Signal Agency brand guide for usable report and content production. This specimen is a show-and-tell system for assets, colour, typography, badges, notifications, evidence, recommendations, worker handoff, and export QA.
Signal Agency content should feel editorial, evidence-led, squared, and decisive. Use warm paper, black rules, mono provenance, and one terracotta signal accent. Never round the core report components.
Use on covers, title pages, and formal client handoffs. Keep the mark on warm paper with strong clearspace and one terracotta rule.
Use only for intentional chapter openers or high-contrast presentation frames. Do not mix with soft gradients.
Use for proof points, source-led pages, and dossier dividers. It should feel like a stamp, not an app icon.
Use when pairing Signal Agency with a client or project name. Align on a rule; keep both marks squared and balanced.
Default page colour. It makes reports feel reviewed, printed, and evidence-led.
Primary text, rules, card headers, and table dividers. It creates hierarchy without decoration.
Use for decisive moments: cover accent, critical decision, or “read this first” marker.
Use for passed checks and protected patterns. Keep it subdued but legible.
Use for partial evidence, stale facts, and dependencies. It must be visibly different from green and red.
Use for blockers, missing citations, or privacy risk. Pair with a required action.
Client AI-search dossier
Large Bricolage-style title. Use for covers and chapter openings.
What answer engines can verify
Use for the question the section answers. Keep it concrete.
Use normal body text for evidence. Bold the conclusion. *Italic* marks caveats, assumptions, or interpretation.
Use mono for source IDs, owners, dates, engine names, and verification commands. Provenance must be easy to scan.
Use quotes for client voice, source excerpts, or reviewer observations. The quote should support a finding, not replace the finding.
Badges are small provenance tags. They sit beside claims, sources, and table cells. They are not notifications and should not occupy a whole page alone.
Evidence: Evidence:Verified Evidence:Partial Evidence:Inferred Evidence:Missing
Priority: Critical High Medium Low
| Badge | Purpose | Use example | Signal Agency rule |
|---|---|---|---|
| Evidence:Verified | Directly observed evidence | “AI Overview cites the comparison page” | Attach source ID |
| Evidence:Partial | Mixed or incomplete support | “Gemini sees pricing but misses warranty” | Explain the gap |
| Evidence:Inferred | Judgement from pattern | “Likely schema/entity mismatch” | Keep caveat visible |
| Evidence:Missing | No evidence captured | “No source card for claim” | Convert to an action |
Purpose: stop the reader and require action. Use for privacy exposure, missing source IDs, or evidence that invalidates the recommendation.
Example: “Raw prompt transcript appears in the public export. Replace with a redacted source summary.”
Purpose: highlight risk that affects confidence. Use for stale citations, partial retrieval, or owner ambiguity.
Example: “Two engines cite the old service name; update corroborating profiles before the next crawl.”
Purpose: explain how evidence was collected or scoped.
Example: “Prompts were run from a clean browser profile and compared against first-fetch HTML.”
Purpose: mark a verified pattern to keep.
Example: “Source IDs now appear beside every factual recommendation.”
3/5
Engines with at least partial visibility. Source: C001.
27
Evidence references captured. Source: ledger.
6
Roadmap items sized. Source: priorities.
0
Private URLs in public export. Source: privacy check.
| Component | Purpose | Signal Agency treatment | Example content |
|---|---|---|---|
| Manifest | Scope and evidence rules | Square field block with ink rule | “Raw transcripts stored securely” |
| KPI card | Executive metric | Huge numeral, short source line | “3/5 engines visible” |
| Source ledger | Claim traceability | One width, dotted row rules | C001 — prompt capture batch |
| Brief card | Worker handoff | Light ruled panel, mono fields | Task / files / acceptance / verify |
C001 — Prompt capture batch — High confidence; raw transcripts stored securely.
C002 — Rendered crawl — High confidence; confirms first-fetch visibility.
C003 — Analytics export — Medium confidence; prioritises commercial pages.
C004 — Parity review — Medium confidence; checks third-party fact drift.
Use when one finding needs executive visibility, owner, due date, source IDs, and verification.
Owner: Editorial. Due: 2026-W23. Verify: C001 and C002 show first-fetch retrieval.
Use when facts exist but are too far from the claim, table, or source card.
Owner: Content. Due: 2026-W25. Verify: source IDs appear beside each claim.
Use for shipped work that should be protected in the next iteration.
Owner: Engineering. Verified: C001. Preserve: no private URL appears in public PDFs.
Task: Move critical comparison facts into crawlable HTML and attach source IDs.
Files: comparison template, source-card component, pricing facts module.
Acceptance: direct answer, source IDs, updated date, and criteria table appear in first-fetch HTML.
Verification: rerun per-engine prompt set separately and compare citations.
Do: put source IDs beside factual claims.
Do not: publish raw transcripts, local paths, or private URLs in public reports.Public artifact rule
Signal Agency public examples must use placeholders only. Raw transcripts, screenshots, private URLs, client names, and local paths stay in approved secure storage.
Signal Agency brand style guide specimen · usable production guide · public-safe placeholder content