ARPIE User Guide

1. Welcome to ARPIE

ARPIE (Automated Real Property Investment Engine) is a commercial real estate underwriting and portfolio intelligence platform built for investment professionals. It ingests deal data, runs a fully deterministic underwriting pipeline, scores every deal on a 0–100 scale, stress-tests key assumptions, and surfaces actionable insights — all without relying on AI black-box models.

What makes ARPIE different

• Deterministic underwriting. Every score, metric, and recommendation is produced by transparent, rules-based logic. You can trace exactly why a deal scored the way it did. There are no large-language-model guesses in the underwriting pipeline.
• Portfolio-level analysis. ARPIE does not just analyze deals in isolation. It ranks holdings against each other, highlights allocation concentrations, and generates investment-committee-ready comparison materials.
• Capital-stack mechanics, not just pricing. ARPIE's Max-Bid Solver tells you the maximum price the deal supports for a target investor IRR. The Capital Stack Modeler then shows how that money actually distributes across N capital classes and a multi-tier promote ladder — pari-passu math, class sequencing, promote leakage, all visible.
• Operating Agreement drafts. After underwriting, the OA Generator turns your firm's template into a draft .docx populated with deal-specific values, with source provenance on every field. Drafts start in shadow state and never bind your firm until your attorney confirms review.
• Speed. Upload a deal package, and ARPIE returns a full underwriting memo, scorecard, stress scenarios, and broker follow-up questions in seconds — work that typically takes an analyst several hours.

2. Getting Started

2.1 Accessing ARPIE

Open your browser and navigate to https://arpie.masegoinc.com. Supported browsers: Chrome, Firefox, Edge, and Safari (latest two major versions).

2.2 Signing in for the first time

If you received an invite link from your organization admin, click the link to set your password and activate your account. If you were given a username and temporary password, enter them on the login page.

Where you land after sign-in. New users with no deals yet land on the Bring in Deals page so they can get started. Returning users with at least one deal land on the Deals page directly — no extra clicks to reach your pipeline.

2.3 Setting up two-factor authentication (MFA)

1. Sign in to ARPIE.
2. Click your name in the sidebar, then select Account Settings.
3. Under Security, select Enable Two-Factor Authentication.
4. Choose your preferred method:
   • Authenticator app (TOTP): Scan the QR code with Google Authenticator, Microsoft Authenticator, or Authy. Enter the six-digit code to confirm.
   • Phone verification: Register a phone number for SMS codes.
5. Store the recovery codes in a safe place — they are the only way to regain access if you lose your device.

2.4 Changing your password

Navigate to Account Settings → Change Password. Enter your current password, then your new password twice. Requirements: minimum 8 characters, at least one uppercase, one lowercase, and one number.

2.5 Navigating the dashboard

Every nav item has an icon plus a plain-language label. Collapse the sidebar via the floating « handle on the sidebar / main edge to free up horizontal space; click it again to re-expand. Your collapsed/expanded preference is remembered across sessions.

3. Bringing Deals into ARPIE

3.1 The four ways to bring in deals

The Bring in deals page is your hub. The hero card at the top is Upload deal files — the fastest path when you have files in hand. Below it, four other options handle different sourcing patterns:

3.2 What data you need (for a single-deal upload)

At minimum you need basic property and financial information:

• Property details: Name, address, property type, square footage, unit count, occupancy rate.
• Financial data: Purchase price, loan amount, interest rate, loan term, NOI, effective gross income, operating expenses.
• Optional but recommended: Rent roll, T-12 operating statement, cap rate, year built.

ARPIE's underwriting readiness score tells you how complete your submission is and which missing fields would improve the analysis.

3.3 Supported file formats

3.4 How upload + classify works

ARPIE auto-classifies each file you drop into the correct slot — rent roll, T-12, deal text/OM, or PLA workbook — using filename signals and a content scan. Each filled slot shows a “Detected as X · confidence” badge so you can see exactly what ARPIE thought.

