×ChurchOps.AIPrepared for focus412
×ChurchOps.AIIn response to the handoff and the assessment archive you sent on 28 July. Everything here was built from your own files. Each page works on a phone.
Your whole archive read at respondent level in a single pass: 528 scored responses, 1,663 written answers, 25 engagements. There is a finding about Culture in here that we think validates the four D’s from your own data rather than from experience. It also flags five things the archive cannot do yet, each with a small fix.
Start hereAll nineteen constraints triaged into phases, filterable. Our leading hypothesis for why the dashboard reads 0.00, with the single query that would confirm or eliminate it. What we would keep, what we would rewrite, and the two places we would move slower than your document suggests. The free build is spelled out with a definition of done.
Read the planWe read all twenty-five of your reports and wrote down how you write one: the structure, the reveal mechanic, the fifteen rules your voice actually follows, and the recommendation-plus-outcome move from one of the older decks. This is the document that makes generated writing sound like focus412 rather than like everything else.
See the methodEvery figure was recomputed from your respondent-level rows and checked a second time in a different tool. Where a claim rests on a judgement call rather than on the data, the page says so plainly. Where we took something out on confidentiality grounds, it says that too.
Nothing here is estimated, and nothing is padded. If a number looks wrong to you, it is worth a conversation, because we would rather correct it now than have it end up in front of a church.