Guide 1

Your morning briefing: what needs you today, at a glance, every morning

The guide from the video, with the full AI instruction ready to copy.

1. What you'll end up with

A morning briefing for your business. One overview that shows you every morning:

which customers are still waiting for a reply, and which of them are urgent,
which quotes you need to chase before the customer goes elsewhere,
what's overdue or due today, for you and your team,
what got done since yesterday.

At the very top: the three things to start with.

You build the briefing once, in a conversation with an AI. It asks where your enquiries, quotes and tasks live today, and when something counts as urgent for you. At the end you get your own morning instruction. Every morning you paste it into an AI chat along with your lists, and the AI sorts everything by your rules. You also get a version for your team, and a plan for what someone would need to set up so the briefing later arrives every morning on its own.

Who it's for: business owners and the self-employed who start every morning digging through their inbox, spreadsheets and sticky notes before they know what matters today. No coding needed.
Time: a quiet fifteen minutes for the conversation. The AI asks you 15 questions at most, one at a time. After that, every morning: paste your lists, send, read.
What you need:
- an AI chat such as Claude (claude.ai) or ChatGPT (chatgpt.com). Both have a free plan. On Claude, every plan has usage limits that, according to its pricing page, reset on a rolling five-hour window. If the conversation stops halfway, pick it up later in the same chat.
- your lists: open customer enquiries, quotes you've sent, tasks. It doesn't matter where they live today: inbox, invoicing software, Excel, sticky notes.

2. Step by step

Set it up once

Open your AI chat and sign in. In the browser on your computer (claude.ai or chatgpt.com) or in the app on your phone.
Before you start, stop your chats from being used for training.
- Claude: go to claude.ai/settings/data-privacy-controls and switch off "Model Improvement". Or use an Incognito chat: Anthropic says those aren't used for training.
- ChatGPT: your profile → Settings → Data controls → switch off "Improve the model for everyone". Or use a Temporary Chat: OpenAI says those aren't used for training and are kept for up to 30 days.
Copy the AI instruction from section 3 into your notes app and fill in the "What you already know about me" block. Replace everything in [SQUARE BRACKETS] with your own details. Leave out what you don't know; the AI will ask.
Open a new chat and paste the whole instruction as your first message. Send it.
Answer the questions. Keep it short and honest. "Not sure" is a perfectly good answer. On your phone, you can use your keyboard's dictation and just talk.
- "skip" skips a question.
- "done" ends the interview straight away. You get the result with whatever is there so far, and any missing numbers show up as placeholders.
Save your morning instruction (section 4 of the result) in your notes app. You'll use it every morning from now on.
Write your lists once using the template from section 3 of the result: one line per item, customers as "Customer A", staff by their role, always with a date.

Every morning

Open a new chat.
Paste your morning instruction and put today's date at the top. This matters: the AI counts every day from that date.
Paste your lists underneath. Only what's still open, plus what got done since yesterday.
Send it.
Look at the last line first: "Check: X items in · X items in the briefing". If both numbers match, nothing went missing.
Start with "Do first today".

3. The AI instruction, ready to copy

Copy everything in the box, fill in the placeholders, and paste it as the first message in a new chat.

You're my advisor on how small businesses run day to day. Together we'll build my morning briefing: one overview that shows me every morning, at a glance, which customers are waiting for a reply, which quotes I need to follow up on, what's open or overdue for me and my team, and what got done since yesterday. At the end you write me a ready-made morning instruction that I paste into an AI chat every morning, together with my lists.

== WHAT YOU ALREADY KNOW ABOUT ME ==
(Ask about anything left empty. Don't ask again about anything filled in.)
- Industry and what we sell: [FOR EXAMPLE: ACCOUNTING PRACTICE, JOINERY, AGENCY]
- Who works here (roles only, no names): [FOR EXAMPLE: ME, 1 OFFICE MANAGER, 3 FITTERS]
- Where customer enquiries come in: [FOR EXAMPLE: EMAIL, PHONE, WHATSAPP, WEBSITE FORM]
- Where my quotes live: [FOR EXAMPLE: INVOICING SOFTWARE, EXCEL, A FOLDER IN MY INBOX]
- Where tasks live, mine and the team's: [FOR EXAMPLE: TO-DO APP, EXCEL, NOTEPAD, IN MY HEAD]
- What slipped through the cracks recently: [IN YOUR OWN WORDS]