• High / medium confidence: the file is auto-routed to that slot. You can still override manually by clicking the slot and picking a different file.
• Low confidence or unknown: the file lands in a “Needs review” panel where you pick the target slot manually (or skip the file entirely).
• Slot conflict: if a drop would replace a file already assigned to that slot, ARPIE routes it to the review panel so you don't accidentally overwrite.

This is the fix to the old “all .xlsx files become rent roll” behavior. A T-12 dropped as Excel now lands in the T-12 slot, not silently in rent roll.

3.5 What happens after upload

You see a live progress strip with three stages:

1. Files received — upload progress percentage displays during the transfer.
2. Analysis running — ARPIE's pipeline ingests, normalizes, runs underwriting math, scores, and stress-tests.
3. Deal ready — a success card appears with a direct Open Deal CTA.

Behind the scenes the pipeline runs:

1. Ingestion — File parsed and fields extracted.
2. Normalization — Data mapped to canonical deal schema.
3. Underwriting — Metrics calculated (NOI, cap rate, DSCR, LTV, cash-on-cash, IRR, debt yield).
4. Scoring — Composite deal score (0–100) generated.
5. Stress testing — Three scenarios applied automatically.
6. Artifact generation — Underwriting memo, scorecard, follow-up questions, and scenario comparison produced.

3.6 Outlook / Gmail intake

When you connect Outlook or Gmail, ARPIE scans the selected folder for deal emails and packages them automatically. The intake screen shows a skipped-emails count so you know exactly which messages did not look like deals (and why). A deterministic broker/OM classifier runs in front of Outlook intake to keep noise out of the pipeline.

4. Working the Deals Page

4.1 Page layout

The Deals page is now optimized for fast triage. From top to bottom:

1. Hero metrics — pipeline-wide signals at a glance.
2. Status tabs — Active / Rejected / All.
3. Intent tabs — pre-built filter combinations that match how analysts actually triage (covered below).
4. More filters — collapsed by default; expand for sort, readiness, action, and attribute filters when you need them.
5. Deal table — the actual queue. Click any row to jump to the deal detail.
6. Pipeline summary — collapsed by default; holds the package-mix / readiness-mix / blocker-mix / next-action / investment-stance charts and the workflow cockpit cards. Open it when you need the analytics; otherwise it stays out of the way.

4.2 Intent tabs (pre-built filters)

4.3 Drilling into a deal

Click any row to open the deal detail page. The first thing you see is the Decision Desk at the top — a one-screen summary of where the deal stands.

5. Understanding Your Analysis

5.1 The Decision Desk

Every deal detail page starts with a Decision Desk — four panels that answer stance / why / what blocks / next step at a glance:

• Stance — current investment stance (Proceed / Proceed with Caution / Reprice / Decline / Hold) plus badges for next action and blocker count.
• Top reasons — up to 6 reasons grouped by tone: Strength (green), Watch (gold), Risk (red). Pulled from the analysis commentary, gate failures, and risk flags.
• Blockers — critical missing items with a prominent Upload missing files CTA that jumps you straight to the right uploads slot for that deal.
• Next action — what to do next in plain language.

Below the Decision Desk, the hero shows lifecycle / execution badges, quick links to Max-Bid and OA Generator (covered below), and the deal header. Then a Quantitative signals strip with four numeric scores: underwriting readiness %, portfolio score, underwriting score, portfolio fit.

5.2 Deal score breakdown

Every deal receives a composite score from 0 to 100, calculated from four categories:

5.3 Risk classification

5.4 Recommendation bands

5.5 Key metrics

• NOI — Net Operating Income: total property income minus operating expenses.
• Cap Rate — NOI / purchase price. Unlevered return on the asset.
• DSCR — NOI / annual debt service. Above 1.25x is healthy; below 1.20x raises concern.
• LTV — Loan / purchase price. Most lenders target 65–75%.
• Cash-on-Cash — Annual pre-tax cash flow / total equity invested.
• IRR — Projected annualized return accounting for time value of money.
• Debt Yield — NOI / loan amount. Lender-focused coverage metric.
• Expense Ratio — Operating expenses / effective gross income.

