EIRA

Urban intelligence for sustainability leaders

Adaptation you can verify, asset by asset.

Catastrophe models tell you where the risk is and how big it is. EIRA tells you what can be changed there, at what cost, and whether it worked.

EIRA turns the underused surfaces of an urban portfolio — rooftops first, then courtyards, façades and residual spaces — into measurable climate value. For an insurer that is a variable your models do not have yet: which assets can create resilience they are not creating today, and what actually changed once they did.

Start with a portfolio screen Propose a research pilot Four ways to start

What we do

The layer after the catastrophe model.

In your models, vulnerability is a curve you estimate. We want to make it a variable you can move. EIRA sits on top of the catastrophe model; it does not replace it. You identify the risk. We ask what can be done about it, and then we check whether it was.

We screen

Which assets can move

Every roof is different, but nobody goes roof by roof. For every building in Spain there are public data — cadastre, LiDAR, satellite, rainfall — about fifteen numbers per roof. Rules tell us what fits: flat roof, enough load, no heritage protection. Physics tells us roughly what it does.

We specify

Which measure, at what cost, with what effect

For each asset, the interventions that are physically, legally and economically possible, a cost range, and an expected effect on heat and water. Expressed as a variable you can put next to your own models, not as a certificate on a wall.

We verify

Whether it worked

Before-and-after measurement on the built intervention, checked by someone who is not the owner, so the number can enter an underwriting conversation. Every verified intervention improves the reading of the next one.

Portfolio screensIntervention feasibilityHeat & water metricsCost rangesBefore / after verificationAuditable, AI Act-readyCSRD-ready outputs

Ways to start

Four ways in. None of them touches your pricing model.

We are not asking anyone to price resilience. We are asking where insurers already spend — their own buildings, their corporate clients' prevention, their research budgets — and offering something that can be tested in weeks, not years.

01 · Portfolio screen

Your own real estate

For the buildings the insurer owns and reports on.

What it is
A ranked list of the assets in a portfolio that can create resilience they are not creating today: which surface, which intervention, cost range, expected effect on heat and water. Screening is software; it runs on data that already exist.
What you receive
A prioritised list, a viability map per asset, a cost–impact matrix and a pilot shortlist. Reporting-ready for CSRD and your own sustainability disclosures.
What we need
The asset list — addresses or cadastral references — and one counterpart on your side.
Timing
Four to six weeks.
The test
Put the list next to your models, keep the models exactly as they are, and ask whether it tells you something you could not see before. If not, you have lost a spreadsheet.
02 · Prevention layer

For your corporate clients

Alongside the loss-prevention engineering you already do.

What it is
Your corporate underwriting already sends an engineer and recommends protective measures; often the incentive is insurability itself, not a discount. The prevention layer is the same logic one level up: for a client's portfolio, the measures that reduce heat and water exposure at the asset and the block, with cost and expected effect, that your engineers can verify.
What you receive
A prevention report per client, with interventions ranked by effect and cost, ready to attach to the risk survey.
What we need
One client segment to start with — hotels, logistics, retail — and the survey format your engineers use.
Timing
Six to eight weeks for a first segment.
03 · Research pilot

One insurer, one portfolio, one question

Run through the Chair in AI for the Built Environment, UDC · CITIC.

What it is
The question is simple: does verified adaptation tell you something your models do not see today? The Chair screens the portfolio and publishes the method. EIRA designs and delivers the interventions you choose to move. Whoever measures does not sell the intervention; that separation is deliberate, and it is what makes the evidence worth anything to an underwriter.
What you receive
An independent screen, a documented and auditable method, and the first before-and-after evidence on real assets. You keep the results.
What we need
One portfolio, a named counterpart, and patience: the screen takes weeks; the first measured results take a year.
Timing
Screen in six weeks. Evidence in six to twelve months.
Where it works best
Where the insurer and the owner are the same balance sheet. Nobody has to move first.
04 · Verified interventions

Build the history

The first rooftops, measured before and after.

What it is
Design, delivery and monitoring of the first interventions on the assets the screen ranks highest — rooftops first — with sensors before and after. A ten-centimetre green roof keeps about seventy per cent of a normal rain out of the sewer and cools the floor below by four to six degrees; in a big storm, much less. That difference is exactly what has to be measured where the asset stands.
What you receive
Built resilience on your own assets, measured, reportable, and a data point you own. Every verified intervention sharpens the screen for the next one.
What we need
One to three assets and a decision to move.
Timing
First results within a season.
Start with a portfolio screen Propose a research pilot

What we have, and what we do not.

Insurers should be sceptical. A new resilience score, offered for a pricing model, deserves a no. What we offer instead is three things that in 2026 matter more than history.

Traceable

Every number can be rebuilt

Every score is reconstructed from open or auditable data. Under the AI Act a model an insurer cannot audit is a liability, not an asset. Ours is built to be checked.

Complementary

A variable you do not have

We add what can be changed, at what cost, with what effect. We do not touch the variables your models already carry, and we do not replace the catastrophe model.

Measured

Before and after, on real assets

Performance measured on the built intervention, at the asset and the block, checked independently — the only kind of evidence that can start an underwriting conversation.

What we do not have is thirty years of claims on interventions nobody has built yet. That history can only be built by building. Which is why the first step is a screen, not a score.

Why it matters

Not five per cent off a premium. Insurable at all.

In parts of Spain the live question is no longer the discount; it is whether an asset stays insurable. Where catastrophe risk is socialised, the private insurer has little economic reason to reward adaptation — and every euro of verified adaptation is a euro of loss that may never be paid. That is a design question for regulators as much as for underwriters. Either way, the proof of prevention has to exist first.

How much resilience could this asset create that it is not creating today?

The platform

One platform. Two layers.

The intelligence layer

EIRA Core

Core reads every property through cadastral, geospatial, climate and regulatory data — satellite imagery, LiDAR, rainfall, planning constraints — with expert validation, and produces structured intelligence at single-asset and full-portfolio scale. It answers what is possible, where, under what conditions, and with what measurable effect.

The design & delivery layer

AIRE Studio

AIRE Studio turns Core's intelligence into interventions: concepts, technical coordination, delivery and monitoring. Core identifies the opportunity. AIRE builds it and measures it. Whoever measures does not sell.

Research-born. Built for implementation.

Start with one portfolio.

A screen takes weeks and asks one question of your models. A research pilot takes a year and gives a metric its first history. Both start with an asset list and a conversation.

Start with a portfolio screen Propose a research pilot Ask about the prevention layer