# Founder walkthrough: ready to record

## Recording brief

Build n Bloom · version 1.0 · 26 September 2026. **Script and storyboard, not a recorded video.** Suggested presenter: Richard for the buying decision, with Rayan reviewing the technical boundary before recording. This does not assert either founder has approved or recorded it. Target runtime about six minutes; rehearse to confirm. Open the self-guided walkthrough and exact versioned records in advance. Use a real founder voice and screen; no generated likeness or invented testimonial.

**Opening on-screen label:** 'Fictional advisory example. Authored records and decisions. No live model or client system.' Keep that label visible when showing the case. Mark the separate public-source review clearly. Show original passages and the contrary case; do not edit away the reviewer return. Do not present local calculations as observed savings.

**Files to open:** sample-blueprint.html; working-records.html; walkthrough.html; synthetic-effort-ledger.csv; buyer-approval-pack.html. Before publishing, confirm all links, current offer wording, actual presenter permission and the rendered figures. Recording remains outstanding.

<!-- PAGE -->
## Script and storyboard: first half

**00:00-00:40 · Presenter, then case map**

'Every new engagement can add more preparation, review and correction to the same people. Here is how we examine that constraint. This is a completed fictional example of a Firm Capacity Blueprint. We have not installed a system for this firm, and these figures are teaching assumptions. The point is to make the design and the decision inspectable.'

Screen: Blueprint §01 then §02. Point to the whole revenue path and three designs. Do not introduce an agent count as the product.

**00:40-01:30 · Brief and source register**

'Northbank sells an operating-model diagnostic. The delivery brief tells us the scope, source permissions, method and professional owner. Before looking at the polished report, open its evidence. These two records identify responsibilities and a handoff delay. They say nothing about how many leaders the client needs.'

Screen: working records B-01, S-01, S-02. Read the exact missing-receipt sentence. Explain that this example contains no real client material.

**01:30-02:30 · Challenged conclusion and return**

'The draft says the client can safely remove two leadership roles. It sounds definite. Its cited source does not support it. The reviewer returns that inference because it is unsupported and staffing advice is outside the scope. The supported handoff observation stays. Any staffing request goes to the partner for a separate scope decision. The revised finding describes the unresolved handoff and makes no staffing conclusion. This decision was authored to show the method. It is not evidence that a deployed model passed this test.'

Screen: X-01 original → disposition → revised result. Leave the sources visible. No animation implying live model execution.

<!-- PAGE -->
## Script and storyboard: second half

**02:30-03:20 · Proposed working capability**

'The proposed capability carries evidence gathering, preparation, checks and coordination to a reviewable result. The professional keeps the consequential decision. Source changes, missing permission and reviewer absence must have explicit stop and return paths. Those controls need actual execution tests and receipts in the client's approved environment.'

Screen: Blueprint W2 failure/tests, shared boundary. Highlight proposed/not run. Rayan should check that the recording does not imply a verified provider or security certification.

**03:20-03:55 · Simpler close-out**

'We do not recommend an AI agent for every opportunity. Here the close-out job is mostly structured: exact approved version, acceptance, milestone and finance owner. An existing approval tool and checklist can do it. Finance controls invoices and collection. Reuse has a separate rights decision.'

Screen: W3 no-custom-AI recommendation and checklist C-01.

**03:55-05:10 · Economics and contrary case**

'We count the complete accepted case, by role. Only the proposed report/review change is in this model. It releases senior and analyst capacity but adds operations effort, setup and oversight. Under the fictional demand and headroom assumptions, six added cases fit. With the illustrative fees and incremental costs, that is 54,000 dollars before unknown one-time investment. It is not a quote or a forecast. This compares the proposed package with the unrepaired baseline; incremental AI value against the repaired method still needs measurement. If senior review remains at ten and a half hours, the cap drops to two cases and 14,000 before investment. With no demand, the added-case benefit disappears.'

Screen: calculator at nine then 10.5 senior hours; demand six then zero. Restore central assumptions after recording. Do not say 54,000 is a recommended price or a client saving.

**05:10-06:00 · Shareable decision and next step**

'The Blueprint gives you three complete designs and a reasoned first decision. You can use it with another competent implementer. Installation has its own scope, price and acceptance. Share the approval pack with the people who need to assess economics, delivery and data. If you recognise a comparable constraint, start with a free 25-minute Firm Capacity Fit Call. You do not need confidential files for that conversation.'

Screen: approval pack, then actual booking link. Optional cutaway: the separate real public-source review, explicitly labelled public-source work, to show our source method.

**Before release:** actual founder records and approves the final video; source/version labels remain legible; no confidential tab/notification appears; captions match the spoken figures; links resolve; use-specific claims remain scoped. No video exists as part of this script deliverable.