5.6 Stress scenarios

5.7 Insight pillars

5.8 Deal detail tabs

Below the hero, the deal page has tabs that group deeper analysis:

• Overview — underwriting snapshot, scorecard, readiness breakdown, leverage analysis, structural risk, and commentary.
• Artifacts — preserved underwriting files, previews, and export-ready deliverables.
• History — timeline of lifecycle changes, execution events, and decision context.
• Operating Agreement — OA drafts for this deal with state transitions and amendment chain (covered in section 9).
• Workflow — committee notes, ownership, approvals.

5.9 Comparable Sales & Leases

Every deal detail page pulls nearby sale comps from HERA, our county-level comp mirror. Because a county pull can return buildings that are very different from your subject — a 3,000 SF single-tenant suite and a 60,000 SF multi-tenant tower are both technically "office" — every comp is automatically scored 0 to 100 for similarity to the subject and binned into four tiers:

Each row's Why cell shows the top two reasons the scorer assigned that tier. Hover it to see the full rationale plus the numeric score.

Practical shortcut: if you underwrite off Primary + Secondary rows only, you're following standard institutional comp-selection practice. Excluded rows are there for transparency, not for math.

Manual comps you add via Add comp or Upload CSV are flagged "Manual" and are never overwritten by a HERA refresh — your broker comps stay untouched across re-runs.

6. Market Intelligence

ARPIE automatically enriches every deal with county-level market intelligence drawn from 14 federal data sources covering all 3,200+ US counties. When you upload a deal, ARPIE resolves the property location to its county and pulls a comprehensive market profile — no extra steps required.

6.1 Composite market scores

Each county is scored on five dimensions (0–100 scale with national percentile ranks):

6.2 Economic indicators

Beyond the composite scores, each county profile includes:

• Per capita income and GDP growth trends (5-year CAGR)
• Average weekly wages by sector
• Net migration — how many people (and how much adjusted gross income) are moving in or out annually (IRS SOI data)
• Building permits — new construction activity signaling demand and supply pipeline
• Fair market rents by bedroom count (HUD FMR data)
• Broadband coverage — percentage of households with broadband access

6.3 Risk profile

ARPIE displays the FEMA National Risk Index for each county, including:

• Overall risk score and rating (Very Low to Very High)
• Individual hazard breakdowns: earthquake, flood, hurricane, tornado, wildfire
• Expected annual loss score and social vulnerability

This helps you assess natural disaster exposure — critical for insurance costs, business continuity planning, and long-term asset value.

6.4 How it feeds the analysis

The Market Support insight pillar (20% of the overall insight score) uses county intelligence data to evaluate market-level support for the deal. A property in a county with strong economic vitality, positive migration, and active construction will score higher on market support than one in a declining county with outbound migration and elevated natural disaster risk.

7. Max-Bid Solver

The Max-Bid Solver inverts the underwriting question. Acquisition underwriting asks “at this price, what IRR do I get?” The Max-Bid Solver asks “given my required IRR and waterfall, what's the highest price I can pay?”

It returns max value, the achieved IRR at that value, and a plain-English rationale so you can see why the ceiling is what it is rather than treating it as a single fragile number.

7.1 Two modes

7.2 Class A / Class B = LP / GP

The two equity classes use the same waterfall. Class A (LP) is preferred / passive capital; Class B (GP) is the sponsor co-invest plus the promote earned above hurdles. Pick which class's IRR to target with the Investor class to target dropdown.

7.3 Waterfall structure presets

• IRR-hurdle ladder (multi-tier). The Class A / Class B split changes at IRR hurdles. Default seed is the LaPour 3-tier structure: 75/25 below a 16% LP IRR, 65/35 between 16% and 18%, 50/50 above 18%.
• Simple pref + fixed promote (one tier). One row, no upper bound — e.g., 8% pref then 70/30 above. Useful for simpler institutional structures.

You can edit any row, add tiers, or remove tiers. The final tier always has “no upper bound.”

