Same seven you gave me after coffee, nothing added and nothing quietly dropped. I put them in the order I would do them, easiest first, so the big tile is always the thing to start on.
Drag them around if you disagree, it is your board. And if an hours number looks wrong, tap it and fix it.
Keegan/ Handled
The board
Easiest first. Drag to reorder.
What happens next
Start with whatever is sitting in the big tile. If you have not moved anything, that is the Whisper Flow one, and it is twenty minutes.
Then tell me you have done it and I will build the email digest against your real inbox. That is the one with a September deadline, so that is the one I want moving.
Handled.
"Do one session with Whisper Flow to share who you are as a person, essentially making a Michelle .md file to save time with processing."
You talk for about fifteen minutes and it types. What comes out is one file about you: how you work, who your clients are, what you refuse to do. You paste it into a project once and stop starting from scratch every single time.
It is the same idea as your voice doc, except that one is how you sound and this one is what you know. You need both, and the voice doc is already built and waiting.
Yours
Install it, hit record, talk. Send me the raw dump, typos and all.
Mine
Turn the dump into the file and pair it with your voice doc.
Start here: install Whisper Flow, open a new Claude chat, paste the prompt below, then talk through these five.
Who you are and how you got to running ImageUp
What you actually sell, in your own words, not the website version
Who your clients are and what they are usually worried about
How a month works for you, batch by batch
What you will not do, no matter who asks
"Set email triage / one digest (slack included)."
One digest instead of an hour in the inbox. Four buckets, the ones we settled on: client is waiting on me, needs a decision from me, I am waiting on them, and money. That third one is the one that quietly costs you.
Promotional never reaches it. Newsletters get pulled out before triage sees them and come back separately as a few lines on the week, with the two or three worth opening in full flagged. That is a second pass, not triage, which is why one prompt could not do both.
Slack is stage two of this same card, a channel per client, once email is running clean. Same shape, and "I am waiting on them" pays for itself even faster there.
Yours
Filters are done, you did those. Next is letting it pull senders itself instead of you pasting them.
Mine
The digest, then the drafting pass with a hard approve gate, then Slack.
The cruise, honestly. Nothing sends on its own, not even while you are away. The realistic version is triage plus queued drafts plus an auto-reply plus a ten minute daily pass. True zero touch means somebody else holds send authority, and that is a different conversation. Calibration is the slow part, not the build, so we want this running in August so early September is just watching it.
Start here: run the prompt below on today's actual inbox instead of a pasted sample. You have done enough reps to trust it.
"I want to automate my FB ad analytics to where it lands in my inbox every week with ad spend. I want to be as hands off as possible. This is both internal and client facing."
A weekly email, per client, showing spend against budget and whether you are on pace, landing without you touching anything. Both versions you asked for: one you read, one you can forward.
The number it leads with is pacing, not clicks. When your subscription covers their ad spend, underspending leaves their budget on the table and overspending eats your margin. Everything else is supporting detail.
Most of this is already built. I made a start on it after you mentioned it, so there is something to look at rather than a promise.
Yours
Get me partner level access, and confirm whether your monthly figure per client is ad spend or the whole subscription.
Mine
The report, the email, the schedule. Already mostly standing.
Two honest notes. One, it has never touched a real ad account yet, so I am not going to tell you it works until it has. Two, the access is the long pole and it is genuinely yours: partner level Business Manager access to each client's ad account, roughly six steps in Business Settings. Being added as a person on their account is not the same thing and will not work.
Why this is third and not second. You listed it second and I moved it down one. Not because it matters less. It is the only card on this board gated behind something you have to go ask other people for, and I did not want the whole list stuck behind a permissions screen.
"Content strategy." One of the two worst hours you named back in July.
Shot lists, the content calendar, and captions. The three pieces you named. This is the card that most needs the voice doc, because a caption in the wrong voice is worse than no caption.
Do it on one real client first, not a hypothetical one. And run your Rella prompt against this one on the same brief so you can see the difference instead of taking my word for it.
Yours
Pick the client. Send me both outputs side by side and tell me which one you would actually post.
Mine
Tune it off your corrections until the captions stop needing a rewrite.
Start here: paste your voice doc, then the prompt below, then one real client's month.
