AI Tooling Dashboard
The platform lens on AI usage for an engineering manager: which tools the team uses, how heavily, what they cost, and how that maps to output. One of two dashboards carved out of a single meeting.
- Jun 22 intake meeting
- 138 traceability tags
- 7 sections
Origin: one meeting, two dashboards
A "Reports" meeting on June 22, 2026 (the CEO, Trent, the Head of Product, and the Dev Lead) was pruning an executive-facing AI adoption report section by section. Two things fell out that were too granular for executives but exactly right for a line manager, and each became its own stream: this one took the per-tool cut, its sibling (AI Usage) took the per-developercut. The core question, in the CEO's words: "generally, what tools am I using? how much am I using them?"
Data grounding
The binding constraint: stored activity collapses to a daily grain, so the hour-by-hour calendar heatmap all three stakeholders endorsed cannot be served. The data-relevance doc proposes the honest partial: a day-of-week pattern ("Fridays are quiet, Tuesdays are peak") with the hour axis red-bracketed. Second gap: per-tool valueoutput isn't exposed by any aggregation, so the platform comparison degrades, honestly, to cost and usage.
I don't have to convert or figure out between Mexico and central and eastern and Poland.
Design decisions
A distinctive lens, not a duplicate.Where objectives overlap with sibling surfaces, this dashboard shows the per-tool cut and cross-links to the per-developer cut. It never duplicates a sibling's panels.
Tools are visually trackable.Every tool gets an identity color held constant across every chart on the page, straight from the CEO's ask that platforms be comparable at a glance.
Shared vocabulary.The meeting's agreed work-model nomenclature (human only / human with AI / agentic) is adopted verbatim rather than inventing labels.
Self-review by annotation
This stream skipped the persona panel; instead the spec carries its own review layer: every claim tagged to its source, plus thirteen "manager takeaway" strips, two of them self-flagged as weak with a written fix. A coverage table maps every objective and transcript ask to where on the page it is answered.
138
Traceability tags
47
From the CEO
24
Data-relevance findings
13
Takeaway strips