May 12
Same engine, five readers — the unifying bridge page
Capital Refinery's structural claim is that the operator's self-assessment, the banker's pitch packet, the buyer's IC memo, the lender's covenant forecast, and the LP's verification page are all rendered by the same engine grading against the same 10-axis framework. The new /same-engine page makes the architecture explicit — five readers, five evidence levels, one underlying record. The diagram and the ladder argument that the price reflects evidence cost, not feature differences, ship as a dedicated surface.
May 12
Verified Financial Artifacts — category creation
The /p/<token> public-verification primitive ships today on two artifact types — the Continuous IC Memo and the IRA — each carrying a deterministic fingerprint and a public verification URL the receiving counterparty resolves independently, with no SaaS dependency, no login, and no opportunity for post-issuance editing. The architecture extends to further classes (stress run, covenant attestation, board export, QoE and lender artifacts), which are in development. The category page (/verified-artifacts) names the primitive cluster and the structural argument behind portability.
May 12
Vertical layer — 4 industry pages + hub
The IRA grades 10 sub-lanes with industry-parameterized thresholds. Until now, none of those sub-lanes had a public marketing surface. Four vertical pages shipped: HVAC services (roll-up acquirer framework), dental multi-location (DSO acquirer benchmarks including recall completion), distribution (Cedarbrook canonical proof case), and multi-family real estate (refi wall + lender engagement framing). Each uses engine-prediction language rather than fabricated observed-frequency claims. Each cites real industry sources — ACCA, ADA HPI, NAW, MBA, Trepp, NMHC.
May 12
Audience hubs — bankers, lenders, boards
Three new audience hubs joined the existing /for-advisors and /for-lps. /for-bankers covers sell-side mandate cycle integration (pre-CIM diligence compression, post-LOI fall-through reduction). /for-lenders covers credit officers underwriting facilities, distinct from /solutions/private-credit which serves portfolio funds. /for-boards covers fiduciary infrastructure for portfolio governance, with explicit Caremark/Marchand/Boeing legal framing and a 'not legal advice' disclaimer.
May 12
Anatomy of a decision — system transparency, anchored on Falcon
Most platforms describe what they do. The new /anatomy-of-a-decision page shows what one decision looks like as it moves through the Capital Refinery engine — from T+0 IC approval through T+540 LP review. Every primitive visible at each stage: Investment Decision Ledger, KPI Lineage, Continuous IC Memo, Risk Signals, Scenario Propagation, Decision Fragility, Decision Lock Outcome, /p/<token> public verification. Anchored on Falcon Services Q1 2026 (the canonical buy-side proof case) so the artifacts are reproducible from the running platform.
May 12
Risk Signals — feature deep-dive surface
The /risk-signals page documents the consolidated firm-policy scoreboard rendering on every IC memo, every export, every portfolio rollup. Six signal categories with observed/policy/verdict-typed entries: portfolio LTV, customer concentration, T-3 expense creep, debt-service runway, lease at-risk rent, title exceptions. Change a firm policy threshold once in firm_assumptions.py (or via tenant override); every existing memo re-grades on the next render. Includes the before/after threshold-change worked example.
May 12
Case Studies hub — two-section gallery
The /case-studies hub formally distinguishes named cases (Cedarbrook + Falcon, methodology fixtures disclosed as fixtures rather than customers) from pattern examples (the 4 industry verticals, explicitly labeled illustrative). Refuses a customer-logo wall on principle — the artifact's portability claim depends on standing alone, not on customer logos to ratify it. Two-section layout is structural, not editorial: a sophisticated reviewer should be able to tell at a glance which artifacts are real customer engagements (none currently) vs methodology fixtures vs pattern examples.
May 12
Trust Center + Glossary
/trust ships as a routing hub (not a compliance marketing page) covering seven primitives: data handling, tenant isolation, evidence retention, model boundaries, public verification URLs, security posture, refusal discipline. Each links to the detailed surface where the primitive is documented. /glossary defines 15 terms in brutally practical language — locked primitive names where they exist, transparent cross-references where buyers may encounter common alternate names (e.g., 'Decision Record' → 'Decision Ledger').
May 12
/learn library doubles — 3 → 8 pieces
Five new framework pieces shipped alongside the three existing seller/borrower/PE-evaluation pieces. Flagship: 'How a buyer's IC actually evaluates an acquisition target' — the 7-question framework, what gates vs negotiates, the patterns that quietly reprice every initial bid. Four supporting pieces: how an LP's IC evaluates a fund (vs how they should), evidence grades (management-prepared → audited financials), decision quality vs decision outcome (Kahneman, Mauboussin, Duke on resulting), the anatomy of a covenant breach (T-12 → T+180), and why deal memo regret is structural not personal.
