> ## Documentation Index > Fetch the complete documentation index at: https://invoca-5bd45748-mintlify-17ed87db.mintlify.site/llms.txt > Use this file to discover all available pages before exploring further. # Identifiers: Iconography > Whether an agentic feature needs its own identity glyph, and why that glyph must not also mean 'generating' or 'uncertain.' **No visual identity has been decided for any Invoca agentic feature yet.** This page proposes a framework for making that decision — it is not the decision itself. See [Coverage, stated honestly](/invoca-design-system/ai-experience/overview#coverage-stated-honestly). ## What it governs Whether an AI feature uses a dedicated glyph to identify itself as AI — separate from any glyph used to show it is generating, or any glyph used to flag uncertainty. It governs only the **identity glyph**: the one that answers "is this AI," not the ones that answer "is it working" or "should I trust this." It does not govern [Foundations: Iconography](/invoca-design-system/foundations/iconography) itself — the closed 154-icon set, the four size tokens, and the color pairings every icon in the product already draws from, including any identity glyph this page eventually adopts. ## Vocabulary Three glyphs get confused for one because they can all sit in the same corner of the same card. This page owns exactly one of them. | Term | Answers | Owned by | | -------------------- | -------------------------------------------------------------- | ---------------------------------------------------------------------------- | | **Identity glyph** | "Is this AI?" | This page | | **State glyph** | "Is it working right now?" — a spinner, a generating indicator | [Governors](/invoca-design-system/ai-experience/governors/stream-of-thought) | | **Confidence glyph** | "Should I trust this?" — a caveat or warning mark | [Trust builders](/invoca-design-system/ai-experience/trust-builders/caveat) | ## Choosing a value | If you need to… | Use | Not this, because | | ------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Mark a surface as AI-made with a small, consistent icon | An icon from Titan's existing 154-icon set that fits, or a formal proposal to add one | An inline SVG or an ad hoc mark outside the closed set — `TITAN-ICO-01` already forbids this for every other icon in the product | | Show the AI is actively generating | An outcome-state treatment on the relevant pattern page, per [Governors](/invoca-design-system/ai-experience/governors/stream-of-thought) | Reusing the identity glyph to also mean "in progress." One glyph now carries two meanings, which `TITAN-ICO-07` already prohibits for every other icon | | Flag that an answer is uncertain | The [Trust builders](/invoca-design-system/ai-experience/trust-builders/caveat) Caveat pattern | Overloading the identity glyph for this too — the same one-meaning problem, with higher stakes, since a user who can't tell "AI" from "AI, and I'm not sure" is the confident-and-wrong case the whole section exists to prevent | | Render AI output before any glyph has been decided | A text label only | A placeholder icon — an icon reads as a decision the moment it renders, even an unofficial one | ## Reference Nothing is decided or shipped. No identity glyph exists for any Invoca agentic feature today. This section will be generated from a real decision once one exists. ## Relationships An identity glyph, if adopted, comes from [Foundations: Iconography](/invoca-design-system/foundations/iconography)'s closed set — `TITAN-ICO-01` — or from a formal proposal to extend it, exactly like any other icon in the product. It also inherits that page's color rules: an identity glyph is colored with an `icon-*` token, never a hex value or a text-color token, and if it carries a status-adjacent meaning it uses the `-alt` variant per `TITAN-ICO-04`. It pairs with [Identifiers: Color](/invoca-design-system/ai-experience/identifiers/color) — if a color treatment is decided, the glyph and the treatment are the icon-plus-color pairing `TITAN-COLOR-03` requires everywhere else — and with [Identifiers: Avatar](/invoca-design-system/ai-experience/identifiers/avatar), since a small identity glyph is one of the plausible shapes an AI mark could take. ## Accessibility * An icon-only identity mark has an accessible name stating what it is — "AI-generated," not a description of the glyph — the same requirement `TITAN-ICO-05` places on every icon-only control. * A decorative use of the glyph (appearing beside text that already discloses AI content) is hidden from assistive technology with `aria-hidden="true"`, per `TITAN-ICO-06`. * The glyph is never the sole carrier of the AI disclosure. Pair it with text, per `TITAN-ICO-08` and the overview's disclosure requirement. ## Constraints | ID | Constraint | Rationale | | ------------------ | ---------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **TITAN-AIICO-01** | An identity glyph, if used, comes from Titan's existing 154-icon set or a formal proposal to it — never an inline SVG or an ad hoc mark. | Mirrors `TITAN-ICO-01`. An icon outside the closed set has no size token, no color token, and no contrast guarantee. | | **TITAN-AIICO-02** | An identity glyph carries exactly one meaning — "this is AI" — and is never reused to also mean "generating" or "uncertain." | Mirrors `TITAN-ICO-07`. Those are separate glyphs owned by Governors and Trust builders respectively; reusing one glyph for two meanings makes both unreadable. | | **TITAN-AIICO-03** | An identity glyph is never the sole signal that content is AI-generated. Pair with text — the feature's stated name. | Matches `TITAN-ICO-08` and the overview's disclosure requirement — icon comprehension without a label is low, and worst for the users who need the disclosure most. | | **TITAN-AIICO-04** | An icon-only identity mark carries an accessible name stating what it is, not what it looks like. | Mirrors `TITAN-ICO-05` — a screen reader user needs the fact, not a shape description. | | **TITAN-AIICO-05** | Where no identity glyph has been decided, AI-touched surfaces render with a text label only, never a placeholder icon. | An icon reads as a decision the instant it renders, even when nobody intended it as one. | ## Divergences Not applicable. Nothing is shipped yet to diverge from. ## Gaps * Whether a glyph vocabulary is needed at all, given that Name and Color together might disclose AI content sufficiently on their own, is undecided. * Whether one identity glyph serves all of Invoca's agentic surfaces or each needs its own is the same per-feature-or-once question flagged on the [overview](/invoca-design-system/ai-experience/overview#gaps). ## Related [Identifiers: Avatar](/invoca-design-system/ai-experience/identifiers/avatar) · [Identifiers: Color](/invoca-design-system/ai-experience/identifiers/color) · [Identifiers: Name](/invoca-design-system/ai-experience/identifiers/name) · [Identifiers: Personality](/invoca-design-system/ai-experience/identifiers/personality) · [AI Experience overview](/invoca-design-system/ai-experience/overview) · [Foundations: Iconography](/invoca-design-system/foundations/iconography) ## Why it works this way Icons are learned, not read — `TITAN-ICO-07`'s one-meaning-per-icon rule exists because a glyph that means two different things on two different screens has to be relearned every time, which means it is carrying no information at all. That problem gets worse, not better, under an identity glyph, because the three candidate meanings — is this AI, is it working, should I trust it — are exactly the three questions a confused user is most likely to be asking at the same moment. Splitting them into three glyphs owned by three different categories is what keeps any one of them answerable at a glance.