A funder-ready narrative pairs every figure with a verbatim quote. Here is the prompt that grades each claim green, amber, or red so nothing unsupported ships.
In short: A funder narrative is only defensible when every claim carries both a number and a verbatim quote from the same people. Run one prompt over your cleaned dataset that writes the narrative and grades each claim green, amber, or red — so a self-reported figure shows as amber and an unsupported claim shows as red before it ever reaches the funder.
Point the Sopact Sense Assistant at your dataset so it works from clean records with persistent contact IDs, then have it load your Decision Brief before it does anything. That brief — your decision, audience, outcomes, indicators, and evidence standard — is what lets the assistant judge whether a claim is actually supported.
You are the Sopact Sense Assistant working over the DEMO-03 · Workforce Cohort dataset (clean data + persistent contact IDs). Load my Decision Brief (decision, audience, outcomes, indicators, evidence standard) first, then wait for my task.
For the workforce cohort, write an evidence-led narrative of change across cycles; anchor every claim to a quant citation AND a verbatim quote; nothing ships red. Grade green/amber/red.
Five elements make this prompt work: it runs over your dataset; it is a decision-first narrative rather than a list of stats; every claim pairs a number + quote from the same person; the rule is nothing red ships; and it ends with a grade of green, amber, or red on each claim.
Run on the Workforce Cohort dataset (DEMO-03) already loaded in Sopact Sense.
GRADE: green | Skills claim | cited; amber | Confidence claim | self-report; red | Sustained-jobs claim | unsupported
The green skills claim is anchored to a measured indicator and a matching quote. The amber confidence claim rests on a self-reported rating with no triangulation — usable, but flagged. The red sustained-jobs claim has no indicator behind it at all, so it cannot ship as written.
Take the lowest-graded element above and fix it using only what the program could realistically measure. Show the before → after grade and the single indicator/edit that moves it to green.
Create a 'missing & incomplete' report from this analysis in Sopact branding [or paste your website URL / brand guideline to apply your own]. List every element graded amber or red, what is missing, and the one input that fixes each. Lead with the decision this report informs.
Create a shareable link for this report and open it in a new tab.
Anchor on IDs, not aggregates. A quote only supports a number when it comes from the same person the number describes. Ask the assistant to cite the contact ID alongside each claim so the pairing is auditable.
Treat self-report as amber, not green. Confidence and satisfaction ratings are real signal, but on their own they are self-reported. Keep them amber until a second source — a behaviour, an outcome, an observer — backs them.
Never let the narrative outrun the data. The most common red is a sustained-outcome claim with no follow-up measure. If you cannot cite it, soften it or drop it.
Make the grade do the editing. Once each claim is graded, the amber and red items are your revision list. Fix the lowest one first.
List every red claim in this narrative and, for each, name the single indicator I would need to collect to turn it green.
Write it claim by claim, pairing every figure with a verbatim quote from the same person, and grade each claim green, amber, or red against your evidence standard. Ship only what is green or defensibly amber; rewrite or cut anything red.
A cited claim has two anchors from the same contact ID: a quantitative indicator and a verbatim quote that says the same thing in the person's own words. One without the other is, at best, amber.
Mark it red and either add the indicator that would evidence it or rephrase it as a hypothesis rather than a result. The one-fix prompt will tell you exactly which single measure moves it to green.
Open Sopact Sense, paste your program description, and put it to work.
Try in Sopact
