Hire an AI production yield and cost reviewer

Everything it can hand back

Your task sets the actual outputs. The specialist picks the methods that cover them.

• The real cost per good unit for each product, after scrap recovery
• A contribution table with profit per unit and for the planned volume
• A capacity check with hours needed against hours available, and any unused hours or overload
• A cost bridge for an input price change, showing the part recovered from customers and the part you absorb
• What changed since the last review, a record for the next one, MISSING lines and any checks that could not run

How it does the job well

• Reconcile to source documents, such as bank statements, ledgers and invoices, before reporting.
• Show the calculation, inputs and period for every figure.
• Label actuals, estimates and forecasts clearly.
• Send tax, legal and regulatory questions to a qualified professional.

Mistakes it avoids

• Mixing cash and accrual figures, or different periods, in one comparison.
• Rounding or formula errors that do not tie back to totals.
• Giving personal investment advice.

What a good result looks like

An example, not your task: A new batch comes in with a lower yield than the last one.

• Changed since last review. Each yield, scrap, price or quantity that moved, with its old and new value and source date.
• outputs/step-1-manufacturing-yield-cost-bridge.md. Input weight tied to good output and scrap, net batch cost after proven scrap recovery, cost per good unit, and a lost-yield sensitivity.
• outputs/step-2-product-mix-capacity-checker.md. Contribution per unit and for the planned volume, bottleneck hours needed against available, and unused hours or overload. The mix stays your proposal.
• Commodity pass-through bridge. Extra input cost per unit, the share passed on to customers, the share you absorb, and the rebuilt contribution. Only when an input price changed.
• Record and MISSING lines. The record for the next review, every gap with the line it stops, and CHECKS MISSED.

How the work is checked

• For every batch, input equals good output plus scrap, in one physical unit.
• Scrap value counted is backed by a dated sale record. Unproven scrap counts at zero value and is listed.
• Cost per good unit, selling price and planned quantities all use the same selling unit.
• Contribution is price less direct cost per unit. Shared fixed costs are not in both the unit margin and the period overhead.
• States which checks passed, failed or could not run, and names missing inputs.

These checks have not been run against a model yet. Your AI uses them to check its own work.

Capabilities

• Compares with the last review. Reads the saved record if it belongs to these products and this period, or sets a first baseline. It compares batch IDs, measured yield, input price dates, the agreed pass-through share, the bottleneck scope and the plan version. If nothing changed and no review is due, it says so and stops.
• Reconciles each batch. Runs the Manufacturing yield cost bridge step. It ties input weight to good output and scrap, values only scrap you can prove, and subtracts that from input plus conversion cash. It divides by good output and converts to your selling unit, keeping the evidence for the conversion.
• Builds contribution and checks capacity. Runs the Product mix capacity checker step. It builds each product's contribution from that cost and your selling price, multiplies by planned units, and compares the hours needed with the hours the bottleneck has. It reports unused hours or overload and lists other limits it did not test.
• Bridges an input price change only if needed. If a supplier invoice changes an input price, it runs the Commodity pass through cost bridge. It works out the extra input cost per unit, applies only the agreed pass-through share to the price, and rebuilds contribution. Old stock keeps its original cost date.
• Checks and saves. Tests the result against the acceptance rules, retries the affected calculation at most twice, then writes the dated files, which steps ran or were skipped, and the new record. The old record is kept, and planned volume is never raised on its own.

How the role works

1. Checks your task, context and available tools.
2. Picks the methods needed and prepares a work plan.
3. Carries out supported tasks, with each method's limits and review gates.
4. Checks the real deliverables and corrects supported faults.
5. Hands back files, evidence and unresolved gaps.

Your specialist's working setup

Download the base specialist without a task form or account. README.md and SETUP.md explain how to use the folder. SOUL.md defines its scope. INTAKE.md asks missing questions inside your AI. PLAYBOOK.md, SKILLS.md, BEST_PRACTICES.md, CHECKS.md and HANDOFF.md guide the work and its checked deliverables.

Complete associated skills, supporting files and licences are included. Choose only the methods needed for the actual task. Keep context, working files and accepted feedback in your AI project.

Use this specialist through MCP or the API

Call prepare_specialist with role_id production-contribution-evidence-reviewer only. No task is required to load it. Your AI asks what job you need done, what you want back and which relevant context it can use.

For task-led hiring, call find_specialist, explain the suggested specialists and let the user accept or change the team.

Complete ZIP: Download base specialist. API: GET /api/v1/specialists/production-contribution-evidence-reviewer/pack. The host AI supplies its own model, tools and execution; a pack does not connect accounts or run a task.

Inputs and helpful context

The specialist starts from the task brief. Only active methods require their inputs.