7.4 Bound to a deal (the round trip)

Click Run max-bid solver from a deal's detail page and the solver opens with the deal pre-bound. A banner at the top reads “Bound to deal: [name]”. After the solver converges, the result card adds a “Generate OA draft for [deal name]” button that jumps you to the OA wizard with the deal already in scope.

8. Capital Stack Modeler

Where Max-Bid solves for the price ceiling, the Capital Stack Modeler shows what happens to the cashflows after they cross that ceiling. It runs N capital classes through a multi-tier promote ladder and reports per-class IRR, equity multiple, capital returned, pref paid, and promote share.

8.1 When to use it

• Sophisticated deals with multiple capital classes (LaPour-style Initial Mandatory + Operating Additional + Subordinate + Priority).
• Smaller firms modeling a graduated capital structure for the first time — the page is seeded with LaPour's Creekside Centennial defaults you can adapt.
• Validating how a Max-Bid output actually distributes once you specify class sequencing.

8.2 Inputs

The page is seeded with sensible defaults so you can hit Run immediately and see results before changing anything:

• Capital classes table. Each row: name, contributed capital, pref %, sequence, restriction, promote side.
• Promote ladder. Same multi-tier editor as Max-Bid. The residual after every class is fully repaid flows through this ladder.
• Cashflows. One per line. Year 0 is typically 0; interim periods are operating cash; the last period usually includes the exit / sale proceeds.
• Period years. 1.0 for annual, 0.25 for quarterly.

8.3 The fields that matter

8.4 LaPour-style reference structure

The seed defaults model the LaPour Creekside Centennial structure — the same one used as a benchmark in the underlying waterfall tests. Six capital classes (Initial Mandatory 10% / Operating Additional 10% / FM Additional 10% / Subordinate 0% (manager-only) / Optional 10% / Priority 14% (non-default cure)) feeding a 3-tier promote ladder.

8.5 Results panel

The results panel shows:

• Aggregated investor IRR — the headline return number across all investor-side classes.
• Side totals — investor capital and distributions versus sponsor capital and distributions, plus promote-paid-to-investors and promote-paid-to-sponsor.
• Per-class breakdown table — name, side, sequence, contributed capital, pref %, pref paid, capital returned, promote share, total payout, class-level IRR, equity multiple.
• Rationale — expandable details of every period's payment decisions.

9. Operating Agreement Generator

The OA Generator turns your firm's existing OA template into a draft .docx populated with deal-specific values. ARPIE does not draft legal language — your firm's template is the source. ARPIE fills the variables, surfaces what was substituted, and tracks the shadow → legal_reviewed → executed lifecycle.

9.1 Setup (admin, one time)

Go to Operating Agreement in the sidebar (admins only) and complete two steps:

1. Upload a template. Drop in a .docx file containing {{placeholder_name}} markers anywhere ARPIE should fill in values. ARPIE detects markers at upload time and shows the count. Re-uploading creates a new version; old versions stay on disk so drafts that pinned them still resolve.
2. Configure firm defaults. Set legal-review mode, securities posture, capital classes, operating waterfall, fees, governance threshold, tax allocation methods, formation state. The form is pre-filled with LaPour-shaped starter values to save typing for sophisticated firms; new firms should customize before generating any draft.

Legal-review modes.

9.2 Generating a draft

Two entry points:

• From a deal's hero: Generate OA draft button.
• From a Max-Bid result card (when bound to a deal): Generate OA draft for [deal name].

