Turn Whiteboard Photos into a Real Document

Turn Whiteboard Photos Into a Real Document

Photograph the flip-charts and whiteboards from your retreat and hand them over, and they come back as a structured document covering themes, decisions, owners, and the parking lot, while people still remember what the arrows meant.

Nothing to flag, if the boards are your own team's working notes. The assistant checks this with you first
What it will ask you

One line per photo about what that board was for, who else was in the room and can check the guesses, and where the finished document needs to go. You don't write the labels into a prompt or work out any structure: you answer as you attach, in the words you'd use out loud.

Copy the whole block below and paste it into your AI chat. Nothing to fill in. Have the board photos to hand, but wait to attach them: the assistant confirms it can actually read photos in this chat, asks what each board was for as you go, and then walks you and one other person who was in the room through checking every guess it made about an arrow.

The whole recipe, in one paste
You are my research assistant. I work at a nonprofit. I am not
technical. Your job is to run the following FOR me, asking me
only questions a non-technical person can answer, and keeping a
hard line between what the boards actually say and what you
worked out from them.

WHAT WE'RE DOING
Many AI chats can't actually see photos you attach, so we check
that first, before anything else. Turning photos of the
whiteboards and flip-charts from our working session into one
document the people who weren't in the room can use: a
transcription of each board, then themes, decisions, action
items with owners, and the parking lot. Every word you couldn't
read and every arrow you interpreted stays labeled, so the two
of us can check those and nothing else. What this takes: about
an hour if this chat can read photos; if not, your phone's
photo-text tool first.

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. Send that as its own message.
- Then, in a separate short message, check whether you can
  actually see and read photos I attach in this chat, and tell
  me plainly either way before laying out what happens next. If
  you can't, offer the fallback: I run my phone's photo-text
  tool on each board first and paste the text instead. If you
  can, tell me how many photos to send in one message and to
  send the rest after.
- Never ask me a technical question directly. Ask the everyday
  version and work out the technical answer yourself. For
  example: do NOT ask "how should the boards be keyed and
  segmented?" Instead, take the photos one at a time and ask
  "what was this board for?", accept a one-line answer like
  "the fundraising breakout", and use that as its label.
- Don't assume what software I use. Ask me where this document
  needs to end up and who has to open it, and hand it to me in
  a form that pastes cleanly into 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.

QUESTIONS YOU'LL NEED ANSWERED (in your own words, one at a time)
1. What the session was and who this document is for. Ask me
   whether anything on these boards is about a named person, a
   client, or a staffing decision. If it is, tell me to leave
   that board out or crop it, and to keep it out of anything
   that gets shared widely. Say plainly that this check comes
   from the recipe instructions I pasted, not something you're
   assuming about me.
2. What each board was for, one line each, taking the photos
   in turn as I attach them.
3. Who else was in the room and can look at this with me,
   since your reading of the arrows needs a second person who
   was actually there.
4. Where the document has to go and by when. Tell me that
   sending it the same day matters, because a week later
   people find the arrows just as hard to read as you did.

HOW TO DO THE WORK (this part is for you, not me)
Transcribe every board separately, under my own label for it,
and keep them separate all the way through: they were different
conversations and merging them loses the thread. Preserve lists
and groupings exactly as they sit on the board. Where an arrow
or a circle connects two things, describe the connection in
words rather than reporting that a line exists. Then, across
all the boards together, organize everything into four
sections: themes discussed; decisions made; action items, with
an owner wherever one is written down; and open questions, the
parking lot. Mark anything you cannot read as [UNREADABLE], and
anything you worked out rather than read, including every arrow
and circle, as [INFERRED]. Give me the per-board transcription
first, then the four sections, with every tag left in place.
Never resolve a tag quietly to make the document read better.

AFTER THE DOCUMENT, WALK ME THROUGH
1. Every [UNREADABLE] and [INFERRED] tag, one at a time, so I
   can confirm, correct, or delete each one. Ask me about them
   rather than waiting for me to notice them.
2. Sitting down with the other person who was in the room and
   reading the four sections against the photos. If they
   correct more than a couple of the [INFERRED] items, tell me
   to go back through every arrow together before we send it.
3. Getting it into the document I named, sending it today, and
   filing the photos somewhere the team can find them, because
   the photos are the original and this is the write-up.

RULES
- Never merge two boards, and never smooth over a board that
  contradicts another one. Two boards disagreeing is a finding,
  not a mistake to fix.
- Don't invent an owner for an action item. If no name is
  written next to it, leave the owner blank and list it for me
  as something to assign.
- Before we finish, remind me of the habit the whole thing
  depends on: photograph every board before anyone erases it.
  Nobody can reconstruct a half-erased board, you included.
- 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
do together, then ask your first question.
The transcription is reliable, but the interpretation of arrows and circles is a labeled guess. The assistant is instructed to tag every one and to walk you through them individually rather than leaving you to spot them. Confirming them is still a job for two people who were in the room, and the prompt asks who your second person is up front for exactly that reason.
Photograph before erasing. That habit is the one thing the assistant cannot do for you, and it is told to remind you of it before you close the chat. A half-erased board is gone.
What you walk away with

A structured retreat document covering themes, decisions, owners, and open questions, checked by someone who was in the room and sent out while people still remember what the arrows meant.

Previous
Previous

Who's Missing? (The Three-List Reconcile)

Next
Next

Where Do Our Volunteers Work?