Setup review
Before any content is built, an agent reviews the Canvas sub-account and course shell against the programme’s requirements, so problems are caught while they are still cheap to fix.
- Theme
- FCA_DA CSS v2.3 · JS v2.3
- Course navigation
- 11 items visible
- Grading scheme
- FCA standard (pass / not yet)
- Term dates
- 2 Nov 2026 – 29 May 2027
- Module locks
- Not set
- Template pages
- 6 pages imported
The setup-review agent reads the live configuration and the template pages, and compares both against the programme’s documented standards: the approved theme version, the six-item navigation, the grading scheme, and the markup the template pages should use.
It changes nothing — the output is a findings list, ordered by how expensive each issue becomes if it reaches the build stage.
- Theme current. CSS and JS both v2.3 — matches the programme standard.
- Grading scheme matches. FCA standard scheme applied at course level.
- Two heading styles in use. Three template pages carry manually-styled H2s instead of the themed class — these would propagate into every built page.
- Navigation shows 11 items; standard is 6. Files, Outcomes and three tool links are visible to learners.
- Module locks not set. All modules currently open; the programme releases Module 1 only at launch.
Canvas build
An agent converts the approved storyboard into the programme’s Canvas HTML — pages, assessments and discussions in the right module — behind a dry run, with a reload to verify every page it builds.
3.4 Choosing the right chart
H2: Match the chart to the question
Body: Every chart answers a kind of question. Comparisons want bars; change over time wants a line; parts of a whole want care…
Activity: Before moving on, sketch the chart you would choose for last week’s sales data, and note the question it answers.
Image: chart-decision-tree.png — alt: “Decision tree from question type to chart type”
Download: Chart checklist (PDF)
Match the chart to the questionH2 → themed heading
Every chart answers a kind of question. Comparisons want bars; change over time wants a line; parts of a whole want care…
- Dry run first. The agent lists every page, assessment and discussion it will create — nothing is written until the plan is approved.
- Copy is verbatim. Authored content is checked byte-for-byte against the storyboard; the build never paraphrases.
- Verified by reload. Each built page is reloaded and read back to confirm the markup survived Canvas’s editor.
- Queried, not guessed. The storyboard’s glossary note (“term modal: categorical data”) had no approved definition — raised with the author instead of invented.
Video placement
Course videos live in Wistia and are listed in a media register kept separately from the storyboards. An agent matches each row to its video, retrieves the embed code and transcript, and places both into the correct page markup.
| Page | Video title | Host ID | Status |
|---|---|---|---|
| 3.2 | Why charts mislead | fca_9d31k | Placed |
| 3.4 | Choosing the right chart | fca_4f8k2 | Approved |
| 3.6 | Dashboards that work | fca_7m20q | In edit |
The agent works through rows marked Approved, and leaves the rest.
- Right video, right slot. The host ID in the register matches the embedded player, and the page matches the register row.
- Transcript present. Every placed video carries its transcript — a video without one is left unplaced and flagged.
- Title mismatch flagged. Row 3.6’s register title differed from the video’s title in the host — held back for a person to resolve.
Feedback actions
QA feedback arrives as tickets linked to course content. An agent gathers the outstanding tickets, checks each against the approved storyboard, and separates actionable corrections from items needing clarification — then applies approved fixes behind a dry run.
Caption refers to “Figure 2” but the image above it is Figure 3.
Page 3.4 · Choosing the right chart · below the decision-tree image
The dry run shows the current page content and the proposed change, side by side, before anything is written:
Figure 2: Decision tree from question type to chart type
Figure 3: Decision tree from question type to chart type
- Verified after writing. The page is re-read, the publication status is confirmed unchanged, and ticket #214 is returned as actioned, ready to move to review.
- Held for clarification. Ticket #219 asked for a stat the storyboard doesn’t contain — returned to QA with a note instead of applied.
Build review
Before anything is published, an agent compares every completed page with the approved storyboard and the build standards: content accuracy and order, components and media, accessibility details, navigation, completion requirements, and links — including whether internal links survive course cloning.
3.4 Choosing the right chart
H2 · body · activity · image + alt · download · video + transcript
Banner · themed H2 · body · activity well · image with alt · download card · video with transcript
- Content matches, in order. Every storyboard element is present, verbatim, in sequence.
- Accessibility details present. Alt text on the image, transcript with the video, one H1 per page.
- Clone-safe links. Internal links use course-relative paths, so they stay valid when the course is cloned for the next presentation.
- Completion requirement missing. Page 3.4 has no “mark as done” requirement; the module’s standard requires one.
- External link redirects. The chart-checklist source link returns a redirect — update to the final URL before publishing.
The review ends as a report, not a publish action. The two flagged items above are investigated, fixed through the feedback workflow, and re-checked — then, and only then, a person publishes the module. The time the agents recover is spent here: deeper checks on content, accessibility, navigation and functionality than a compressed schedule would ever have allowed.