Part I · The Opening — What "We Actually Do" Means at a Property

Most AI proposals begin with what the system could do. This plan begins with what it is doing — on your data, for your properties, with outputs that arrive on Tiffany's desk on specific days of specific weeks.

Adam, you have read enough vendor decks to know the pattern. Big promises on slide three, ninety-day "discovery phase" on slide seven, eighteen-month implementation starting on slide eleven. The industry trained you to translate those slides into months-of-nothing-happening and dollars-out-of-the-P&L-before-dollars-in. That translation is correct for almost every vendor you have ever met. It is not correct for this plan.

So the plan is organized not by phase numbers but by named days. Day 30. Day 90. Day 180. Day 365. Each day has a specific output, a specific owner, a specific measurement. Each day's output decides whether the next day's work gets commissioned. That is the shape.

Part II · Day 30 — Five Named Deliverables, Each Owned

Not five "areas of focus." Five deliverables. Each has a named week, a named output, a named recipient at Equinox, and a named Genesis agent responsible for producing it. No cost to Equinox. The only input required is read-access to PMS and the OTA account.

Why Phase 1 is free — said honestly, not coyly

Because Genesis does not need an Equinox-scale engagement to produce the Phase 1 output — it needs a Phase 2 decision, which requires a Phase 1 that is genuinely useful on its own. The five deliverables at Day 30 are Equinox's to keep whether or not Phase 2 is ever signed. If the WiFi ROI is wrong, the competitive intelligence is thin, or the quick win does not materialize, Equinox owes Genesis nothing and walks away with less damage than thirty minutes of a Monday sales call would cost. That is the deal. It is that shape on purpose.

Named decisions at Day 30

1. Did the Phase 1 output arrive on time, at quality? Yes / no.
2. Would Tiffany act on it even without Genesis in the room? Yes / no.
3. Is the quick win measurable? Yes — here is the number / no — here is why.
4. Does Adam want Phase 2 at Richardson? Yes — sign outcome-based terms / no — thank you, keep the deliverables.

Part III · Day 90 — Four Intelligence Capabilities Live at Richardson

If Day 30 produces the five deliverables and Adam says yes, Day 90 is when four intelligence capabilities go into live operation at Sonesta Select Richardson. "Live" means: running in production, integrated with PMS, outputs arriving at named people on named days, measured against named KPIs.

The six success KPIs — tracked weekly, reviewed at Day 90

Phase 2 90-day pilot value range: $93K–$171K to Equinox. Genesis compensation: percentage-of-incremental, capped; the full number is inside the commercial addendum attached to this document, not in the proposal body. Every dollar paid to Genesis in Phase 2 is a dollar that was not in Equinox's P&L before Phase 2 existed. That is the line.

Part IV · Day 180 — Portfolio Extension Begins

Day 180 is the moment Richardson's Phase 2 data decides what the portfolio rollout actually looks like — not what was forecasted at the start.

If Phase 2 delivered against the KPI bar, Genesis extends the four Phase 2 capabilities to the three remaining Sonesta DFW properties between Day 90 and Day 180. Not all three at once. Sonesta ES Suites Dallas Richardson first (same market, deepest data overlap), then Sonesta Simply Suites Richardson, then Simply Suites Fort Worth. The interval between each rollout is two to three weeks, not parallel — because what Richardson learns in week 14 gets baked into the ES Suites rollout in week 16, and what the ES Suites rollout learns in week 17 gets baked into Simply Suites in week 19. The portfolio compounds.

Part V · Day 365 — Full Portfolio, Compounding

Day 365 is Tribute Arlington (near AT&T Stadium, converting to Marriott Tribute Portfolio) and Tribute Marin (Marin County, California — 235 keys, still under renovation). Both properties are strategic, both are in transition, and both benefit from a brand-transition intelligence layer that Genesis has already been running in the background since Day 30 — because brand transitions are data-heavy decisions that reward a year of context.

Success at Day 365 — what it looks like, named by property

Part VI · Priority Matrix — High-Impact / Low-Effort First, Ruthless About the Rest

Not everything that could be done should be done. The list of capabilities Genesis could deploy at a Select-Service property is longer than any single year can absorb. So the plan stays disciplined by a 2×2 priority matrix. Everything lands in one of four quadrants; only Q1 gets attention in the first ninety days.

Part VII · Expected Lift per Intervention — Drawn as Bars, Not Bullet Points

The same numbers that live in Document 04's ROI model, re-drawn here per intervention at Equinox-portfolio scale, so the plan can be read without flipping to another doc.

Part VIII · What Only Genesis Does

Everything above could, in theory, be produced by a sufficiently-motivated combination of a revenue-management consultant, a review-response agency, a corporate-sales team, and a good CIO. It would be slower, more expensive, and it would not compound. It also would not be the thing Adam is actually being offered. What Adam is being offered is a system with three structural properties that nobody in the hospitality-AI market provides today.

What each of the three structural properties means for Equinox

The bio-mimetic memory graph. Traditional AI systems have short memory. They answer the question in front of them and forget. Genesis remembers — every guest interaction, every pricing decision, every comp-set shift, every corporate RFP cycle — in a graph that compounds across time. By Day 365, the system has more operational memory of Equinox's properties than any GM who has worked at the company for less than five years. That memory is Equinox's, not a vendor's.

The dual-model actor-critic. Every recommendation Genesis produces is generated by one model (the actor, Qwen3.5-397B on a 397B-parameter MoE architecture) and critiqued by a second independent model (the critic, GLM-4.7) before it reaches the GM's inbox. The golden-ratio weighting — 61.8% analytical, 38.2% creative — is the architecture of the whole system. What it means in practice: the Tuesday-night rate recommendation that shows up in Tiffany's email has already been adversarially reviewed by a second intelligence before anyone at Equinox ever sees it. Nobody else in the market does this.

The sovereign data path. Equinox's data never leaves the Genesis organism. There is no "phoning home" to OpenAI, no Anthropic API call, no data shared with any third party. Genesis runs on an 8×H200 GPU cluster, privately hosted, with operator data flowing in and operator intelligence flowing out — period. The data Adam contributes to Genesis is Adam's competitive advantage, not a training corpus for a competitor's next product.

Part IX · The Close — One Conversation, One Pilot, One Decision

This plan is long on specifics because the planning philosophy is short: name the work, sequence the work, own the outcome. Every page above is an extension of those three sentences.

What Genesis is asking for is small on purpose. Not a capability contract. Not a multi-year commit. Not a signing ceremony at a conference.

The commitment from each side — written symmetrically

What Equinox keeps no matter what

Everything in this package is yours whether or not the conversation ever happens. If this plan, the Corporate Vision at operator scale, the Deep Dive on Richardson, the DFW Portfolio ROI, the Tax Credits & Incentives brief, and the FIFA Playbook all land in a folder and stay there, Equinox still has approximately one and a half to five million dollars of unclaimed tax credits identified, a Richardson ROI model built on your actual competitive set, and a FIFA 2026 playbook for your DFW portfolio. That alone more than covers the time it took to read.

If it goes further — we grow together.

— C.