For the primary method, when used

• For each product, the measured input, good output and scrap quantities for a batch, in the same unit (from your production records)
• Cash received for scrap, with proof such as a buyer's invoice (optional per batch; with no proof it counts no scrap value)
• The batch's input cost and the cash cost of making it (from supplier invoices and your cost records)
• Selling prices and the unit you sell in (from your price list)
• Planned units for each product (from your plan)
• Bottleneck hours needed per unit, and hours available in the same period (from routing sheets or the machine schedule)
• The new invoice price and the share your customers agreed to pay (optional, only if an input price changed)
• The last review's record and the earlier source versions (optional, needed to show what changed)

Helpful context (optional)

• Supplier invoices dated to the stock that was used.
• The agreement or email that sets the pass-through share with customers.
• A named person who owns the plan and will check the yield and the hours.

Tools and access

• A spreadsheet engine and file-capable AI: Builds a workbook with inputs and formulas you can click into, and saves the record file for next time. In plain chat it returns the same tables as text.
• Your batch and machine records: Measured weights and hours are facts. Estimates change the answer, so use records where you have them.
• A spreadsheet engine and file-capable AI: Builds a workbook with inputs, formulas and a checks tab you can click into. In plain chat it returns the same tables as text and says no workbook was made.
• A spreadsheet engine and file-capable AI: Create, recalculate, render and reopen the requested native workbook.
• Node.js 20 or later: Run the bundled dependency-free reference calculation against explicit numeric inputs.

Functions and deliverables

• Production yield and cost reviewer: Connect measured production yield with product costs and capacity. Show changed assumptions without treating a ranking as a production order.
  • The real cost per good unit for each product, after scrap recovery
  • A contribution table with profit per unit and for the planned volume
  • A capacity check with hours needed against hours available, and any unused hours or overload
  • A cost bridge for an input price change, showing the part recovered from customers and the part you absorb
  • What changed since the last review, a record for the next one, MISSING lines and any checks that could not run
• Production yield and capacity reviewer: Turn measured good output and scrap into unit costs. Compare product contributions and check the proposed workload against capacity.
  • Mass reconciliation per batch: input, good output and scrap
  • Net batch cost and cost per good kilogram, converted to your selling unit
  • Contribution per unit and in total for each product
  • Bottleneck hours needed against hours available, with unused hours or overload
  • Missing facts, and checks that could not run
• Raw material price impact calculator: Share material costs, product quantities and price plans. See how cost increases affect the margin left.
  • The extra input cost per unit
  • The price recovered, and the part left unrecovered
  • The new selling price and the margin per unit
  • The margin for the period, beside the original
  • Sources, assumptions and a list of anything missing
• Manufacturing yield and unit cost calculator: Share batch weight, yield, costs and scrap prices. Get net batch cost and cost per good kilogram.
  • Good kilograms and scrap kilograms, which add up to the input
  • The money back from scrap
  • The net batch cost, input plus conversion less scrap
  • The cost per good kilogram
  • A check that good plus scrap equals input and that scrap money is counted once
  • A list of missing facts and checks it could not run
  • A small workbook with the same sums, if your AI can make files
• me.md builder: Answer a few questions, or let your AI use what it remembers. Get one private file that tells any AI who you are.
  • One me.md file with up to ten dated sections
  • A list of facts taken from memory, for you to check
  • A short note of what it removed for privacy
  • The two next steps, to save it and set it up

Tasks you can use it for

• A new batch comes in with a lower yield than the last one.
• You sell scrap and want its value counted correctly, with proof.
• A supplier invoice raises the price of a main input, and part of the rise can go to customers.
• You are choosing how many units of two products to make, and the bottleneck may not have enough hours.

Limits to check before starting

• You make one product with no scrap. A simple margin sum in a spreadsheet is enough.
• You want the best mix. It checks the mix you propose and does not search for a better one. Ask your finance lead or use an optimiser.
• You have no measured batch quantities. The yield lines stop and show MISSING, because a guessed yield changes every later number.
• You want a forecast of demand or sales. It uses your planned units and does not predict what will sell.

Best way to use this role

Give it one clear outcome, the actual files and a definition of done. Keep the work in one AI project so it can reuse checked context and earlier files.

• Add this agent to Claude, ChatGPT or Gemini, with the two pieces it uses. They are the Connect production yield to mix contribution workflow and the Commodity pass through cost bridge skill. An AI that reads files works best. In plain chat it still works, but it cannot save files.
• Say if this is the first run. If not, paste last time's record. Then attach batch records, price lists, supplier invoices and the machine schedule.
• Say which unit you sell in. Check that costs per good unit and planned quantities use that same unit.
• Check that the input weight equals good output plus scrap. Check that scrap value appears only where you gave proof.
• The capacity step covers two products at a time. If you have more, say which two to compare, or run it again for the others.
• Fix every MISSING line and run it again when a yield, an invoice price or a plan changes. Adding it sets up no schedule.

Read it as data

Read the complete role setup as JSON. Connect your AI to the AI Blocks MCP library and read aiblocks://roles/production-contribution-evidence-reviewer for its role instructions and pinned sources.

This is a prepared AI role. No task has run. Your AI carries out authorized work using its available tools and reports the actual deliverables and checks.