Foundry phase 02 · static-html
AcquiScope Phase 1 prototype
Every screen the Phase 1 build has to produce, wired together so you can click the whole underwriting path end to end. The numbers are a worked fictional example carried consistently across all 20 screens. Nothing here fetches, stores, or computes anything.
Demo data. Company names, company numbers, financials and Companies House references on these screens are invented for the walkthrough. They are not real businesses and no real listing or filing was read to produce them.
Where a real product would show a fetched figure, the prototype shows the provenance tag the real figure would carry.
Start here
Step 1
Opportunity dashboard
All 20 columns, 7 live rows, filters, and the distressed toggle sitting off by default.
Step 2
Opportunity workspace
One business, worked from entity match through to the recommendation and the next action.
Step 3
Stage 2 investment report
The 50-section output, including the sections that render an information gap instead of prose.
All screens
Set-up and sourcing
| Screen | What it proves | Prompt section |
|---|---|---|
| Mandate | Every default editable per search, nothing hardcoded in an engine | 5.1 |
| Sources and authorisation | Adapter registry, authorisation basis gate, robots and rate-limit state, both public-web adapters off | 5.2 |
| Ingest | Manual entry, CSV import, Companies House lookup, rate budget in view | 5.2, 5.3 |
| Dashboard | 20 columns, core actions, provenance on every figure | 5.17 |
| Distressed band | Detection signals stored, hidden by default, never blended into the shortlist | 5.9 |
Underwriting the opportunity
| Screen | What it proves | Prompt section |
|---|---|---|
| Overview and entity match | Confidence 0-100 with the stored evidence behind it, not just the score | 5.3 |
| Three-year financials | Revenue, growth, CAGR, EBITDA series, margin, abridged-accounts handling | 5.4 |
| Quality of earnings | Adjustment register with verdicts, and the three-column EBITDA bridge | 5.5, 5.6 |
| Revenue quality | Recurring revenue, concentration bands, cash conversion, capital efficiency | 5.7 |
| Quality score | 11 weighted components, each with its inputs and its reasoning | 5.8 |
| Screening | Rejection rules, per-rule override, recorded override reason | 5.9 |
| Valuation | Multiple build-up line by line, EV to equity bridge, full price set | 5.10 |
| Funding and returns | Structure, DSCR, base and downside IRR, MOIC, equity payback | 5.11 |
| Stress tests | Five stresses, two of them breaking the DSCR covenant | 5.11 |
| Forecasts | Downside, base and upside over five years, every column labelled | 5.12 |
| Narrative and number guard | Fact pack in, prose out, ungrounded numbers rejected before render | 5.14 |
Decision and output
| Screen | What it proves | Prompt section |
|---|---|---|
| Approval gate | Server-side gate with an audit trail, refusing because no channel is wired | 5.16 |
| Stage 1 deck | 12 slides: mandate, funnel, shortlist, snapshots, ranking, actions | 7 |
| Stage 2 report | 50 sections, gaps rendered as questions rather than filler | 7 |
| PDF report | Print layout for buyer, accountant, lender and solicitor | 7 |
What this prototype does not do
- No FastAPI app, no database, no Alembic migrations. Nothing persists between clicks.
- No source adapter runs. The Companies House panel shows a rate budget it never spends.
- No engine computes. Every figure is a hand-worked constant, consistent across screens so the walkthrough holds together.
- Provenance is styled, not enforced. Type-level enforcement lands with the real domain layer.
- No deployment. dev-acquiscope.innovatexperts.co.uk comes in the refinement phase.
AcquiScope produces analysis and modelling, not regulated investment advice. Every figure requires independent verification and professional due diligence before any offer is made. This prototype additionally contains no real data of any kind.