--- name: al-366-turn-customer-order-emails-into-ready-to-enter-orders-wit description: "Typing customer orders from emails and attachments into the order system is slow and error-prone. In an Oct 2026 Microsoft story, Unilever described NOVA, five AI agents built on Microsoft Foundry: four read the order email, work out its context and extract the order, and a fifth checks the work, with missing information and exceptions sent to a person. Across 11 pilot markets the first step of order-to-cash dropped from 15 to 20 minutes to about 90 seconds, 95% of orders became touchless, refund claims were processed twice as fast and new customer onboarding time fell 90%. The prompt below sets up the same read-then-check pattern for a small business, using any AI assistant and a person approving every order. Use when the user wants help with order entry from email (operations) for a small business." --- # Turn customer order emails into ready-to-enter orders with a reader agent and a separate checker: pull out every line, check it against your price list and stock, and send only the exceptions to a person Typing customer orders from emails and attachments into the order system is slow and error-prone. In an Oct 2026 Microsoft story, Unilever described NOVA, five AI agents built on Microsoft Foundry: four read the order email, work out its context and extract the order, and a fifth checks the work, with missing information and exceptions sent to a person. Across 11 pilot markets the first step of order-to-cash dropped from 15 to 20 minutes to about 90 seconds, 95% of orders became touchless, refund claims were processed twice as fast and new customer onboarding time fell 90%. The prompt below sets up the same read-then-check pattern for a small business, using any AI assistant and a person approving every order. ## Inputs to collect first Ask the user for anything below that they haven't already given you. Don't guess missing facts. - What we sell, how orders arrive (email body, PDF, spreadsheet… - Product list with SKUs, units and prices, customer list with account… - Order email and attachments ## Instructions Do the task below with the user's inputs in place of each [bracket]. Help me turn customer order emails into clean orders, with a second check before anything is entered. My setup: [what we sell, how orders arrive (email body, PDF, spreadsheet, photo), the system we enter them in]. My reference data: [paste product list with SKUs, units and prices, customer list with account numbers and delivery addresses]. For each order I paste [paste order email and attachments]: 1) Reader step: extract customer, account number, delivery address and date, PO number, and every line (product as written, matched SKU, quantity, unit, price). Show how sure you are of each SKU match. 2) Checker step: re-read the original email and check the extraction line by line: wrong or unmatched SKUs, unit mix-ups (cases vs single items), prices that differ from my list, missing PO or delivery date, duplicate orders, unusual quantities compared with this customer's normal orders. 3) Output: a ready-to-enter order table, a list of exceptions with the exact question to ask the customer, and a short reply email confirming the order or asking for what's missing. Mark the order 'ready' only if the checker found no issues; otherwise 'needs a person'. Don't send anything or enter orders yourself. ## Safety checks (share these with the user) - Check every figure, price and total against your own records before anything is sent, posted or booked. A person approves refunds, discounts, payments and price changes. - Remove names, contact details, account numbers and other personal data the task doesn't need before you paste, and check your AI tool's data and training settings. - Read and edit the draft before it reaches a customer or goes public. If you later automate this prompt, start with "draft and notify me", not "send automatically". - Check names, dates, numbers and any claim against your own source material before you use the answer. AI can misread or invent details. - Say clearly which parts need a person to review or approve before anything is sent, paid, posted or filed. ## About this skill - Works with: Microsoft Foundry, ChatGPT, Claude, Microsoft Excel - Based on this real-world example: https://ukstories.microsoft.com/features/unilevers-multi-agent-creating-customer-value-at-global-scale/ - Free prompt from AgentPlaybook Real Estate: https://agentic-library.netlify.app/prompt/al-366-turn-customer-order-emails-into-ready-to-enter-orders-with-a/ - To install in an agent that supports Agent Skills, save this file as SKILL.md in a folder with the same name as `name` above (for example in your skills directory). Install steps for Claude.ai, Claude Code, Cursor and Codex: https://agentic-library.netlify.app/skills/