== HOW TO RUN THE INTERVIEW ==
1. Explain in two sentences what we're about to do. Tell me: 15 questions at most, short answers are fine, "not sure" is allowed, "skip" skips a question, "done" takes us straight to the result. Then ask the first question.
2. Always ask exactly ONE question per message. No forms, no lists of questions. With each question, give two or three short example answers from my industry so I can answer faster.
3. Order, each question in its own message:
a) Ask about whatever is missing in the block above.
b) "How do you know today that a customer is still waiting for a reply from you?"
c) "After how many days without a reply does a waiting customer become urgent for you?"
d) "How do you know a quote has gone out and you haven't heard back?"
e) "After how many days without a response do you follow up on a quote?"
f) "Who on your team should get their own overview every morning?" Roles only, no names. If I work alone, skip f.
g) "How do you know what got done since yesterday?"
h) "What else needs to be at the very top every morning?" (for example today's appointments, callbacks, deliveries)
4. If an answer is unclear, ask once more. If I contradict myself, point it out. Never guess.
5. Keep it short. No summary after every answer, no compliments.

== HOW THE BRIEFING SORTS ==
You write these rules into my morning instruction. The day limits C (customer urgent) and Q (follow up on quote) come from my answers.
- Count calendar days, starting from the date at the top of the morning instruction.
- Every item from my lists goes into exactly ONE section:
1. Overdue: the due date is before today. Add how many days overdue.
2. Customers waiting: an enquiry we haven't replied to yet. From C days on, put "urgent" in front. Oldest first.
3. Quotes to follow up: quote sent, no response for Q days or more. Oldest first.
4. Due today.
5. Waiting on others: we can only carry on once someone else delivers something.
6. On the radar: anything still open that fits none of the sections above (for example due later, or a quote younger than Q days).
7. Done since yesterday: done, dated yesterday or today. Anything done earlier you only count: "Done earlier: X items".
8. Date missing: if an item lacks the date the sorting needs, it goes here, even if it would otherwise fit another section. The same goes for done items without a date.
- If an item fits several sections, the section with the lower number wins.
- Team items go into the matching section with the role in front, for example "[Office]".
- At the very top: "Do first today", at most three items, picked in exactly this order: urgent waiting customers (oldest first), then overdue items (most overdue first), then quotes to follow up (oldest first). These three also stay in their own section.
- Right below it comes a block for whatever I named in question h (for example today's appointments and deliveries), with a heading in my words. If I named nothing there, this block is left out.
- Last line of the briefing: "Check: X items in · X items in the briefing" (not counting the repeats in "Do first today"). If the numbers don't match, the AI finds the mistake before it answers.

== FIXED RULES ==
- Don't invent numbers or items. C and Q come from me. If one is missing, the morning instruction says [C: YOUR NUMBER] or [Q: YOUR NUMBER], and it goes into the last section.
- Don't claim that a specific piece of software can do something (for example export a list or send emails on its own). Describe what's needed and mark open points "to be checked".
- No promises like "fully automatic" or "never again". The briefing shows, I decide. The AI sends nothing to customers.
- Don't answer legal questions (data protection, employment law, for example whether I'm allowed to track my staff's tasks this way). Mark them "check with an expert".
- Write so that someone without any tech background understands every word. Use a technical term only when you have to, with an explanation of five words or fewer.
- In my lists, customers appear as "Customer A", "Customer B" and staff by their role. If I paste names, phone numbers, addresses or other personal data, tell me once that they don't belong here.

== WHAT THE RESULT LOOKS LIKE ==
# Your morning briefing
## 1. Your briefing at a glance
The sections in order, one line each: name of the section · what goes in, with my day limits.
## 2. Where the items come from
Table with these columns: Section | Where it lives today | How it gets into the chat each morning | Who keeps it up to date
## 3. How to write each item
One template line: Type | What | For whom (Customer A or role) | Date | Due | Status. Plus three example lines from my industry, marked "Example".
## 4. Your morning instruction to copy
The finished instruction in one single copy block. It starts with "Today is [DATE]." and contains: every section and rule from above with my day limits, "Don't invent anything. Don't leave anything out.", the check line, and at the end "[PASTE YOUR LISTS HERE]". It has to work today in an AI chat, with nothing to install.
## 5. The team version
For each role from question f, a short instruction: the same rules, but only that role's items. If I work alone, write "Not needed".
## 6. Later, on its own
What someone would have to set up so the briefing arrives every morning by itself. Level A = I paste it myself each morning (works today) · Level B = my lists live in one fixed place, always laid out the same way · Level C = a program collects the lists and sends the briefing every morning by itself (someone has to set that up). For each source from section 2: what's needed, open points marked "to be checked".
## 7. What you assumed and what's still open

== BEFORE YOU SEND THE RESULT ==
Check every point. If one isn't met, fix it first:
- C and Q come from me or appear as placeholders.
- Every section has exactly one rule, and the tie-break order is in the morning instruction.
- The morning instruction starts with "Today is [DATE]." and contains "Don't invent anything. Don't leave anything out." and the check line.
- The morning instruction works today with just an AI chat.
- No software is credited with a capability. Open points say "to be checked".
- No real names, phone numbers or addresses, not even in examples.
- Anything I didn't answer is marked "to be checked" or "open", not guessed.
- No technical term without an explanation.
End with one line: "Self-check: all points met", or which point is open and why.

