Navigator

Who it is for

This section is for project managers, project administrators, and analysts: fast access to accumulated context — where a decision lives, who owns it, what has already been agreed — without manually combing through the tracker and knowledge base.

Operational planning, issue structure, and control of the approved scope — in Digital project administrator.

The Navigator is a working mode built on universal search: the answer to "where is it written?" and "who is responsible for this?" with links to primary sources. The Navigator's strength is that it combines three layers of information: your data, joint work on the results, and the outside world.

The three layers of the Navigator

1. Your data: the knowledge library and document base

The Navigator searches by meaning across all your sources at once:

• Tracker: issues, epics, statuses, comments, worklogs;
• Confluence knowledge base: pages, protocols, procedures;
• Uploaded document base: you upload project documents (PDF, DOCX, MD, and other formats) — the system parses them, including tables, and indexes them. The agent finds the relevant fragments by meaning and cites them with a source reference.

Documents are distinguished by level:

Level	What it is	Who has access
Session documents	drafts of the current conversation	conversation participants
Project documents	the durable project base	the project team
Global corporate corpus	PMO standards and materials shared across projects	all host users

Every answer includes a Sources block — links to specific issues, pages, and document fragments so you can verify a claim in one click.

More details: Knowledge library and Universal search.

2. From answer to document: the collaborative editor

The Navigator does not just tell you "where it is written" — it opens the result:

• a Confluence page — for reading and agreed edits;
• a project working document (charter, scope, requirements) — in the same editor with change review;
• a schedule (Gantt) — when the question is about deadlines and dependencies;
• a web page — when the source lives in an external portal.

In the working area, you and the agent work on one live document: the agent proposes changes, and you accept or reject them. This turns a Navigator answer into an executed action — an updated page, a current schedule, a prepared document.

All surfaces are described in Collaborative editors.

3. External research

When the answer is not in your data — technology selection, market practices, regulatory requirements, benchmarks — the Navigator conducts external research:

1. The agent formulates search queries and searches the web through your perimeter's model gateway (see Data sovereignty).
2. It gathers material from several sources and cross-checks them against each other.
3. It synthesizes an answer with source citations: every conclusion is backed by a link, not an "AI opinion."
4. If needed, it opens the found sources in the collaborative browser to break down their content with you.

Important: external research is a supplement, not a replacement, for your data. In the hierarchy of sources, your systems always take priority: first the facts from the tracker and documents, then recorded decisions, and only then the external context. If the external data is insufficient, the agent says so honestly instead of guessing.

Key competencies

• Removing information noise: a precise answer instead of reading dozens of pages.
• Synthesis across sources: the issue tracker + knowledge base + uploaded documents in one answer.
• Hidden links: issues and documents about the same functionality without explicit links in the system.
• Context onboarding: quickly pull up a module's status, owners, and decision history before a stand-up or escalation.
• Research work: gathering information on new topics with verifiable sources.

How to use it

There is no dedicated command: ask a question about search, context, or research — the agent applies the search, indexing, and external-query tools.

Example requests on your data:

• "What has already been agreed on the release date of module X, and where is it recorded?"
• "Who owns the payment-gateway integration, and what open blockers does it have?"
• "Collect the architecture decisions from meeting protocols over the last month".

Example requests on uploaded documents:

• "What does the uploaded specification say about the acceptable API response latency?"
• "Find the force-majeure notice-period requirement in the contract".

Example external research:

• "Compare approaches to organizing work acceptance in projects of this scale — what is practiced today?"
• "What regulatory requirements apply to storing access logs in our sector?"

Skills used

At its core — Universal search: semantic access to the single source of truth across the project and the uploaded document base. Working with results goes through the collaborative editors. Requirements gathering and formalization — in Requirements work (for analysts).

Why it works

The Navigator searches by meaning, not just by words, and works with a hierarchy of sources: the facts from your systems always take priority over external ones. Every answer carries a Sources block — you can verify it before making a decision on the plan, risks, or status. And when the data is external, the agent marks that explicitly and does not mix it with your facts.