Each report is a named AI agent with a defined scope — not a static dashboard. Holiday-aware, weather-aware, and grounded in the sales plan.
⚡
Daily & Weekly Flash
Live
The original anomaly report — chain, region, district, store rollups. Pre-computed bps deltas, 28-day baselines, traffic noise floor. Delivered as Adaptive Cards in Teams; ask follow-ups in the same thread.
@Reporting — weekly flash
Why did Atlanta region underperform?
🔔
Loss Prevention Report
New
22 LP signals against the JMC POS database with scope-aware rollups (chain · district · region · store) and per-employee analysis. Volume floors prevent false positives. Monthly grading.
@Reporting — LP report for last month
Top 5 cashiers flagged this period
📅
Outlook Report
New
Forward-looking weekly briefing. Anchors against the sales plan, frames tailwinds and headwinds, holiday-aware, last-year weather. Anti-single-cause rule — weather and holidays are contributors, never whole explanations.
@Reporting — outlook for next week
What's the chain plan vs. forecast?
🌡️
Ambient Context
Every report ships with the surrounding weather chip (period hi/lo + precip) and holiday awareness. The narrative changes when context changes — product-first openers, traffic noise floor, KPI vs. anomaly distinction enforced.
Every commit, every PR, every job run feeds AI-authored documentation back into the knowledge base. The system gets smarter every day — without anyone writing docs.
Claude-Powered Doc Pipeline
CI workflow runs on every push — reads diffs, generates summaries, publishes to Confluence
Owner-based routing — each repo's docs land in the right space and parent page automatically
Auto-generated repository summary workflow — new contributors land on a current map of the codebase, not stale READMEs
Hangfire job consoles surface LLM token usage and dump malformed responses for triage
The Recursive Loop
More AI usage → more documentation generated → better context for the next AI run
DTLR IQ pulls from the same Confluence — every shipped feature becomes answerable, automatically
DevAgent reads the same docs back — the codebase trains itself on its own evolution
drop a Confluence example screenshot into Images/auto-docs.png
Live Demo · Customer-Facing AI
Where Customers Meet Our AI
🛒
Agentic Commerce & UCP
Live
Product feeds streaming to ChatGPT, Google Search AI, and Microsoft Copilot. Built on the Universal Commerce Protocol — AI agents browse, recommend, and check out on behalf of consumers. We are ahead of the curve as UCP rolls out industry-wide.
Automated customer service across email, chat, and voice. Connected to DTLR systems via MCP & API for live order status, returns, and process documentation — resolving routine cases without a human in the loop.
AI shopping assistant on dtlr.com — intelligent product search, natural-language Q&A about fit and style, and personalized recommendations driven by behavior.
Customers stop visiting websites — their AI agents shop for them. Our product catalog, fulfillment, and service are already wired for that future. The pipes are in place; the volume just keeps growing.
Live Demo · In-Store Experience
Marquee — In-House Signage Platform
Multi-Tenant Digital Signage
Shipped · In Production
Built in-house — manages kiosk displays and audio playback across every retail location
Replaces our third-party digital signage contract — hundreds of thousands per year back on the bottom line
Now integrates with the POS Customer-Facing Display — second-screen experience tied to live transactions
Rust agent on kiosk hardware + Cloudflare Workers + Durable Objects for real-time tenant coordination
We own the platform — new features ship in days, not contract cycles
$100K+
Annual savings vs. the legacy signage contract — plus full creative and technical control.
