Purpose of this article
Scaler includes a set of AI features built around Lumi, your in-platform assistant. This article explains what AI in Scaler does, where to find each feature, and how it keeps your data secure.
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 you can do with AI in Scaler
Scaler's AI features cover the full ESG data lifecycle — from input and validation, through analysis and forecasting, to action planning and reporting. They are designed to reduce manual review effort and bring decision-relevant insights closer to the work you're already doing.
Lumi works in two ways:
- As a chat assistant for natural-language questions about your portfolio, scores, and how Scaler works
- As an embedded layer across the platform — AI Analysis buttons, AI suggestion buttons, and contextual prompts surface inside the workflow you're already in
The features available today group into three areas, each covered in a dedicated article:
- Data quality and analysis — outlier detection, meter prioritisation, forecasts, contextual analysis, asset list queries, AI-assisted data collection, and bill scraping
- Action and reporting — AI CapEx measure suggestions, the AI report reviewer, and Investor DDQ answers
- Prompt-based interaction (Lumi) — natural-language access to all of the above, plus general portfolio and platform questions
The capabilities at a glance
Capability | Where to find it | What it does |
AI chatbot (Lumi) | Home → Welcome plus the Lumi icon, available from any page | Ask anything about your data, scores, or how to use Scaler |
Contextual AI suggestions | Every Analytics page (suggestions adapt to the current view) | Page-specific prompts tailored to what you're looking at |
Lumi data analysis | Analytics Portal → Portfolio → Asset List | Ask natural-language questions against the asset database |
AI data verification | Analytics Portal → Portfolio → Asset List → Data Outliers | Detects outliers and inconsistencies across the portfolio |
AI CapEx measure suggestions | Data Collection Portal → Asset → Assessments & Measures → Add Measure | Recommends retrofits with cost and payback estimates |
AI meter analysis | Data Collection Portal → Portfolio → Meter List (via Lumi prompts) | Prioritises meters by impact on data coverage and scoring |
AI-assisted data collection | Data Collection Portal → Asset → Asset Characteristics | Auto-suggests climate zones today; energy labels and certifications coming |
AI report reviewer | Data Collection Portal → Portfolio → Saved Reports | Checks framework completeness and consistency before submission |
Investor DDQ answers | Home → Investor DDQ | Drafts data-backed responses to LP and lender questionnaires |
Bill scraping | Data Collection Portal → Asset → Meters & Consumption (paperclip icon) | Extracts consumption fields from utility invoices |
Tip: Most AI features have a counterpart you can also access by asking Lumi a natural-language question. For example, ask Lumi to "show meters with the biggest data coverage gap" instead of opening the Meter Analysis view directly.
Where to find Lumi
Lumi is available from the Home page in the left-hand navigation. The default landing is the Welcome page, which has a Lumi prompt entry and four category tabs — Analyze, Report, Compare, Improve — with suggested prompts under each.
You can also open Lumi from the flower icon in the corner of the sidebar, visible from anywhere in the platform. The icon position is configurable — drag it between the bottom-left and bottom-right corners to suit your workflow.
How AI in Scaler works
Hosted on AWS Bedrock, inside Scaler's infrastructure
All AI features run on AWS Bedrock — Amazon's hosted-models service. Practical implications:
- Models run inside Scaler's AWS environment. Client data does not leave Scaler's data centres when processed by the model.
- EU data stays in the EU. For EU clients, the model and the data it processes remain in EU data centres.
- Models are not trained on your data. AWS Bedrock does not use customer data for model training.
Authorisation scoping
Lumi inherits the same access controls as the rest of the platform. When you ask a question, Lumi can only see the data you are authorised to see — at the company, portfolio, and asset level your permissions allow.
Accuracy controls
Scaler tunes model parameters — specifically, low temperature — to minimise hallucination and produce deterministic, accurate responses rather than creative ones.
Audit logging
Every Lumi interaction is logged: which data was accessed, what the agent decided, and what the user asked. Logs are retained for auditability.
Compliance
AI features sit within the same compliance framework as the rest of Scaler — GDPR, SOC 2, ISAE 3000, and ISO 27001.
How Lumi handles your prompt under the hood
- Your prompt is wrapped in a prompt-management layer that adds context — your active portfolio, your current page, and relevant Knowledge Base content
- The wrapped prompt is sent to a Bedrock-hosted model running with a low-temperature setting
- Lumi authenticates the data lookup using your session token, so it can only fetch records you have permission to see
- The response is logged, then returned to you — either in the chat panel or in the relevant in-page surface
Your prompt allowance
Every Scaler user gets 15 free Lumi prompts per month, regardless of which Scaler package their organisation has. This is designed to give every team member a way to try Lumi and build a feel for what works.
For higher-volume use, contact your Account Operations Manager about licensing the full Lumi package.
If Lumi does not appear in your platform, your account may not have it enabled yet. Contact your Account Operations Manager — activation typically takes minutes. In rare cases, your organisation's agreement with Scaler restricts AI use; if so, check with your legal or admin team.
Getting comfortable with AI
A few patterns that have helped client teams adopt Lumi:
- Use your 15 free prompts every month. Building a habit matters more than any single answer. Even if your organisation has not licensed Lumi yet, the free allowance gives you room to practise.
- Ask domain-specific questions, not generic ones. Lumi is grounded in your portfolio data and Scaler's reference guides, so it performs best on questions like "which assets are losing the most data coverage in 2025?" rather than abstract questions about ESG concepts.
- Treat answers as drafts, not final. Verify AI-suggested measures, forecasts, and DDQ answers before acting on them.
- Iterate. If the first answer is not useful, refine your prompt with more specific scope, a clearer time period, or an explicit output format.
For prompting techniques and example libraries, see Using Lumi: writing effective prompts.
Where to go next
- Using Lumi: writing effective prompts — Write prompts that get accurate, useful answers
- AI for data quality and analysis — Outlier detection, meter analysis, forecasts, asset list queries, and bill scraping
- AI for action and reporting — CapEx measure suggestions, AI report reviewer, and Investor DDQ answers
- Finding help & support — Knowledge Base, Lumi, and contacting support