The company knowledge platform: documents, curated facts, and 60+ secure read-only connections into databases, code, and documentation feed every AI surface — Teams, DREAM, and developer tools — with single sign-on and role-based control over who can teach it. Built on the Model Context Protocol (MCP), the emerging industry standard for connecting AI to business systems. The #1 contributor to development-team productivity gains this period.

One AI reporting engine now covers both sides of store performance: twenty-two automated shrink and fraud signals rolled up by cashier, store, and district with monthly graded report cards, alongside a daily sales flash and forward outlook measured against plan and enriched with weather and local-event context. Upgraded to a frontier-class model in July.

A dual-purpose in-store platform in active development: a visual merchandising tool and an in-store utility for associates’ handheld devices. Not yet launched, but with substantial progress this period.

An in-house, multi-tenant digital signage and audio platform now running across stores, replacing a six-figure annual vendor contract. Owning the platform only penciled out because AI-assisted development made building it affordable for a small team.

A shared knowledge base spanning the JD North America banners (DTLR, Finish Line, Shoe Palace, Hibbett), built so each company’s AI assistants draw on the same institutional knowledge. In use today for the 905 Cross-Dock Project — the first phase of the JDNA North America Supply Chain Project.

An AI triage layer for engineering work: it watches project boards and error feeds, classifies incoming issues, gathers context, drafts scoped implementation plans, and keeps boards current on its own.

DTLR product feeds are live in ChatGPT, Google AI, and Microsoft Copilot shopping experiences through Shopify, and a structured-data layer across the storefront makes the catalog readable to AI shopping agents. We are partnering with these platforms to continuously monitor and expand reporting and to leverage every capability as agentic shopping matures.

The AI customer-service agent now resolves issues end-to-end on real orders — updating shipping addresses, resending store-pickup PINs, writing notes back to order systems, and canceling orders — governed by rules that decide when the bot is allowed to act.

An “Ask AI” conversational shopping assistant launched on dtlr.com in June, backed by AI-powered search, suggestions, and a custom filtering experience refined continuously through August.

The company assistant is now embedded in DREAM with single sign-on identity, conversation history, answers aware of the page you’re on, and one-click export of results to Excel — the same brain employees already use in Teams.

Managed AI agents running in the cloud take on development and research work directly — offloading it from local developer machines, so development and ongoing research continue outside business hours.

An automated pipeline generates and centralizes documentation across every code repository, keeping the system knowledge that AI tools depend on current without engineer hours.

Organization-managed settings for developer AI tools and per-developer usage dashboards — adoption is measured and governed rather than ad hoc.

AI coding agents now draft and open changes directly from the team’s ticketing and code-review workflow; dozens of AI-authored features have merged across DTLR systems since May.

Analyses that used to take weeks now take a day — most recently, the 30-store DoorDash Marketplace pilot was selected through AI analysis of a full year of delivery demand, fulfillment speed, and geographic coverage data.

An AI markdown-and-clearance application — recommendations with confidence scores, risk heatmaps, and scenario modeling (your strategy vs. AI-optimal vs. do nothing). No new work this period; the next phase will be taken up in partnership with the broader JDNA team.

The store register platform advanced through five version upgrades this year, with payment, receipt, and gift-card hardening throughout, was extended to run the Shoe Palace banner, and began its move to the cloud — architecture, identity integration, replication, and pilot-store readiness are complete, with cutover activities underway. The headline capability — customer order entry inside the POS, with split tender, returns by order, and shipping options — is nearly ready, bringing omnichannel orders onto the register itself.

TikTok Shop was built from scratch and went live June 16 — orders flow straight to the OMS with full catalog, inventory, dispute, and finance reconciliation, already roughly 3% of e-commerce. The DoorDash Marketplace integration is built and launching with a 30-store pilot, Instacart expanded with City Gear onboarded, and Klarna and Zip buy-now-pay-later joined checkout.

Customers now receive invitations to reserve high-demand release product ahead of launch. Orders are captured in advance, so release-day fulfillment ships immediately and far more efficiently — while safeguards and protocols already in place ensure nothing ships early. The program is live, on the web and in the Tapcart app.

A six-month storefront redesign shipped — new navigation, cart, product pages, a rebuilt Releases experience, and a full accessibility pass — with the Tapcart mobile app brought to parity ahead of holiday.

The internal DREAM portal was rebuilt end to end across all 164 screens — new design system, mobile, offline support — and gained major operational utilities: carton tracking with cost-and-retail financials, an RFID inventory management suite with shrink analytics, and rebuilt KPI dashboards with channel-level sales breakdowns.

The Aptos OMS was upgraded to the latest version with a clean July cutover, and the order pipeline from Shopify into Aptos was rewritten for resilience — the backbone under BOPIS, TikTok, DoorDash, and BNPL. The remaining core-apps upgrade is the next phase.

The company-wide carrier conversion from FedEx to UPS completed, with every label — warehouse, e-commerce, DREAM, and in-store operations — now funneling through our own Label API. Automated address verification holds undeliverable addresses before they ship, and shipping visibility and configuration options were extended into DREAM. Owning the label layer positions us for any future shipping changes.

Complete the cloud migration through pilot validation and production rollout, then launch customer order entry inside JumpMind — putting the full omnichannel catalog on every register.

The next-generation order management platform — one of the largest strategic investments on the FY27 roadmap, underpinning future omnichannel, inventory, and fulfillment capability.

Open DTLR product to DoorDash’s customer base, starting with a 30-store pilot selected on demand, fulfillment speed, and regional coverage — with Uber fulfillment optimization and additional marketplace integrations to follow.

Cross-banner supply-chain integration with Finish Line, Shoe Palace, and Hibbett off the shared Alabaster DC, beginning with the 905 Cross-Dock Project — alongside Hanover conveyor modernization and continued warehouse improvements.

Customer-initiated cancellations, richer customer-service tooling, search and recommendation improvements, continued promotions and discount modernization, and buy-now-pay-later expansion.

An automated new-store opening pipeline in DREAM — playbooks, task sequencing, and coordinated workflows for every new location.