Nicky reads the thread
Nicky looks for missing context, stale risk, weak descriptions, poor evidence, unresolved attention and other signals that make complex work harder to trust.
In active development
Nicky is Nolta's built-in context intelligence layer.
Nicky is no longer a distant idea. The first working version is alive: it can scan project context, detect weak areas and generate reviewable proposals for humans to inspect before anything becomes part of the record.
The goal is not a generic chatbot. Nicky reads the Nolta context layer: graph structure, timeline, heat, links, risks, promotions, evidence and context quality. It helps teams find what deserves attention.
What Nicky does
Context intelligence before AI theatre.
Nicky starts with deterministic sensors and Nolta's own context model, then turns findings into explanations and proposals that people can review. It exists to improve the thread, not to bypass it.
- Scans Nolta context using the graph, timeline, heat, links, risks, promotions and context-quality signals.
- Finds weak or thin context, stale areas and nodes that need better evidence, descriptions or follow-up.
- Creates inspectable findings and proposals instead of silently changing project records.
- Works with Nolta's human review loop: proposal in, approve or reject, trusted record after review.
- Keeps the deployment principle: customer data stays in the customer environment.
Product direction
Nicky follows the context. AI work stays reviewable.
Nicky is always part of Nolta. It is the Nolta-native intelligence layer that understands project context. AI execution is different: optional work worker or AI profile for teams that need a local brain to perform tasks.
Humans stay in control
Findings and proposals are reviewable. Nicky should help people improve the project record, not silently rewrite it behind their backs.
Next hardening work
From working loop to polished product.
The core loop exists. The next step is making the review experience clear, calm and trustworthy enough for real users.