PiP · Portfolio Intelligence Platform
Your whole estate, current, and honest about what it doesn’t know.
The front door for people who own or oversee many hotels: the estate as a tree of brands and regions, a league table per node, the exceptions first, and Boardroom’s read of the position with its receipts visible. No venue switcher, no operational chrome. Read monthly, on a big screen, with someone else in the room.
For Investors and boards, owners and CEOs, regional and brand managers, group finance. In preview.
- Performing
- Required
- Gap
- Levers
Performing and required, across the estate. PiP reads every property’s position from the same spine the venue runs on, rolls it up venue-weighted rather than summed, and puts the exceptions ahead of the narrative. Property-level levers are pulled in ViP; PiP is where the owner sees which property needs them pulled.
The whole loop →An owner reads the estate monthly, often with someone else in the room.
PiP is built for that reading, not for a shift. Scope, coverage and freshness are said out loud; the exceptions come before any AI voice; the position is one paragraph with its receipts visible. The test is whether you can look up after thirty seconds and say one true sentence about your estate that you did not know when you sat down. Click a citation to see the row it came from.
Exemplar Hotels Group
Estate · 26 properties · 4 nodes · month to August 2026 · read 06:00 todayExceptions
3 open · covenant first · deterministic, no model involvedInterest cover on the Lakeside facility. Headroom under 0.05×. A covenant, not a cost line.
Second month over the rule. A roster setting, not a bad night.
A coverage gap. Excluded from every estate figure on this board until it reports.
Estate measures
this month · stated on the properties reporting, or not at allLeague by node
portfolio → brand → region · GOP against the estate, in points| Node | GOP % | vs estate | Exceptions |
|---|---|---|---|
| Harbour brand | 33.0 | +1.8 | 1 |
| Coastal brand | 30.4 | (0.8) | 2 |
| Regional · south | 28.9 | (2.3) | 3 |
| Regional · north | 31.7 | +0.5 | 0 |
Boardroom reads the estate
AI · portfolio read · generated 06:00 today · 6 figures checked · 6 of 6 matchThe estate is ahead on rate and behind on margin: ADR $3 over budget1, GOP margin 1.1 points under2. Harbour brand carries the month at 33.0% GOP3. Regional south is where all three exceptions sit4, and one of them is a covenant, not a cost line: Lakeside at 1.31× against 1.35×5. Two properties have not reported, so the margin is stated on 24 of 26 and marked partial6. Nothing in this paragraph is a forecast.
- Read 26 nodes · 24 reporting
- Checked 6 figures against the metric layer · 6/6 match
- Exceptions raised before narration · 3
- Confidence Definitive · primary data · coverage marked
- Definitive
- Interpreted
- Advisory
The estate, whole and current
Scope, coverage and freshness said out loud on the first screen: how many properties, how many reporting, how old the oldest figure is. Intensive metrics are venue-weighted, never summed, so an estate occupancy means what an owner thinks it means.
A tree, not a dropdown
Portfolio → brand → region → property. A league table at every node, and a compare view for two to six properties side by side. The context control is a breadcrumb; there is no venue switcher in PiP at all.
Exceptions before any AI voice
A deterministic exceptions board sits first-class on the home screen: cost lines over threshold, covenant headroom under a floor, coverage gaps. No model is involved in raising one, so none can be talked out of it.
Boardroom, at portfolio scale
The portfolio canvas narrates the estate’s position in one paragraph with a provenance strip beneath it, generated on a schedule and dated. It says which properties it is stated on and marks anything modelled or partial.
Property rooms, read-only
Open any property from portfolio context and read its month-end, market share, rate intelligence and P&L as projections of the venue’s own governed metrics. Nothing is recomputed on the way up. Where you also run the venue, the door into ViP is drawn; where you don’t, PiP says so.
Your own AI, connected
The ViP Portfolio Connector exposes six read-only MCP tools, so Claude or any MCP client can ask the estate a question and cite the drillable original. Every figure it returns is served from the persisted metric layer, never computed by the model.
Boardroom · the AI brain, in PiP
Built for the reader who has to repeat one true sentence.
An operator scans ViP many times a day. An owner reads PiP once a month and must leave able to say one true thing about the estate. Boardroom’s portfolio read is written for that: dated, stated on the properties actually reporting, citing the rows behind each claim, and marking anything modelled or partial. It will not fill a gap with an estimate nobody measured.
How the AI treats your data
- Read-only by construction: PiP holds no data of its own and computes none. One number, one computer.
- Grants are per estate node; an investor persona can read the scorecard and open no property.
- The portfolio read is a persisted snapshot with a generation time, never implied to be live.
- Every AI action and every connector call is written to the append-only audit log.
Ask the portfolio
Free-form questions over the estate with portfolio-scale claims. Parked by rule until the venue record carries at least two resolved months; the read-only canvas ships first.
Personal AI connection
Per-user keys so a board member’s own Claude account can connect to the estate through the connector, minted by an admin. Designed and ruled; building.
Freshness on every layer
The venue canvas already carries live / stale / missing bands per feed. The portfolio canvas is to show its own age and per-property coverage the same way, so a 25-day-old read can never pass as current.
Public host
PiP is served today alongside ViP for the venues in preview. Its own address, pip.hotelstack.ai, lands with the cookie-domain cutover.
If it’s in this section it isn’t shipped yet. We’d rather tell you that here than have you discover it in a demo.