Keep glass and door jobs moving from quote to payment
South Jersey Glass & Door ran nine job hand-offs on a whiteboard and two spreadsheets. Salesforce quoted $4,000 to implement plus per-seat licensing forever. Five days after the scoping call the company was running on a custom field-service platform built around its own nine stages: an office pipeline that shows where jobs stall, crew scheduling that cannot double-book, a phone-first technician surface and a database-enforced audit trail. It now carries $4.3M of quoted work, at 86% less than the Salesforce route over five years.
The owner & the bottleneck
Where it started.
An install crew's job changes hands nine times between the first phone call and the check clearing. At this fabricator those hand-offs lived on a whiteboard, in two working spreadsheets and in the managers' heads. Nothing crashed; things went quiet. Quotes sat for weeks waiting on customers nobody re-called, and finished installs sat unbilled because the paperwork hand-off happened out loud or not at all. The owner went looking for software. Salesforce Field Service came back at $4,000 for implementation plus per-seat licensing on top, and the field-service SaaS tools that fit landed between $250 and $600 a month, each carrying features the company would never use and missing the ones it needed.
What we built
The system we put in.
A custom platform that speaks the company's own vocabulary: the nine stages, priority levels, service areas and job-ID formats came straight out of the spreadsheets they already ran on. The office sees a nine-stage pipeline with per-stage aging thresholds and a daily sweep for jobs that have gone quiet. Scheduling checks availability on screen, re-checks on the server and enforces a no-double-booking rule in the database itself. Technicians get a phone-first surface with start and finish, hour logging and photos, and a job cannot close without both. Every change writes an audit entry; roles are enforced in the database; the platform is multi-tenant from day one so acquired companies onboard on the same monthly fee. Built in nine phases with independent review at each gate: 386 commits, 617 automated tests, 49 issues fixed before launch.
The numbers
What changed, measured.
The full story
How it played out.
This is the same company whose search program took it from invisible to page one. See the SEO proof entry.
Resources
Keep learning
Claude Code Plugin Evals: The Traps the Docs Don't Mention
We ran Claude Code's new plugin eval harness on a real plugin: 8 cases, 46 agent runs, about $12. Here are the three ways the results JSON misleads you, what the eval sandbox can and cannot reach, and the five-line hook that fixes the plugin-root variable Bash never sees.
The Benchmark Pointed at Itself
The first paired benchmark of our procedural-memory registry measured a deterministic share under 0.5% on one workload and about 55% on another, caught its own row-append step failing silently, and ran under two conditions that rule out any arm-versus-arm verdict. Here is what held up, what did not, and the six variables the next cohort holds fixed.
Your Subagents Aren't Using Your Skills
We counted how often our AI subagents invoked skills from a several-hundred-skill library: zero of fifteen dispatches. Here is why cold-started agents never browse the library, the three injection levers that work, the spike methodology that proved the hook surface, and the log-first measurement instrument now watching every dispatch.
AI Agent Memory: Why Our Systems Stopped Re-Solving the Same Problems
Most AI automations re-derive the same work every run. Here is how procedural memory, storing the working code rather than a description of it, makes AI agent workflows faster, cheaper, and more predictable.
Rhize Tasks: A Local-First AI Task Planner for Jira and Calendar
Rhize Tasks turns Jira work into a realistic daily plan across Google Calendar and Apple Reminders, with local-first privacy and human approval controls.
We gave our AI agent 500 skills. It needed a graph, not a list.
Flat lists of AI agent skills fail silently as they grow. Every failure we hit was a relationship failure, so we built a generated graph, and the case for generating it came from our own code drifting within hours.
What could this look like
in your business?
Start with a conversation.
Tell us about your operation and what you want to improve. We'll talk through whether a similar approach could help.
No pitch, no obligation.