Agent Workload Balancer

The problem this kills

Every morning the same scramble: one agent is buried under thirty open tickets while another has a quiet queue, and nobody notices until an SLA is about to blow. The lead eyeballs a spreadsheet, drags a few tickets around by gut feel, forgets to tell the agents, and accidentally hands a billing escalation to someone who's never touched billing. Work piles up unevenly, the fast agents burn out, and "who's got capacity?" gets answered by whoever shouts loudest.

A workload balancer takes the guesswork out. It looks at who actually has open load right now, what kind of work it is, and how much each person can really take - then it suggests a small, sensible set of moves to even things out, with a reason attached to each one. Your lead stays in charge: nothing moves until a human says yes.

What you'll build

A private web app for your team that:

• Imports your open tickets (with owner and complexity) and your agent roster (with capacity and availability) from a CSV or Google Sheet.
• Calculates each agent's real load - not just ticket count, but weighted by complexity - and compares it against their capacity.
• Detects imbalance and proposes a short list of ticket reassignments, each with a plain-English reason ("Agent A is at 140% of capacity; this medium ticket moves to Agent C, who has room and the right skill").
• Respects skills and current ownership, and deliberately avoids churn - it won't shuffle tickets around just to make a number look tidy.
• Shows the lead a review screen to approve, edit, or reject each proposed move - one human gate before anything is final.
• Records every approved move in an audit trail and emails the affected agents through Resend so they know what landed in their queue.

What's inside the Implementation Plan

• A complete, copy-paste runbook you hand to an AI coding agent (Claude Code) - you don't write the code, you paste prompts and answer questions.
• It opens by interviewing you about your business. Before a single line is built, the plan makes the agent ask how your team actually works - your ticket fields, how you measure complexity, your SLA rules, your peak volumes, your skill tags, and your messy exceptions - then reflects back a short tailored spec for your thumbs-up. The tool is shaped around your team, not a generic template.
• Step-by-step build instructions, each ending in a ready-to-paste prompt.
• The balancing logic, the lead's review-and-approve screen, the audit trail, and the agent notification emails.
• A "No API yet?" fallback so you can build and use the whole thing today with just a spreadsheet - and export a clean reassignment plan as CSV.

The governance it includes (this is the point)

This isn't a toy that yanks tickets around on its own. Every build includes:

• Login so only your team can open the tool.
• Row-level security so people only ever see their own organization's data.
• A full audit trail - who proposed, who approved, what moved, and when.
• A hard human-in-the-loop approval gate - the tool drafts the moves, your lead reviews and approves, and only then is anything recorded or anyone notified.
• Duplicate guards so the same ticket can't be reassigned twice by accident (the dedupe key is the ticket ID per move).

Who it's for

Team leads and dispatchers who own the daily distribution of support work - the people who currently balance the queue by hand every morning and want a faster, fairer, more defensible way to do it.

You've got this - paste the first prompt and let the interview tailor the tool to your team.