Circular Commerce
90%
verification accuracy with AI-powered photo analysis
LOOPSMART
AI for
circular operations.
We build the systems that run circular business and research models:
order processing, return reconciliation, compliance and fraud prevention. Designed for operators who deal with circular services.
Built for operators, not demos
Our systems handle the messy parts of circular business models before it costs you.
We've run these operations ourselves
Every product started from a real problem we faced: reconciling millions of deposit transactions, verifying returns at scale, or managing finance and risk. We built tools to solve them, then packaged what worked.
PRODUCTS
AI that runs in production, not slide decks.
We license software we built and use. Pricing is typically per-transaction or revenue-share - you pay when it works.
PROVEN RESULTS
Real outcomes from our technology.
Our R&D delivers measurable impact. Here's what partners have achieved with our solutions.
E-Commerce Ops
3x
faster ticket resolution with agent-assist workflows
Finance Automation
50%+
annual savings through automated reconciliation
Risk & Fraud
60%
reduction in chargeback losses via proactive scoring
Results based on partner deployments. Actual outcomes depend on implementation scope and data quality.
xCommerce
Orders, warehouse, shipping, and returns for circular operations.
xTicketing
Verify returns with photos before you ship. Approve, reject, or request more info.
xFinance
Match payments to orders, prepare VAT filings, and build audit trails.
xPricing
Set deposit prices, run experiments, and see what moves margins.
xRisk
Flag suspicious orders and generate evidence packs for chargebacks.
xMobile
Mobile apps for warehouse staff: scanning, picking, quality checks, works offline.
Want a custom track?
If you have a high-leverage problem, we are ready to help.
HOW WE BUILD
AI-native development for AI-first organizations.
Our state-off the art research enables all development and operations phases.
03 Aug 2026
Forecasting Returns Before They Happen: Can LLM-Augmented Models Cut Reverse-Logistics Waste?
AI-driven reverse logistics is already a named McKinsey and Bain playbook, not an unproven idea, and it's one narrow instance of a much bigger shift toward circular business models as a real profit lever. What's still open: does LLM-embedding signal add anything on top of what's already deployed, and why aren't most companies capturing value they already have access to?
21 Jul 2026
Who Owns the Cart? Agentic Shopping Assistants and the Coming Disintermediation of eCommerce Funnels
As AI agents start browsing, comparing, and checking out on behalf of consumers, which parts of the eCommerce funnel are exposed, and which are quietly becoming more defensible? Grounded in real 2025-2026 adoption data, including why OpenAI's in-chat checkout stumbled while Google's Universal Commerce Protocol kept expanding.
UNDER THE HOOD
What we actually use.
The libraries, patterns, and infrastructure that power our products. All built to be reused across deployments.
Automation that works
We use ML where it helps and rules where it doesn't. Everything is tested against real data.
- Search and retrieval: find orders, match documents, answer questions
- Classification: route tickets, detect fraud, categorize transactions
- Accuracy tracking and alerts when things drift
Clean integrations
Our AI connects to what you already use. APIs, webhooks, and documented behavior.
- REST APIs, webhooks, and event feeds
- Reference implementations and migration scripts
- Test suites for integrations
Runs reliably
We know what things cost and what breaks. Everything is logged and monitored.
- Cost tracking and capacity planning
- Logs, metrics, and alerts
- Security controls and audit logs
Documented
You get versioned releases, changelogs, and docs - not a handover meeting and a Slack channel.
- Versioned releases with changelogs
- Test data and accuracy benchmarks
- Runbooks and ownership boundaries
CAREERS
We're hiring.
Small team, real problems. If you like building software that gets used, not just demoed, we should talk.
Applied AI Engineer
Build automation that works: search, classification, and verification systems.
Research Engineer
Run experiments, measure what works, ship what passes.
Full-Stack Software Engineer
Build the ERP, finance, and ticketing products from end to end.
Platform Engineer
Keep things running, track costs, and make deployments boring.
Security Engineer
Audit logs, access controls, and compliance - the stuff that matters.
CONTACT
Let's talk.
Running a circular business model (e.g. refill, exchange, rental) program? Dealing with deposit accounting headaches? Tell us what's broken and we'll show you what we've built.
Tell us about
- What you're running today (refill, exchange, deposit scheme)
- Your current pain points (reconciliation, returns, fraud)
- Rough volume (orders/month, transactions)