"Proposal writing." Your other worst hour.
This one needs a second document before it can work: what you sell, what it costs, and what is in scope. It is already on your build ladder from July.
Without it you get one of two bad outcomes. Either it invents a price, which is the worst possible thing for it to invent, or it tags every other line "confirm with Michelle" until reading the draft takes longer than writing one.
Yours
A rough services and pricing doc. Bullet points are fine, it does not need to be pretty.
Mine
Shape it into something the drafts can rely on, then wire it up.
Start here: write the pricing doc badly and fast. Fifteen minutes, no formatting. That unlocks this whole card.
"Notion replacement."
Before anybody builds anything: what is Notion actually doing for you? Half of what people keep in Notion turns out to be three documents and a list, and the other half is genuinely load bearing.
I would rather leave you with three files that work than a careful rebuild of a tool you were already fighting.
Yours
Open Notion and tell me what you actually touch in a normal week.
Mine
Tell you honestly which parts do not need a tool at all.
"Replace Dubsado."
Dubsado is doing five separate jobs: forms, contracts, invoices, scheduling, and client workflows. "Replace Dubsado" is really five projects wearing one name, which is exactly why it sits at the bottom.
Some of those we can do better. Invoicing probably is not one of them. I will tell you which is which instead of pretending the whole thing is easy, and there is no rush on it at all.
Yours
Which of the five do you actually use, and how often.
Mine
An honest read on each one, including where Dubsado wins.
I'm going to talk for about fifteen minutes about myself and my business. Don't reply until I say I'm done.
When I'm done, turn everything I said into one reference document about me, written in my words. Do not summarize it into corporate language and do not tidy up my phrasing.
Organize it as: who I am, what ImageUp does, who my clients are, how I work month to month, what I care about, what I refuse to do.
At the bottom, list anything you'd want more detail on before you could use this to write as me.
Read today's inbox and sort it into these four buckets, nothing else:
1. Client is waiting on me
2. Needs a decision from me
3. I'm waiting on them
4. Money (invoices out, proposals out, anything unpaid)
Rules:
- Work off sender and subject line. Don't guess at what's inside.
- Anything promotional or newsletter shaped, leave out entirely. I handle those separately.
- If something doesn't fit a bucket, put it under "not sure" rather than forcing it.
- Give me the shortest version that still tells me what to do.
- No em dashes, no emojis.
I want my Facebook ad analytics to land in my inbox every week without me touching it. Ad spend against budget for each client, in two versions: one for me internally, one I can forward to the client.
Before you build or suggest anything, tell me:
1. Exactly what access you need on each client's ad account, and what you cannot do without it.
2. Which steps are mine and which are yours.
Then give me the plan in order, shortest path first. Lead with pacing against budget rather than clicks, because my subscription covers their spend.
My voice doc is above. Use it for every word you write, including the words you write to me.
Client: [name]
Month: [month]
What they have going on: [your notes, messy is fine]
Give me three things:
1. A content calendar for the month.
2. A shot list I could hand to a photographer without explaining it.
3. Captions I would actually post.
If you need something from me to make one of those good, ask before you write it rather than guessing. No em dashes, no emojis, no hustle language.
Above: my voice doc, and my services and pricing doc.
New prospect: [name]
What they asked for: [paste their email or your notes]
What I think they actually need: [your read]
Write the proposal in my voice.
Hard rules:
- Use only prices and scope from my pricing doc. If they asked for something that is not in there, say so plainly instead of inventing a number.
- End with an invitation, not a push. "Let's talk" over "act now".
- No em dashes, no emojis, no "leverage", no "scale".
I want to get off Notion, but don't suggest anything yet.
First ask me what I actually use it for, one question at a time.
Then tell me honestly: which of those things need a tool at all, which could just be a single document, and which I would genuinely miss if it disappeared.
I would rather end up with three files that work than a rebuild of Notion.
Dubsado does forms, contracts, invoices, scheduling and client workflows.
Ask me which of those I actually use and how often, one question at a time. Then tell me what replacing each one really takes, in plain terms.
Be straight with me about the ones where Dubsado is genuinely the better answer. I don't want to rebuild something worse to save a subscription.