials integrates into an operator's existing ERPs, email and logistics systems, meaning virtually zero hardware CapEx. In one deployment, an AI agent found that 2 percent of the client's 2025 FedEx billing was charging the client incorrectly. Freeing skilled finance personnel from manual data review makes payback almost immediate, Ardery said.
How does AI handle DPP and SB 707 compliance?
Regulations like the EU's Digital Product Passport and California's SB 707 are, in Ardery's framing, data-orchestration challenges. Physical sorting capacity counts for nothing if operators cannot digitally prove chain of custody, origin and material composition at every handoff — otherwise the material hits a compliance wall.
AI agents read the unstructured trail of purchase orders, bills of lading, certificates of origin and lab reports generated each time a batch moves from collection to a domestic shredder, onto an ocean carrier and into an EU facility. The agents extract weight, origin and material attributes and assemble a real-time, auditable digital ledger, turning compliance into an automatic byproduct of daily operations.
Can AI match bales to buyers?
For exporters of used clothing under HS 6309, Applied Industrials built AI-driven market intelligence tools, demonstrated as part of the SMART conference, that screen global trade flows and identify vetted wholesale importers down to the city level — in Chile, Uganda or Central America. Connected to real-time demand and pricing intelligence, sorters can dynamically adjust what they pull for domestic resale versus specific export grades, bridging what comes off the sorting belt and what global buyers actively pay a premium for.
What does the next decade look like?
Over five to ten years, Ardery expects end-to-end digital visibility driven by the convergence of physical and agentic AI: robotics, optical sorters and imaging sensors on the warehouse floor, with Applied Industrials acting as the data-orchestration layer translating machine-level data — already including X-ray inventory imaging — into decisions inside legacy ERPs.
For the small and mid-market operators that form the backbone of SMART's membership, his advice was direct: you do not need to wait ten years or spend millions on robotics. Build the AI "harness" around existing operations — automating procurement, export compliance and sales outreach — to protect margins now while constructing the ERP data infrastructure that robotics will later plug into.
via Sourcing Journal (Source)
More from Grace Kim
Show full bio
Correspondent covering regulation, policy and their impact on Light industry and softgoods manufacturing at Softgoods Report.
381 articles