Work

Systems in production, not portfolio pieces.

The systems below run real licensed cannabis businesses today. We label everything honestly: client work is client work, and concept demonstrations are clearly marked.

Client work, in production

Licensed dispensary · New York

Cannabis Cowboy

An ongoing operations partnership covering the systems a modern dispensary runs on:

  • Operations dashboard with per-brand sales, margin, velocity, and inventory analytics pulled from the POS, plus deal planning and an interactive deals-and-events calendar.
  • Deal lifecycle automation: promotions reviewed by a person, then published to the Flowhub POS and the public website together, retired on schedule, and measured against pre-deal baselines.
  • Team alerts in Discord: low-stock, out-of-stock, and restock notifications from live inventory, aggregated by product to avoid batch-level false alarms.
  • Compliance-gated content pipeline: AI-drafted educational articles held for automated compliance checks and human approval before anything publishes.
  • Customer data hygiene: loyalty and consent reconciliation across POS and marketing platforms so outbound messaging counts real opt-ins.

Licensed dispensary · Vermont

Float On Cannabis Company

Website operations and an editorial content system tuned to Vermont:

  • Compliance-reviewed publishing pipeline producing localized educational content with human approval gates.
  • Menu-aware content: articles recommend products actually in stock, driven by live menu snapshots.
  • Site reliability and deploy automation so publishing never depends on a developer being awake.

Cannabis brand

Toke Folks

A complete brand back office, delivered as a subscription platform:

  • Back-office platform covering retail accounts, orders, fulfillment, and reporting in one place, replacing spreadsheet-and-email operations.
  • Publish pipeline connecting back-office data to the public web presence.
  • Built lightweight on purpose: fast, dependency-minimal, and maintainable by any developer.

We do not publish revenue figures or performance metrics for client systems. What we can say is what each system does and that it is in daily use. On a call, we are happy to go deeper into any of these.

Concept demonstrations

Example solutions, clearly labeled.

These illustrate systems we can build. They are not paid client engagements, and we never present them as such.

Concept demonstration

A professional services firm turns inbound documents into structured work

The problem
Intake documents arrive by email, get read by whoever is free, and occasionally slip through entirely.
The system
AI-assisted intake that extracts key fields from each document, opens a structured work item, and routes it, with a person approving every classification.
Connects to
Email, document storage, practice management tools.
Human oversight
A reviewer confirms extraction and routing before anything becomes a commitment.
What improves
Nothing slips, intake time drops, and the team starts each day with a triaged queue instead of an inbox.
Worth knowing
Extraction accuracy varies by document quality; the review step is what makes the system trustworthy.

Concept demonstration

A founder validates a product with a working prototype

The problem
A strong product idea, a willing pilot customer, and no way to test it short of a six-figure build.
The system
A prototype sprint delivering the core workflow as working software: enough to put in front of the pilot customer and learn.
Connects to
Kept minimal by design; simulated where honest.
Human oversight
Structured pilot feedback with defined success criteria.
What improves
A build, adjust, or stop decision made on evidence for a fraction of the full cost.
Worth knowing
Prototype code is for learning; a production build is a deliberate second step.
Case studies

Detailed case studies are being written with our clients now.

We publish case studies only with client review and approval. Until then, the summaries above describe what is actually running. Want specifics? Ask us on a call.

Want a system like one of these?

Tell us which one sounds like your operation. We will walk you through how it works and what a version built for you would involve.

Discuss Your Project