Glossary

The smallest reusable unit of work: a single thing done well, like 'research keywords', 'score AEO readiness', or 'transcreate to es-MX'. Skills do not run on their own; agents bundle them.

A role with a model and context access that bundles skills and can run autonomously or be chatted with (the SEO Agent, the Refresh Agent, the Localisation Agent). One agent is one doer. It runs its own skills in sequence internally, but it does not know what comes before or after it, and it cannot review or publish its own work.

A chain of steps that ties multiple actors together (several agents, human reviews, and a publish step) with handoffs between them. It coordinates who does what, in what order, with which checks. An agent is one step inside a workflow; the workflow begins where work is created and runs until the result is safely live.

The same process as a workflow, but laid out as a batch: steps become columns, deliverables become rows. One row per item (for example 1,847 legacy posts), and you run a column across every row at once. 'Run as Grid' on a workflow opens this view.

A single piece of content (a blog post, a localised page, a brief). It is one row in a grid and the thing that moves through a workflow and is delivered to the client. Its performance is tracked over time.

A step where a vetted Contentoo expert reviews and approves agent output, often a native speaker for localisation. This is the guard against 'AI slop': nothing publishes without it.

The automated quality band inside a workflow: the QA Agent checks facts, hallucinations, and structure, and the Brand Consistency Agent enforces voice and glossary. Both run before a human review, so experts only see work that already passed the basics.

The board that shows every deliverable in flight across stages (Brief, Drafting, Expert review, QA, Awaiting approval, Delivered). A workflow defines the steps; the pipeline shows where each piece currently sits.

A trigger the Signal Agent detects: the sales-call language mix (Gong), search decay (Search Console), or a product release (Jira). A signal starts a workflow, but it only becomes content after it passes the strategy filter in the Context hub.

The knowledge layer that grounds every agent: brand voice, audiences, knowledge base, style guides, and strategy. It is not decoration: the Brand Consistency Agent enforces it, and the Signal Agent uses the strategy in it to filter what becomes content.

Answer Engine Optimization: how visible and citable your content is inside AI answers, versus classic search rankings. Tracked as AI share of voice on a 0 to 100 scale (for example, climbing from 25 to 38).

One client's environment. Switching workspace re-themes every surface (agents, grids, workflows, dashboards, team) to that client's data, so the whole product can be viewed as that client sees it.

The combined roster of AI agents and human experts, scored on the same dimensions (throughput, turnaround, quality, cost, performance lift) so you can compare them and decide who to assign. Agents carry volume and speed; humans carry nuance, markets, and top quality.