The wizard has four steps:

1. Pick template. Choose from your firm's uploaded templates. Each card shows the placeholder count detected at upload time.
2. Review placeholder values. ARPIE auto-fills each placeholder with a value sourced from the deal data, your firm defaults, or a derived helper (e.g., today's date). Each row shows a colored badge:
   • From deal (green) — pulled from the deal's underwriting data, e.g., property_name.
   • From firm defaults (green) — pulled from your saved firm-defaults snapshot, e.g., formation_state.
   • Derived (blue) — computed automatically, e.g., today → ISO date.
   • Manual entry (blue) — you edited the field. Editing any auto-filled value flips its badge to manual.
   • Needs your input (yellow) — ARPIE couldn't match the placeholder and is asking you to type a value. These are blockers; Generate stays disabled until every one is filled.
   The match strategy is alias + path: {{property_name}} resolves via a curated alias map; {{deal.purchase_price}} resolves directly via path; everything else falls through to manual.
3. Review and generate. The summary shows a bar of source counts (X from deal, Y from firm defaults, Z derived, W manual, blockers). Generate is disabled if any blockers exist.
4. Result. The draft is generated in shadow state. You get a Download .docx button plus a diagnostics panel showing any placeholders the template had that you didn't supply (left blank in the doc) and any values you supplied that the template doesn't reference (typos or stale field names).

9.3 The lifecycle: shadow → legal_reviewed → executed

Every draft starts in shadow state. Shadow drafts are not legally binding. They're working documents for your attorney's redline.

The Operating Agreement tab on a deal page lists every draft for that deal with a state badge (Shadow / Legal reviewed / Executed) and inline forms for the state transitions:

• Confirm legal review. Records the attorney name, review date, and an optional redline summary. Flips state from shadow to legal_reviewed.
• Confirm executed. Records the execution date (when all members signed). Flips state from legal_reviewed to executed. This is a one-way lock — executed drafts cannot be modified.

9.4 Amendments

Once a draft is executed, the row changes to show a Create amendment button. Clicking it opens the OA wizard in amendment mode with the executed parent's values pre-loaded:

• The template-picker step is skipped — amendments always use the parent's template (pinned automatically).
• Every placeholder is pre-filled with the value from the parent draft, shown with a green From parent draft badge.
• Edit any value that changes for this amendment. Editing a field flips its badge to Manual entry; the rest stay tied to the parent.
• Click Generate amendment to produce a new shadow draft with all the values (parent's plus your edits) substituted into the .docx.

The new amendment carries forward the parent's firm-defaults snapshot, per-deal overrides, member roster, and placeholder values — then walks its own shadow → legal_reviewed → executed lifecycle. The original executed draft stays immutable. Multiple amendments are allowed; each is numbered (Amendment #1, #2, ...) and chained back to the same executed parent.

Edge case: amendments off drafts generated before May 2026 may show empty placeholder values, because those parents predate the value-persistence feature. The wizard will let you fill them manually before generating.

9.5 ARPIE's legal posture

• Mechanical substitution, not legal drafting. ARPIE fills {{placeholder}} variables in your firm-supplied template. We don't generate legal language.
• Every output is shadow-state by default. No generated draft is binding until your attorney confirms review.
• Templates are firm-supplied. ARPIE may surface starter templates derived from anonymized executed OAs, but firms are expected to substitute their own templates for production use.
• Audit trail. Every state transition records who confirmed, when, and (for legal review) the attorney name and redline summary.

10. Portfolio Management

10.1 Adding deals to your portfolio

Once a deal has been analyzed, click Promote to Portfolio in the deal detail view. You can also import existing holdings via CSV or Excel from the Bring in deals page.

10.2 Portfolio dashboard

• Holdings tab — All portfolio deals with key metrics and scores.
• Analysis tab — Portfolio-level analytics, allocation buckets, weighted metrics.
• Rankings tab — Holdings ranked by score, return, or risk.
• Market Map — Geographic visualization of your portfolio.

10.3 Pipeline comparison

Compare multiple deals side by side via Analysis Runs. The executive workbook export compiles comparisons into a single Excel file for leadership review.

10.4 Investment decision board

A configurable view that categorizes deals by investment readiness based on score, risk level, data completeness, and portfolio fit.

11. Reports and Exports

12. Team Management

Roles

Inviting team members

Go to Team & Access → Invite Member. Enter the user ID, select a role, and share the generated invite link.

Session management

Admins can view access history, active sessions, and revoke sessions immediately from the Team page.

13. Frequently Asked Questions