What can current systems actually do?
Executive summary for the cross-field answer
Capability scorecard → benchmark, vision, skill, or forecast deep dive
Interpret a claim, benchmark, technique, or capability gap
Working map · August 2026
The content map for the state of AI-assisted data visualization.
One evidence base is becoming a comprehensive report, a series of standalone essays, several visual companions, and a set of decision tools. This map shows what each piece owns, how it is built, how the pieces connect, and which promised bridges are still missing.
How to read it
Begin with a question, not a filename. The core research and practitioner narratives split into specialist branches, then reconverge in recommendations, packaging, and audience-specific use.
Current draft corpusOnly substantive August 2026 reports and their visual companions appear as content nodes.
Planned work stays separateMissing chapters and decision tools appear as open bridges, not as completed content.
The content system
Each piece has one primary job in the argument.
The lines show conceptual extensions, not publication order or evidence strength. Choose a piece to inspect its question, boundary, and internal argument. Choose a reader path to emphasize related pieces without losing the shared context.
The complete map: orientation, core narratives, specialist extensions, application, and packaging.
Orientation
Core narrative
Specialist extension
Application or packaging
VISVisual companion exists
Shared evidence and update spine
Applies to every piece · primary-source custody · verified claims · gap ledgers · dated forecasts · refresh triggers
Reader paths
The corpus supports several starts, but not several truths.
Each path begins with a real question and ends at the piece that owns the answer. The executive summary and shared evidence spine keep the routes from drifting into separate narratives.
Reader question
Start
Then follow
Decision supported
Where does AI help or hurt real visualization work?
Practitioner and reader experience
Human skills → learning → context-specific evaluation
Delegate, inspect, repair, stop, or retain direct control
What should we build or test next?
State of the research plus the relevant practitioner job
Skills · vision · forecast → Vizier experiments
Choose a mechanism, baseline, evaluation, and promotion bar
How should this work become a coherent publication?
Editorial architecture
Audience and rollout → canonical hub and decision tools
Package, route, test, update, and eventually publish the right unit
Coverage against the promise
The package is strong where the field can be synthesized; the weak spots sit between stages.
The title promises more than a literature review. It also promises a path from measured capability to human work, delivered artifacts, readers, and a maintained body of evidence.
Area of the comprehensive report
Current owner
Current strength
Bridge still needed
Scope, goals, capabilities, and benchmarks
State of the research · capability scorecard · benchmark crosswalk · specialized vision · forecast
Substantive Ideal and current state, fixed-scale history, shared timeline, evidence ladder, task-and-grader crosswalk, convergence, divergence, and negative findings
Public-safe source index and the link from measured capability to changed work and reader consequence
Techniques and system designs
State of the research · agent skills · specialized vision
Substantive Technique spine and specialist roles are clear
One end-to-end system anatomy with information and authority boundaries
Jobs, tools, creators, and readers
Practitioner and reader experience
Substantive Jobs, tool families, adoption signals, failure costs, and reader evidence
Production cases through delivery, maintenance, mobile, and assistive use
Human capability and learning
Human skills · learning when AI can make the chart
Substantive Capability profiles and performance-versus-learning are explicit
Delayed construction transfer and causal evidence about sustained AI use
Recommendations and experiments
Capability scorecard · forecast · Vizier assessment · gap ledgers
Specified Gates, signposts, experiments, and promotion logic exist
Original equal-budget comparisons and field evidence
Packaging, audience routes, and living updates
Editorial architecture · audience and rollout
Designed Comprehensive and essay renderings plus audience decisions are mapped
Canonical public hub, route pages, citation/version surface, and change log
One comprehensive report
A connective narrative: goals → capabilities → techniques → human experience → conclusions → experiments. It owns continuity, definitions, and complete evidence accounting.
A non-linear essay series
Self-contained answers to specific questions. Each piece carries its own vocabulary, evidence, limits, and “where this fits” link back into the whole.
Missing bridges
The next useful work connects existing strength instead of adding another broad summary.
These are not empty topic slots. Each is a relationship the current package promises but cannot yet demonstrate or navigate cleanly.
Show when a measured gain changes an ordinary job.
Map capability, technique, work surface, accepted artifact, total cost, and reader consequence in one chain.
Give every mechanism an information and authority boundary.
Join intent, data contract, generation, execution, critique, delivery, and human gates in one reference anatomy.
Follow work through delivery, correction, maintenance, and retirement.
Current studies stop early; production value and harm appear after the first render.
Build evidence-bearing cases for journalism, BI, operations, science, and education.
Context changes purpose, authority, update cycle, audience, and the cost of failure.
Bring security, privacy, provenance, disclosure, and incident ownership into one chapter.
These constraints determine whether a technically capable workflow is usable at all.
Build the canonical hub, public evidence bundle, versions, and change log.
The current pieces are substantive sibling drafts. The missing surface must make them navigable and historically citable.
Boundary
This is a map of current ownership, not a publication claim.
The map was reconciled against the report files, existing visual companions, editorial architecture, audience strategy, and private evidence indexes as of 14 August 2026. A report node means a substantive draft exists. A visual marker means a local visual companion exists. Missing bridges remain planned work.