May 12
Comparison sweep — 30 vendor pages updated
Every comparison page on the site (Affinity, iLEVEL, DealCloud, PitchBook, Allvue, Capital IQ, Chronograph, Hebbia, Datasite, Copilot, Carta, FactSet, eFront, Diligent, Juniper Square, Standard Metrics, Intralinks, Workiva, Grata, Tegus, Altvia, Axial, CB Insights, Cobalt, DataSnipper, FundCount, Macabacus, PrivCo, Tenor, Timvero) now carries an institutional-readiness measurement section identifying the dimension the named vendor doesn't operate in. Carta gets partner-framed (complementary trust primitives) per the locked doctrine; all others get the standard 'what they don't do' framing. None of the existing comparison content was removed.
May 12
Citation QA gate — locked after a real catch
Working through Sprint 5 surfaced a real citation failure: /same-engine cited SEC Private Fund Adviser Rules as if active, when those rules were vacated by the 5th Circuit in 2024. Fixed across /same-engine and /anatomy-of-a-decision with explicit vacatur acknowledgment. Also softened unsupported numerical claims (multiple-compression turn counts, MBA $1T headline, Heartland office count) and strengthened the Tides distress citation with the Bisnow-reported specific (~47 loans / ~$1.5B due by end of 2025). Citation QA gate now locked: currency check, numerical specificity check, named platform check, trade press citation check, vacated rules/superseded frameworks check.
May 12
Site-wide nav + sitemap — discoverability fix
All 16 new pages added to the site header dropdowns and the site footer. New Industries dropdown joins Platform, Solutions, Comparisons, Research. Platform dropdown expanded to two columns ('Overview' + 'Architecture & primitives'). Solutions dropdown expanded to four columns including a new 'By audience' column. Research dropdown gained a 'Learn' column with all 8 framework pieces. Sitemap.ts switched from hand-maintained array to filesystem-driven generation — every new page.tsx auto-indexes at default priority, with editorial overrides for the high-priority routes.
May 11
$750 Self-Assessment — operator-attested entry to the IRA ladder
The lowest rung of the operator ladder shipped: a $750 operator-attested intake graded against the 10-axis framework. Produces a 4-page fingerprinted memo with a Self-Assessed badge — explicitly NOT verified, cannot promote past intake without document evidence. The fee credit-forwards 90 days into the $4,500 Readiness Gap Review or the formal IRA, so operators who screen their own readiness do not pay twice for the same review surface. The ladder doctrine: price reflects evidence cost, not feature differences.
May 9
The lifecycle frame, end-to-end
Capital Refinery now serves three audiences from one infrastructure: deal teams underwriting with evidence, portfolio operators tracking which companies are getting more institutionally durable, and owners/advisors preparing for sale, financing, or board review. The homepage rewrite, the Decide / Improve / Verify spine, and the new audience-specific entry points — Portfolio Readiness Monitoring, Modernization Impact Review, Readiness Gap Review, and the new “Measure AI Impact” door for owners considering AI work — make the same engine speak in three voices without forking the methodology.
May 9
Lift Ledger — observed movement, not consultant ROI
Re-IRA outputs now group what moved between two evidence-backed snapshots into four layers: institutional readiness, diligence friction, operating process, and financial movement. The Ledger reads in observation language and refuses attribution claims by design — “EBITDA moved from $8.0M to $12.5M” is provable; “our work caused 18% growth” is not. When canonical KPIs aren’t comparable across both snapshots, the field is marked “not observable” rather than imputed. The discipline is the credibility.
May 9
Portfolio Readiness Monitoring (Tier 1) shipped
PE operating partners now get a single rollup view of every portfolio company with an active assessment — composite verdict, gating-axis status, weakest area, blocker count, last review date, and movement since the prior review. Sorted worst-first so attention-needed deals lead. Open any row to drill into the readiness detail; loop back to the rollup with one click.
May 9
Modernization Impact Review for partners + business owners
AI consultancies, automation shops, fractional CFOs, and operating partners can now attach an independent measurement layer to their engagements: $3,500 baseline before the work, $2,500 Lift Ledger after the work lands. A partner-as-agent legal frame protects both the partner and the client, and a doctrine-locked refusal of attribution claims means the artifact reads as evidence, not theater. Business owners can also commission the same review directly when no partner is involved yet.
May 9
Readiness Gap Review for owners preparing for sale or financing
$4,500, five business days, top-five blockers a sophisticated buyer or lender would surface first — credited toward a full Institutional Readiness Assessment within 60 days. Anchored against Quality of Earnings pricing rather than free SMB scorecards: same diligence-grade discipline, scoped to the deal size.
May 9
Verification language: Institutional Readiness Verified
When all five evidence-confirmed conditions hold, an assessment carries the “Institutional Readiness Verified” badge with co-rendered confirmation and expiration timestamps. Status without timestamps is decoration; status with a 12-month expiry and a deterministic re-grading path is a real verdict.