Monday Morning, Sorted
Monday Morning, Sorted
The AI reads the weekend pileup in your inbox and sorts it: what needs a reply from you, what has a deadline buried in it, and what's just FYI. It drafts replies for the routine ones, and they stay drafts. Nothing leaves without your click.
Connects to your accounts
It reads and drafts, and never sends. The AI reads your mail and writes drafts into your Drafts folder. Sending stays with you, one click per message. The assistant walks you through the permission screen before you approve it, and shows you where the disconnect button is.
What it will ask you
What you read your email in, one sentence about what counts as urgent in your world, when your work week ends, and which routine replies you're happy for it to draft. It helps you write the urgent sentence, which is the part that does all the work. You never paste in an email.
Copy the whole block below and paste it into your AI chat. This is the one recipe that connects to your email, so the assistant starts by working out whether your AI can do that at all, and tells you plainly if it can't rather than letting you find out halfway through. Then it walks you through the permission screen one click at a time, and writes you a Monday instruction you can reuse every week.
The whole recipe, in one paste
You are my setup assistant. I work at a nonprofit. I am not technical, and I don't want to become technical today. Your job is to set the following up FOR me, asking me only questions a non-technical person can answer. WHAT WE'RE MAKING A Monday-morning routine. Each week you read the weekend pileup in my inbox and sort it into three piles: what needs a reply from me, what has a deadline buried in it, and what's just FYI. You draft the routine replies and leave them sitting in my drafts. I send. By the end I should have a saved instruction I can run again every Monday without rebuilding any of this. What this takes: about an hour, and permission to connect my email, which is an all-or-nothing grant. HOW TO WORK WITH ME - Ask me ONE question at a time and wait for my answer before asking the next. Count the questions below and tell me the exact number, and say a follow-up or two may come up. - Before anything else, say which AI product you believe I'm talking to you in (for example ChatGPT, Claude, Microsoft Copilot, or Gemini) and ask me to confirm, then ask whether I'm on a free or paid plan. Never skip the plan question, even when the product is obvious. Then work out whether it can connect to my email at all. If it can't, say so plainly now and offer the fallback: I paste in a list of senders and subject lines and you sort that instead. Don't let me discover a missing feature halfway through. If what we set up is meant for other people too, ask me now who else needs it, and if this product and plan can't share it with them, tell me before we build anything and name the fallback. - Don't assume what software I use. Ask me what I read my email in, and use the exact button names for THAT, not for some other product. - Never ask me a technical question directly. Ask the everyday version and work out the rest yourself. For example: do NOT ask me which permissions or scopes to request. Instead tell me in one plain sentence what the AI will be able to see and do once I approve it, and ask whether I'm comfortable with that. - If I ask you a question at any point, answer it in plain language, then pick up exactly where we left off. - If an instruction doesn't match what I'm seeing, ask me to describe what's on my screen and work from that. - When you give me instructions to do outside this chat, give ONE step at a time and check that it worked before the next. - When you ask me to paste a big block of text into a document or a settings box, tell me first to paste without formatting, and to use a copy icon where one exists instead of selecting the text by hand. QUESTIONS YOU'LL NEED ANSWERED (in your own words, one at a time) 1. What I read my email in, and whether I've ever connected it to this AI before. 2. What counts as urgent in my world, in one sentence. Help me write it: offer an example like "anything from a board member or a funder, or anything about this week's site visits", then push me to name the actual people and topics that can't wait at MY org. This sentence does more work than anything else here, so don't let me leave it vague. 3. When my work week ends, so you know how far back to look. 4. Which routine replies I'm happy for you to draft, and anything you must never draft. WHAT TO SET UP (this part is for you, not me) Two things: the connection, and a reusable instruction. The connection: walk me through connecting my email inside my AI product, ONE screen at a time, using the exact names in my product. Before I approve anything, tell me in plain language what the AI will be able to read and do, that the permission is all-or-nothing so it technically includes send and delete even though you will only ever use read and draft, and say out loud that it will create drafts and never send. Tell me to check which account the picker is showing before I click Allow, since a browser can offer an old signed-in account, and to switch first if it is not my work one. Show me where the disconnect button is. Tell me that disconnecting stops future access but does not erase past conversations, so I should look at my chat-history and data-retention settings while I'm there. Then write me a reusable instruction, filling in my answers, and tell me where to save it so I'll find it on a Monday: "Read my inbox from [WEEK END] to now, unread messages only, and sort every one into three piles. A: NEEDS A REPLY FROM ME, a real person is waiting on my answer. B: HAS A DEADLINE HIDING IN IT, a date, due-by, or RSVP buried anywhere in the message; pull each date out and list this pile soonest-first. C: FYI, no action needed, one line each. In my world, urgent means [URGENT]. Put urgent items at the top of their pile and mark them. For the routine messages in pile A ([ROUTINE]), write a reply and save it as a DRAFT ONLY. Never send anything; I review and send everything myself. If you are unsure whether a message is routine, don't draft it, just flag it for me." Write it so it still works next week and the week after: no dates baked in, nothing that's only true about this week. Keep it short: the instructions box in Projects and similar tools often has a character limit, so write it concisely the first time instead of padding it, to avoid shortening and re-saving it later. AFTER IT'S SET UP, WALK ME THROUGH 1. Running it once now, on my real inbox, with you. 2. Checking the piles against my own judgment. Have me read the FYI pile in full this first time, because that's where a miss would hide. If anything landed in the wrong pile, fix my urgent sentence with me and save the new version. 3. Opening my drafts and reading each reply as if a new staff member had written it: edit freely, delete the ones that miss, send the rest myself. Confirm with me that nothing went out on its own. 4. The habit: where the instruction is saved, and running it Monday morning before anything else. RULES - Nothing goes out without my click. You draft, I send. Never offer to send on my behalf, and never drop the draft-only line from the saved instruction, including later if I ask you to adapt this routine for something else. - The sorting is a smart guess, not a rule. Tell me plainly that for the first few weeks I should read the FYI pile myself before archiving anything. - If my AI can't connect to my email, don't ask me to paste in the full text of messages that might contain personal information about clients. Senders and subject lines are enough for sorting. - If an organizational setting blocks a step (sharing, permissions, an admin restriction), never suggest a personal account or any other way around it. The only options are asking whoever administers that setting, or a different method entirely. Start now by telling me, in two sentences, what we're going to set up together, then ask your first question.
The hard rule of connected work: nothing goes out without your click. The AI drafts and you send. That rule is written into the prompt twice, once for this chat and once for the instruction you save, and the assistant is told to keep it there even if you later adapt the routine to something else.
The sorting is a smart guess. The assistant will tell you this, but it's worth hearing twice: for the first few weeks, scan the FYI pile yourself before archiving anything, since that's where a miss would hide. When you catch one, tighten your "urgent" sentence. Once the piles match your judgment, you can rely on them.
What you walk away with
94 unread messages turned into three ordered piles and a Drafts folder of replies ready to review, before your coffee gets cold, plus a saved instruction that does it again next Monday. You also pick up the habit that makes every connected workflow safe: the AI prepares, you approve, and you send.