AI for data quality and analysis – Scaler

Purpose of this article

Scaler includes a set of AI features that help you understand, clean, and complete your data. This article walks through each one — what it does, where to find it, and when to use it.

For the broader AI overview and security architecture, see AI in Scaler: overview.

✨

Lumi AI is part of the Scale Plan. Every user receives 15 prompts/month free; more requires the Scale Plan — contact your account manager. Some organisations have AI features disabled. Access via the floating icon, the homepage, or inline throughout the platform.

What this covers

Lumi data analysis on the Asset List — natural-language queries against the asset database

AI data verification — outlier detection across the portfolio

AI meter analysis — meter prioritisation for data coverage impact

AI forecasts — filling missing meter consumption values

Contextual AI suggestions on Analytics pages

AI-assisted data collection — climate zones today, more coming

Bill scraping and verification — automated invoice extraction

Lumi data analysis on the Asset List

Purpose

Ask natural-language questions against the underlying asset database without configuring filters or joining columns by hand.

Location

Analytics Portal → Portfolio → Asset List

Type your question in the search bar at the top of the table, then click Ask Lumi.

Example queries

"Which assets have an energy label C or worse?"

"Show all residential assets in Germany with data coverage below 80%."

"Which assets have the highest EUI?"

"List assets without a GRESB Asset ID."

Lumi returns a filtered, sorted view of the assets that match your query, with a short summary and follow-up suggestions. Results can be exported.

AI data verification

Purpose

Scan the portfolio for outliers, missing values, and inconsistencies before reporting or audit.

Location

Analytics Portal → Portfolio → Asset List → Data Outliers tab

Click AI Analysis at the top of the table.

What it catches

Missing consumption data for energy, water, or waste

Implausible energy-to-GHG ratios — for example, values 2.5x above the portfolio average

Contradictory like-for-like values, such as like_for_like nulls despite 100% data coverage

Extreme intensity outliers compared to portfolio benchmarks

Output

A right-side panel with an executive summary and findings categorised by severity. Each finding includes:

A description of the issue

A likely root cause (e.g. meter setup, inconsistent readings)

A specific recommended action

When to use it

Run AI data verification before generating any report, and again before submitting to a third-party verifier or framework body. It significantly reduces the manual data-quality review your team would otherwise do line-by-line.

Tip: Run AI data verification weekly during reporting season — fixing issues earlier in the cycle is much cheaper than finding them at submission.

AI meter analysis

Purpose

Identify the meters where additional data collection has the highest return on investment — both for data coverage and for downstream scoring outcomes.

Location

Data Collection Portal → Portfolio → Meter List, or ask Lumi a meter question from anywhere:

"Analyse my portfolio's meter consumption patterns and flag the meters that need attention."

"Which 20 meters should I prioritise to lift my data coverage the fastest?"

Why it matters

Data coverage is the single biggest scoring driver for GRESB and the foundation for reliable insight. With hundreds or thousands of meters across a portfolio, manually auditing each one is impractical. AI meter analysis surfaces the meters where action will move scores or improve insight reliability the most.

Best use case

Run before any major reporting deadline, and at any point when data completeness is below your target.

Contextual AI suggestions

Purpose

Offer page-specific prompts tailored to whatever you're currently looking at, so you don't have to invent questions from scratch.

Location

Every Analytics page surfaces contextual prompts. The prompts adapt to the current view:

On the Overview page — comprehensive performance assessments, decarbonisation analyses, tenant engagement strategies

On the Scores page — scoring gaps and improvement opportunities

On the Performance page — drivers behind energy, GHG, water, and waste trends

Why it matters

You don't need to know what to ask. Scaler surfaces the right questions at the right time, which makes AI useful for users who aren't sustainability specialists.

AI-assisted data collection

Purpose

Reduce manual setup work by suggesting asset characteristics from location and building data.

Location

Data Collection Portal → Asset → Asset Characteristics → Asset details

Available today

Climate zone — click the Lumi icon next to Climate zone to get an ASHRAE classification suggestion based on the asset's location, with rationale. Click Apply Climate Zone to populate the field.

Coming soon

Energy label auto-retrieval from national registries

Certification auto-detection

Address validation and geocoding

Bill scraping and verification

Purpose

Automate consumption data extraction from utility invoices to reduce manual entry.

Location

Data Collection Portal → Asset → Meters & Consumption → Meter consumption (paperclip icon)

Process

Upload a PDF utility invoice at a meter consumption entry

AI extracts the key fields — meter ID, billing period, consumption, cost

Extracted values populate the relevant fields automatically

Optionally connect to Scaler's verification partners for a second-pass check

Advantages

Saves time on transcription

Reduces transcription errors

Supports auditable, consistent data entry

Works alongside manual and API-based collection workflows

Always review extracted values before saving — bill formats vary across providers and the AI's confidence varies with format quality.

Best practices

Run AI data verification before generating reports and again before submission

Use AI meter analysis before any major data-collection sprint to focus your effort

Treat AI-suggested forecasts and characteristics as drafts to verify, not finished entries

Set up bill scraping for meters where you have a regular invoice cadence — it pays back fastest there

Where to go next

AI in Scaler: overview — Architecture, security, and the full feature list

Using Lumi: writing effective prompts — Prompt techniques and example libraries

AI for action and reporting — CapEx measure suggestions, AI report reviewer, and Investor DDQ answers

Setting up roadmap inputs — Configure AI-suggested and manual roadmap measures