4. Follow-up instructions

Keep going in the same chat, so the AI already knows your business.

a) Fine-tune it after a week

I've used the morning instruction for [NUMBER] mornings now. Here's what I noticed: [FOR EXAMPLE: ITEM X LANDED IN THE WRONG SECTION, THE BRIEFING IS TOO LONG, I NEED A SECTION FOR COMPLAINTS]. Suggest which rule you'd change for each point and explain every change in one sentence. Then rewrite the whole morning instruction in one copy block. Don't invent numbers. Put any new day limits in as [PLACEHOLDERS] for me to fill in.

b) What would need connecting so it arrives on its own?

For my morning briefing: which information has to get from where to where, so it can arrive every morning on its own? Make a list: where it comes from, what form it's in today, who keeps it up to date. For each one, tell me whether it already lives in one fixed place with the same layout, or whether I need to change that first. Don't claim that a specific piece of software can do something. Instead, write down the questions I need to ask my software provider or whoever sets this up for me.

c) The how-to for your team

Write a one-page how-to for my team: what the morning briefing is for, how everyone writes their items (the template from section 3, with two examples from our day-to-day), that every item always has a date, that customers are written as "Customer A" and never by name, and who to ask if something in the briefing looks wrong. Plain language, friendly and direct.

d) Have a second AI check the briefing Open a new chat and paste this:

You're checking another AI's work.
Today is [DATE].
My lists: [YOUR LISTS].
The briefing that was made from them: [THE BRIEFING].
Check:
1) Which item from my lists is missing from the briefing?
2) What's in the briefing that isn't in my lists?
3) Which day counts are wrong?
4) Which item is in the wrong section, by the rules in the briefing instruction: [YOUR MORNING INSTRUCTION]?
Give me the list of mistakes first. If there are none, write "No mistakes found". Don't add anything of your own.

5. Example: what it looks like

MADE-UP EXAMPLE. This practice doesn't exist. Every customer, date and number is invented, just to show how the instruction works.

The filled-in block (three fields deliberately left empty; the AI asked about them):

- Industry and what we sell: Bookkeeping practice for small businesses
- Who works here (roles only, no names):
- Where customer enquiries come in: email, phone, website form
- Where my quotes live: Word documents in a shared folder
- Where tasks live, mine and the team's:
- What slipped through the cracks recently:

An excerpt from the result. This is what came out of our test run with Claude on 15 October 2026. The made-up practice's answers were given by a second AI, working from a profile with invented details. After 11 questions, this was the result. Shortened, otherwise unchanged.

1. Your briefing at a glance

Your day limits: C = 1 day (a waiting customer becomes urgent) · Q = 7 days (a quote needs a follow-up). Q is "for now", as you said.

#	Section	What goes in
Top	Do first today	At most 3 items: urgent waiting customers (oldest first), then overdue items (most overdue first), then quotes to follow up (oldest first)
Top	Client deadlines this week (tax and payroll)	Your own block. It covers the current week, Monday to Sunday (see section 7)
1	Overdue	Due date before today, with the number of days overdue
2	Customers waiting	Enquiry not replied to yet. "Urgent" from 1 day on. Oldest first
3	Quotes to follow up	Quote sent, no response for 7 days or more. Oldest first
4	Due today	Due date is today
5	Waiting on others	We can only continue once someone else delivers
6	On the radar	Open, but fits nothing above (due later, or a quote younger than 7 days)
7	Done since yesterday	Done, dated yesterday or today. Anything older is only counted
8	Date missing	The date needed for sorting is missing, including done items without a date

(Sections 2 to 6 shortened. The morning instruction itself came out at 31 lines in this run.)

7. What you assumed and what's still open (excerpt)

Q = 7 days is a starting value you chose, and you may change it.
"This week" is read as the current calendar week, Monday to Sunday. You weren't sure, so please confirm it.
Website form enquiries weren't covered. I didn't invent a source.
Staff tasks and the team overviews: whether you may track your staff's tasks this way is a question for an expert. Check with an expert.

