RSV Global / Work / LeaseOasis Decision System
Proposals that cannot be compared,
turned into a defensible decision.
Property owners receive short-term-rental operator proposals that cannot be compared: different fee bases, different revenue definitions, different inclusions and different reporting. This turns them into one comparable economic picture and a decision somebody can defend.
Proposals that look comparable and are not.
Operators present economics on incompatible bases. One quotes a fee on gross revenue, another on net. One includes linen and utilities, another does not. One reports occupancy, another reports nights sold. Placed side by side the proposals look comparable and are not, so the choice gets made on presentation quality.
Independence
The system is only useful if it is not run by an operator, because every operator platform compares the market to itself.
Determinism
An investment committee cannot act on a number that changes between runs on the same inputs.
Evidence quality
Proposals arrive with gaps, and the gaps are frequently the point.
Comparability
Useful comparison is cross-operator and same-building, not against a generic market average.
How the system is put together.
The choices that shaped it.
Each of these was a decision with an alternative, taken deliberately.
AI explains the result, it does not calculate it
The economics are deterministic code. The model writes the narrative around the numbers and never touches the numbers themselves. This is the boundary the entire product depends on.
Missing evidence is scored, not filled
An incomplete proposal stays visibly incomplete. Interpolating a plausible figure would make the output look more finished and be worth less.
Independence as the moat
Cross-operator, same-building comparison is only credible from a party with no operator to favour.
Infrastructure defined in code
The estate is Terraform-defined with autoscaling, private administration, federated CI credentials and an operations runbook covering dependencies, troubleshooting and maintenance.
Where AI is used, and where it is not.
- Narrative generation over computed results.
- Summarising qualitative operator differences.
- Drafting the memorandum around fixed numbers.
- Calculating the revenue waterfall.
- Deciding whether evidence is sufficient.
- Scoring an operator.
- Producing any figure that reaches the memorandum.
What it takes to keep it running.
A managed Kubernetes estate on AWS, defined in Terraform, with autoscaling via Karpenter, private administration through a bastion, federated CI credentials via IRSA rather than long-lived keys, and ArgoCD for delivery. The service estate is split across separate application, API, subscription, notification, shared library, migration and assistant services.
What it changed.
Measured outcomes for this system are being compiled from our own operating records. We publish figures only once we can show how they were measured, so this section is deliberately empty until then.
More systems we built and run.
Show us the workflow everyone has learned to tolerate.
We will help determine whether it needs a focused product, an autonomous operating system, a stronger data layer, a review of what already exists, or a simpler fix that involves no AI at all.