Coaching Plan & Follow-Up Tracker

The problem this kills

You run a QA program. You score calls and chats, you sit down with the agent, you have a good conversation - and then nothing. The notes live in a spreadsheet tab nobody reopens. Three weeks later you can't say whether the coaching changed anything, because the "did it improve?" check never happened. Coaching becomes a one-off chat instead of a loop, and the same agents keep missing the same rubric items.

Most teams don't have a system that connects a coaching conversation to the specific QA scores that triggered it, schedules the check-back, and then puts the before-and-after numbers side by side. So coaching feels busy but can't prove it works.

What you'll build

A private web app for team leads and managers that closes the loop:

• Load QA results for your agents from a CSV or Google Sheet (one row per agent per scoring period, with scores per rubric item).
• Get coaching items proposed automatically from each agent's weakest rubric areas - you stay in control and edit them.
• Finalize a coaching plan with a focus area, agreed actions, and a check-back date that you approve.
• Reminders go out by email (via Resend) as the check-back date approaches, so the follow-up actually happens.
• Record the outcome at the check-back, and see the improvement trend - the same rubric items, before vs. after, so improvement is checkable, not a vibe.

It comes with login, so only your team can get in; row-level security, so a manager and agent only ever see their own coaching notes; an approval gate, so a plan isn't active until the manager signs off; and a full audit trail of who changed what and when.

What's inside the Implementation Plan

The plan is a single file you paste into an AI coding agent (Claude Code), and it builds the whole tool with you, step by step.

It opens by interviewing you about your business - your QA rubric, how your scores are named and stored, how you run coaching today, your typical and peak agent volumes, your approval rules, and your messy edge cases. Then it reads back a short tailored spec, waits for your thumbs-up, and shapes the data model and every later build step around your answers. You get a tool fit to your team, not a generic template.

After the interview, the plan walks you through:

• Setting up the project, database, login, and per-team security.
• Importing your QA scores and mapping them to your real rubric items.
• The auto-proposal of coaching items from weak areas, with a manager review-and-approve gate.
• Scheduling check-backs and wiring up Resend email reminders.
• The outcome capture and the before/after improvement trend view.
• A CSV export of the full coaching log as a built-in fallback - so the tool is fully usable today even with no integration to your existing QA platform.

Every build step ends with a ready-to-copy prompt you paste into the agent.

The governance it includes (this is the point)

This isn't a toy. The plan bakes in the controls a real coaching program needs:

• Login so only your team can use the tool.
• Row-level security so a manager and agent only ever see their own organization's - and their own - coaching records. Notes stay private.
• A human-in-the-loop approval gate: the AI drafts coaching items from the scores, but the manager reviews, edits, and approves the plan before it's active, and confirms the outcome at the check-back before it's marked complete. Nothing is committed without a person.
• A complete audit trail: who created or changed a plan, who approved it, who recorded the outcome, and when.
• Duplicate guards so the same agent + coaching cycle can't be logged twice.

Who it's for

Team leads and managers responsible for agent development in a contact center or support team - anyone who runs QA scoring and wants coaching to be a measurable loop instead of a one-off conversation. No coding experience needed.

You've got this - paste the first prompt and let the agent interview you.