GreenSquare AI
Get Frame Free PDF

Frame the decision before AI answers it.

Use Frame Free with the AI you already use to question the decision, separate evidence from assumption and identify the next gap to check.

Frame Free clarifies and tests a decision. It stops before ranking options or recommending a course of action.

Free six-page PDF. Use it in ChatGPT, Claude or Gemini.

Enter your email for Frame Free and occasional Frame product emails, including Frame Pro launch updates. Confirm your address in the email we send to activate updates. Unsubscribe at any time. Privacy policy.

See the Free example

A vendor choice, framed before it is evaluated.

This is the worked example from the Frame Free guide: an operations manager told to choose between two rostering platforms before the current contract expires. The film runs the guide’s six questions in order. Frame Free stops before evaluation: neither platform is scored, compared or chosen.

Worked example from the Frame Free guide, page 5
The worked example from the Frame Free guide, as a muted film; a still frame is shown when motion is reduced or scripts are off. An operations manager is told to choose between two rostering platforms, and the trigger is a renewal notice. The guide’s six questions run in order: the real question turns out to be what the shift rules should be, a third option appears, the load-bearing assumption is named, the deadline and the budget cap are shown to be untested, and each material claim is tagged by its evidence class. The film ends on the reframed decision and the single evidence gap to test next, and no option is scored or chosen. The complete text follows.
Read the worked example as text

The worked example from the Frame Free guide

The software was the surface

An operations manager at a 60-person home-services business is told to choose between two rostering platforms before the current contract expires.

Start with: Use Frame Free in this PDF. Ask me one question at a time.

Find the question behind the question

The first three moves make the decision explicit, expose the deeper issue and reopen the option set.

  1. 1.
    The decision and trigger. What exactly are you deciding, and what made it live now?
  2. 2.
    The decision underneath. If the immediate choice disappeared, what larger question would remain?
  3. 3.
    The missing option. Which genuine third path have you not treated as real, including doing nothing?Do not evaluate it yet.

Separate evidence from assumption

The final three moves identify the belief carrying the decision, test what has been treated as fixed and make the evidence state visible.

  1. 4.
    The load-bearing assumption. Which single belief would change the decision if it proved wrong?
  2. 5.
    Untested constraints. Which deadlines, budgets or stakeholder positions are verified, and which are assumed?
  3. 6.
    The evidence map. What is given, derived, inferred or unknown? Keep approximations visible.

Framed is not decided

1. Presented decision
Choose vendor A or vendor B by the end of the month. The trigger is a renewal notice.
2. Decision underneath
Rosters are rebuilt by hand every week because the shift rules keep changing. The real question is what those rules should be.
3. Missing third option
Extend the current contract month to month, fix the roster rules first, then choose software against stable requirements.
4. Load-bearing assumption
‘The pain is the software.’ If the pain is the rules, even the best platform changes nothing.
5. Untested constraints
The deadline is treated as fixed, but the vendor offers month-to-month extensions.
The budget cap was set two years ago and has not been retested.
6. Evidence map
Given: current licence cost; rosters are rebuilt weekly.
Derived: about 15 hours a week are lost to manual fixes.
Inferred: some turnover traces to roster chaos.
Unknown: whether award compliance is exposed.
7. Reframed decision and the single evidence gap to verify next
Reframed decision. Verify the extension, stabilise the shift rules, then choose tooling against those rules.
Evidence gap to test next. Confirm whether an award-compliance exposure already exists. If it does, the timing and scope of the next step change.

Stop here. Framing prepares a decision; it does not make the decision.

No options were scored. Nothing was recommended. The question became the right one - which is the point.

What the guide’s worked example shows

The vendor choice
Choose vendor A or vendor B by the end of the month. The trigger is a renewal notice.
What it concealed
Rosters are rebuilt by hand every week because the shift rules keep changing. The real question is what those rules should be.
Why switching may not help
‘The pain is the software.’ If the pain is the rules, even the best platform changes nothing.

The question you bring, and the one you leave with.

Frame Free does not answer the question as presented. It returns a framing handoff in seven parts, so you can see what was assumed, what was verified and what to check next.

BeforeThe question as presented

Choose vendor A or vendor B by the end of the month. The trigger is a renewal notice.

AfterReframed, with the gap to verify

Verify the extension, stabilise the shift rules, then choose tooling against those rules.

Confirm whether an award-compliance exposure already exists. If it does, the timing and scope of the next step change.

Framing handoffThe guide’s worked example, page 5

An operations manager at a 60-person home-services business is told to choose between two rostering platforms before the current contract expires.

  1. 1

    Presented decision

    Choose vendor A or vendor B by the end of the month. The trigger is a renewal notice.

  2. 2

    Decision underneath

    Rosters are rebuilt by hand every week because the shift rules keep changing. The real question is what those rules should be.

  3. 3

    Missing third option

    Extend the current contract month to month, fix the roster rules first, then choose software against stable requirements.

  4. 4

    Load-bearing assumption

    ‘The pain is the software.’ If the pain is the rules, even the best platform changes nothing.

  5. 5

    Untested constraints

    The deadline is treated as fixed, but the vendor offers month-to-month extensions.

    The budget cap was set two years ago and has not been retested.

  6. 6

    Evidence map

    Given: current licence cost; rosters are rebuilt weekly.

    Derived: about 15 hours a week are lost to manual fixes.

    Inferred: some turnover traces to roster chaos.

    Unknown: whether award compliance is exposed.

  7. 7

    Reframed decision and the single evidence gap to verify next

    Reframed decision. Verify the extension, stabilise the shift rules, then choose tooling against those rules.

    Evidence gap to test next. Confirm whether an award-compliance exposure already exists. If it does, the timing and scope of the next step change.

Stop here. Framing prepares a decision; it does not make the decision. Fictional business from the guide; not a customer result.

Five decisions it is built to question.

Each links to an illustrative example: the starting decision, what was questioned, the framing handoff and what remains unknown. The PDF is the same whichever you read.

Three steps to the first question.

Under five minutes from download to the first question. Then about ten minutes for one decision.

Read the start guide
  1. 01

    Get the PDF

    Enter your email. The PDF is ready to download on the same page, with no inbox wait.

  2. 02

    Attach it to a new chat

    Open a new chat in ChatGPT, Claude or Gemini, in an account you are permitted to use for the decision, and attach the PDF.

  3. 03

    Paste the starter and describe one decision

    The model asks six questions, one at a time, then returns a seven-part framing handoff and stops.

What has and has not been shown.

Shown here
The guide’s own worked example, and illustrative examples on fictional inputs. They show how the framing behaves, not customer results.
Not shown
Frame Free has not been benchmarked. An earlier study tested a method file that predates Frame: it changed how the model worked, and it did not beat an informed baseline on judgement.
Not offered
Ranking, scoring or a recommendation. Frame Pro, which continues into those, is in development.

Read the research overview