Trust
Visible delegation
How logs, review surfaces, and explicit approval points affect trust in agentic work.
Research
Proof work for continuity software.
Navigic studies how Mac-first AI workspaces can carry memory, delegated action, and visible control through real work without hiding review from the person responsible.
Visible delegationMemory and continuityHuman oversightManaged runtimes
Research programMemory, delegation, continuity, and trust
Early work is labeled as preview or in progress until the method, limitations, and supporting materials are ready.
Program lanes
Questions that shape the lab notebook.
Memory
Memory and continuity
How persistent context changes repeated knowledge work without asking people to rebuild state each session.
Human-Computer Interaction
Human oversight
How people move between planning, delegation, review, and execution when AI work remains visible.
Agent Workflows
Managed runtimes
What users need to understand about hosted workspaces, provisioning status, repair, and teardown.
Featured research
Current studies and previews.
These entries are intentionally conservative. Preview work stays unlinked until a complete research note or detail page exists.
Preview / Evaluation / Evaluation
Visible delegation evaluation
A preview protocol for studying whether logs, status, and approval points help people review delegated AI work.
Preview /
Authors: Navigic
Methods: Protocol design, Interface review, Task walkthroughs
Open research artifacts
In progress / Memory / Field study
Memory and continuity field notes
An in-progress note on how durable context may change repeated planning, review, and follow-through.
In progress /
Authors: Navigic
Methods: Field notes, Workflow observation, Product instrumentation planning
In progressPublication index
Crawlable record of planned and active work.
- Preview /
Visible delegation evaluation
Evaluation / Evaluation
A preview protocol for studying whether logs, status, and approval points help people review delegated AI work.
Open research artifacts
- In progress /
Memory and continuity field notes
Memory / Field study
An in-progress note on how durable context may change repeated planning, review, and follow-through.
- Preview /
Operating loops design study
Agent Workflows / Design study
A preview design study on how users move from planning to delegation, review, and execution.
- In progress /
Managed runtime trust note
Trust / Technical report
An in-progress technical note on what hosted workspace status, repair, and teardown should make visible.
Open research
Keel is the public evaluation repository.
Keel holds NavigicAI's early, public-safe research materials for personal-agent evaluation: methodology notes, benchmark landscape work, and reviewed examples as they become ready.
It supports the research program without replacing this hub, and it should be read as an early repository rather than a complete benchmark leaderboard or production telemetry product.
Explore Keel on GitHubMethod and trust
What qualifies as Navigic research.
Research pages are reserved for artifacts with a method, evidence, or reusable finding: field notes, design studies, evaluations, and technical reports. Claims that depend on benchmarks, customer outcomes, safety, privacy, or compliance need visible support before they are presented as findings.
Published work should include authorship, update history, limitations, and source materials when available. Questions, critique, and collaboration notes can be sent to [email protected].
For practical walkthroughs, visit resources. For the current product path, see product overview and Argo for Mac.
From research to product
Bring one active loop aboard.
The research program informs a Mac-first workspace for continuity, delegated work, memory, and visible review.