A major ASX-listed mining & mining-services group in Western Australia tracks ship-to-ship ore transfers where every minute of delay costs ~$3,000. The manager had no dashboard, no analysis, no way to know where to look first. Just a Word document, emailed around.
Vessel crew tap a cell to timestamp each activity, add a comment, hit send. Data lands in S3, fires an email notification per submission, and rolls up into an automatically generated daily report for the manager.
Every minute of delay costs ~$3,000.
We took visibility from zero to daily. One avoided delay event per week pays for the system many times over.
A large mining company saw demand spread from developers asking about Git and CI/CD to logistics planners wanting to automate workflows against Salesforce, Coupa, and internal databases. They needed someone to actually meet people where they were.
For developers: a structured runbook on agent frameworks, MCP servers, and CI/CD. For non-technical staff: an hour beside them with Claude Code and Claude for Excel, plus scripted setups so the next person self-serves. The rule for everyone — if there's a CLI path, take it.
Logistics planning saves 70 hours / week from autonomous automations alone.
On top of that, on-the-tools work runs roughly 3× faster per team member. Non-technical staff now ship their own automations.
A large mining & mining-services company ran ServiceNow for ~150 IT staff against 74,000 tickets a year — 20k by phone, 65k at the tech bar, 47.5k in the portal. Vague tickets like "my computer has a blue screen" forced manual triage on every one.
We wrapped the ServiceNow API as an MCP server and built an agent on top, shipped as a Next.js app pinned in Microsoft Teams for everyone in the company. The agent asks follow-ups, attempts a fix, and only escalates with a properly-triaged ticket routed to the right group. Managers action approvals through the same surface.
Common requests resolved in ~10 minutes, down from 3+ days.
Cost per ticket cut ~50%. Many requests never become a ticket at all — the agent resolves them inside the chat.
A Perth-based software consultancy for mining, construction & engineering. Designs took 4–5 months to reach production and dev-env setup ran into hours.
We installed AI tooling in their environment, wired it to Jira and Figma, then live-built a receipt-scanning feature in their PHP app inside the session. Then we pulled their MVP epic and planned the full execution order across the team — in one sitting.
Design-to-production collapsed from 4–5 months to days.
Dev-env setup: hours → 30 minutes. Feature planning: multi-day meetings → a single session. They booked a second workshop unprompted.