Advanced AI capabilities, working inside the product you already use — not a chatbot bolted onto a settings page.
From utility-bill scanning to a Scope 3 classifier to an in-app GHG Protocol expert, every claim on this page is something you can open the app and verify in under a minute.
Powered by state of the art LLMs. Cached, audited, and permission-gated.
Three concrete jobs, executed end-to-end inside North Star.
Drop in a utility bill, an invoice, or a Scope 3 spreadsheet. The platform extracts amounts, dates, sites and emission categories — with confidence scores you can review before anything is saved.
Ask the in-app assistant a GHG Protocol question, or have it walk you to the right page. It speaks, it listens, and it knows your role and permissions before answering.
Building a chart, drafting an ESG response, or scoping your inventory? AI proposes the answer; you accept, edit, or override. Suggestions are cached, so identical questions cost nothing.
Click the chat icon in any header. Ask a GHG Protocol question, ask where to enter your Scope 2 data, or ask why a number changed. Navigator answers in plain English, walks you to the exact page, and can speak the answer aloud. Multi-turn history, thumbs-up/down feedback, one-click handoff to support.
Verify it: open the platform → click the chat bubble in the top-right header on any page.
Type “Show Scope 1 and 2 emissions over the last 3 years by site” — the assistant proposes the chart type, time range, scopes, sites, sources and metric in two passes (shape first, values second). You preview, tweak, save. Responses are cached per-prompt so building the same chart twice is fast and easy.
Verify it: Dashboards → Add panel → “Generate Suggestion” with the wand icon.
Drop in a utility bill, carrier invoice, or supplier statement (single or multi-period). The scanner extracts amounts, units, dates, supplier and matches each line to the right site, returning a confidence score on every suggestion. You approve before anything is saved. Document hashes are cached so re-uploading the same PDF is instant.
Verify it: Datalines → Upload Document → drop a PDF.
Drag in your Purchased Goods & Services or CapEx spreadsheet. Each row is auto-classified into the correct Scope 3 category and emission source with a confidence score. Override anything the model gets wrong before importing. Same caching and audit trail as the rest of the AI stack.
Verify it: Datalines → Upload Document → Scope 3 Category Import wizard.
Free-text item descriptions get matched to your emission source hierarchy using an embedding model that runs entirely client-side, accelerated by WebGPU with a WASM fallback. Your raw item names never hit our servers during matching. Configurable similarity threshold per company.
Verify it: No outbound API calls during the matching.
Stuck on a CSRD or CDP answer? AI drafts a contextual response from your previously-entered data and company guidance.
Verify it: open any ESG questionnaire field → AI suggest.
On day one, AI reads your industry and company description and recommends which scopes, categories and sources are material. New users skip a week of consultant calls.
Verify it: lite onboarding → Inventory Boundary step.
Smaller capabilities that ship with every account.
When you build a reduction target, AI proposes actions and strategies tailored to the source type and your baseline-vs-target gap. Permission-gated to MANAGE_SCENARIOS roles.
Find it under Scenarios → New scenario → AI suggest.
Upload your company logo and AI strips the background running in-browser via WASM. A small touch — and a representative one: AI doesn’t have to mean cloud.
Find it under Settings → Company → Upload Logo.
Every AI feature is logged and cached.
Book a 30-minute walkthrough — we’ll run the AI features live against a sample utility bill and a Scope 3 spreadsheet you bring.
