Skill-Based Routing Matrix

The problem this kills

When a ticket lands, who gets it? On most support teams the honest answer is "whoever grabs it first" - or whoever the lead happens to remember. That lottery quietly costs you: French tickets land with an English-only agent, a hard product question goes to someone who's never touched that product, your strategic account gets a stranger instead of its named owner, and your best performer drowns while two teammates idle.

Spreadsheet-based "assignment by hand" is no better. It's slow, it's inconsistent, and the moment your lead is out sick the whole thing falls apart.

You don't need a six-figure routing engine bolted onto your helpdesk. You need a simple tool that knows your agents' skills and current load, proposes a smart assignment for every ticket, and lets a human approve it.

What you'll build

A private web app for your team that:

• Loads incoming tickets and your agent roster (skills, languages, coverage hours, max-load caps) from a Sheet or CSV.
• Pulls in each agent's current open-ticket count so it never piles work onto someone already buried.
• Uses AI to propose a best-fit assignment for every ticket - balancing product knowledge, language, account ownership, and workload, so it doesn't dump everything on your top performer.
• Shows your lead a review screen: the suggested assignment per ticket, with a one-click reassign.
• Commits assignments only after the lead approves the batch, then emails each agent their new tickets via Resend.

What's inside the Implementation Plan

The plan is a single file you paste into Claude Code (a free AI coding assistant). It walks the AI through building the whole tool, step by step, with a ready-to-copy prompt at the end of each step.

Crucially, the plan opens by interviewing you about your business - your real ticket fields, how you name skills and languages, your account-ownership rules, your typical and peak volumes, your max-load caps, and your messy exceptions. It reflects a short tailored spec back to you and waits for your thumbs-up before building anything. You get a tool shaped to how your team actually works, not a generic template.

Inside you'll find:

• The discovery interview that tailors the data model to your team.
• A clean data model for tickets, agents, skills, and assignments.
• The AI routing logic that balances skill match against workload.
• A lead review-and-approve screen.
• Resend email notifications to agents on approval.
• The full governance layer (below).
• A "No API yet?" fallback so you can build and use the tool today with just a CSV.

The governance it includes (this is the point)

This isn't a toy. The plan builds in the controls that make a tool safe to run on a real support team:

• Login so only your team can open it.
• Row-level security so each org only ever sees its own tickets and roster.
• A complete audit trail - who proposed, who approved, who reassigned, and when.
• A hard human-in-the-loop gate - the AI only ever proposes; nothing is assigned until your lead approves the batch.
• Duplicate guards keyed on ticket ID, so the same ticket can't be processed twice and an already-owned ticket is never silently reassigned without an explicit override.

Who it's for

Team leads and dispatchers running a multi-skill support team - multiple products, multiple languages, named account owners, and uneven workloads. If you're the person who decides who works what, this tool turns that judgment call into a fast, fair, repeatable, and auditable process.

You've got this. Paste the first prompt and let the interview tailor it to your team.