And this is what one morning looked like. We pasted the morning instruction from this result into a new chat, with a made-up list of 15 items and the date 15 October 2026. The answer, unchanged:

Do first today

URGENT: Customer A, monthly bookkeeping for a café (received 2026-10-13, 2 days without a reply)

URGENT: Customer B, payroll question (received 2026-10-14, 1 day without a reply)

[Bookkeeper 1] Reconcile bank account, Customer L (2 days overdue)

Client deadlines this week (tax and payroll), Mon 12 Oct to Sun 18 Oct None this week

1. Overdue

[Bookkeeper 1] Reconcile bank account, Customer L: due 2026-10-13, 2 days overdue

Review payroll run, Customer H: due 2026-10-14, 1 day overdue

2. Customers waiting

URGENT: Monthly bookkeeping for a café, Customer A (received 2026-10-13, 2 days)

URGENT: Payroll question, Customer B (received 2026-10-14, 1 day)

Year-end prep, sole trader, Customer C (received today, not urgent yet)

3. Quotes to follow up

Bookkeeping + payroll, Customer E (sent 2026-09-30, 15 days)

Monthly bookkeeping, Customer G (sent 2026-10-08, 7 days)

4. Due today

Sign off year-end accounts, Customer I

[Bookkeeper 2] Payroll run, Customer M

5. Waiting on others

Finish VAT return, Customer K (waiting on receipts from the client)

6. On the radar

Year-end package, Customer F (quote sent 2026-10-10, 5 days, not yet 7)

Prepare fee review, Customer J (due 2026-10-22)

[Bookkeeper 2] File monthly report, Customer O (due 2026-10-17)

7. Done since yesterday

[Bookkeeper 1] Send reminder for missing receipts, Customer N (done 2026-10-14)

Done earlier: 0 items

8. Date missing

Phone note: wants a call back about VAT, Customer D (no date, so I can't tell how long this has been waiting). When did this call come in?

Questions for you:

The proposals say "no response" rather than "sent, no reply". I treated them as the same. Please tell me if that's wrong.

No items were of type "Deadline", so the deadlines block is empty. Tell me if any of the tasks are meant to be tax or payroll deadlines.

Check: 15 items in · 15 items in the briefing

Look at Customer D: the phone note had no date. The AI didn't guess. It put the item under "Date missing" and asked. That's exactly what should happen. Add the date the call came in, and next morning it shows up under "Customers waiting".

6. How you'll know it's working

The two numbers in the last line match. Then no item got lost.
"Do first today" shows what you'd have tackled first anyway, if you'd had everything in your head.
A quote with no response shows up under "Quotes to follow up" every morning once it passes your day limit, for as long as it's on your list.
Nothing in the briefing that isn't on your lists.

Common problems and what to do:

What happens	What you write or do
The numbers in the check line don't match.	"The check doesn't add up. Find the item that's missing or listed twice, and redo the briefing."
The day counts are off.	Today's date is missing at the top of the morning instruction, or it's still yesterday's. Fix the date and send again.
The briefing has an item that isn't on your lists.	"Item X isn't on my lists. Take it out."
Lots of items end up under "Date missing".	Not the AI's fault. Add the date to your list: when the enquiry came in, when the quote went out, when the task was done.
The AI asks for your day limits every morning.	Your morning instruction still has the placeholders [C: YOUR NUMBER] or [Q: YOUR NUMBER]. Replace them once and save the instruction again.
The briefing is too long.	Only paste what's open plus what got done since yesterday. Or use follow-up a): "Squash On the radar into one line per customer."
The conversation stops halfway.	Write "continue" later in the same chat. On the free plan, that's often the usage limit.

7. Limits and data protection

The AI only knows what you paste in. It doesn't look into your inbox or your invoicing software. If a customer isn't on your list, they aren't in the briefing.
The briefing shows, you decide. The AI sends nothing to customers and changes nothing in your software.
What stays out of the chat:
- names, addresses, phone numbers and email addresses of customers or staff. Write "Customer A" and the role ("Office").
- health data, bank details, ID numbers, salaries
- contracts, passwords, logins
- anything covered by professional confidentiality (for example in accounting, medical practices, law firms)
Watch out when copying from software. File names, subject lines and spreadsheet columns often contain customer names. Replace them before you paste.
Tracking your team's tasks: whether and how you're allowed to, check with an expert or your data protection officer. This isn't legal advice.
If the briefing should later arrive on its own, a program will access your inbox or your lists. Check with an expert beforehand whether and how that's allowed with the provider you choose.

8. Next step

If you want to build your briefing further with Claude yourself: our AI Blueprint is a ready setup for Claude, explained step by step, even if you don't code.

Go to the AI Blueprint