Identifiers: Name

No verbal identity has been decided for any Invoca agentic feature except Signal AI. This page proposes a framework for making that decision for the rest — it is not the decision itself. See Coverage, stated honestly.

What it governs

Whether an agentic feature has a name distinct from Invoca’s own product name, and whether that feature speaks in first person — “I found three calls that match” versus “3 calls found.” It governs the decision to name a feature and the decision to give it a voice; it does not govern what the feature actually does (that is Actions) or how it sounds once it speaks (that is Identifiers: Personality).

Vocabulary

Three different names get conflated into one word, and the conflation is where naming decisions go wrong:

Term	Means	Example
The feature’s name	The name a user sees attached to one specific capability	”Signal AI”
The model’s name	The underlying model or vendor powering the feature	Never named — see Constraints
The product’s brand name	The company or product brand the user already trusts	”Invoca”

A separate, unrelated distinction: speaking in first person (“I found…”) is a grammatical choice about voice, not a personality trait. Whether the feature has a persona and how that persona sounds is Identifiers: Personality’s question; this page only decides whether “I” is on the table at all.

Choosing a value

If you need to…	Use	Not this, because
Let users know which feature they’re using	A stated name for that feature, decided deliberately	Leaving it unnamed and letting “Invoca” absorb both the credit and the blame for one feature’s output
Decide whether the AI speaks as “I”	A deliberate, stated decision recorded on that feature’s page	Defaulting to first person because sample or prototype copy happened to use it — per the overview’s `TITAN-AI-04`, disclosure must be explicit either way, and an undecided default is not a decision
Reference what powers a feature	Nothing — the underlying model or vendor is never named to the user or in documentation	Naming the model or vendor, in the UI or in these docs — see Constraints
Name a brand-new agentic surface that has no name yet	No name, until one is deliberately decided	Inventing a placeholder name in documentation, which the next reader will mistake for a real decision

Reference

Nothing is decided for any feature except one: Signal AI is shipped and named, and its transcript-summary action is documented at Actions: Summarize. Whether Signal AI speaks in first person, and whether its name extends to any future assistant or panel, is not recorded anywhere. Every other feature in this section has no stated name at all. This section will be generated from real decisions as they are made.

Relationships

A feature’s name and its speaking register are usually decided together — see Identifiers: Personality for the tone question once a name exists. Named, first-person copy still follows Content: Voice and tone’s three brand traits — Empowering, Eye-opening, Bold — and its plain-language constraints (TITAN-VOICE-01–03); naming a feature does not exempt its copy from the site’s general voice rules. Deciding a name is also frequently the trigger for deciding Identifiers: Avatar and Identifiers: Color for the same feature, since a name without a visual mark is only half an identity.

Accessibility

A first-person voice does not remove the disclosure-at-point-of-output requirement from the overview — a name and “I” together must still make it unambiguous that a system, not a person, is speaking.
A name alone is not sufficient disclosure if it is presented ambiguously — a name that reads as a person’s name (rather than a feature’s) works against the very disclosure this page exists to support.

Constraints

ID	Constraint	Rationale
TITAN-AINAME-01	A feature is named only by deliberate decision — never inferred from copy that happens to use a name informally.	An informal nickname repeated enough in prototypes or support docs becomes a de facto name nobody signed off on.
TITAN-AINAME-02	The underlying model or vendor is never named to the user or in documentation.	Matches the sitewide rule against naming the underlying component library, applied to an actual AI vendor — where the stakes include vendor lock-in perception and competitive exposure, not only editorial cleanliness.
TITAN-AINAME-03	Whether a feature speaks in first person is a stated decision on that feature’s page, not a default.	First person reads as a claim of agency. Per the overview’s `TITAN-AI-04`, the choice changes how easily a user can tell they are talking to a system rather than a person.
TITAN-AINAME-04	A feature’s name never implies a capability broader than the job it does.	Overpromising in the name undercuts the confident-and-wrong problem this whole section exists to design for — a name suggesting more scope invites more trust than the feature has earned.
TITAN-AINAME-05	Where a feature has no stated name, its output is attributed to Invoca generally, never to an invented placeholder name.	An undecided name should read as undecided. A placeholder in documentation is indistinguishable from a real decision to the next reader.

Divergences

Not applicable. Nothing is shipped yet to diverge from, except Signal AI’s existing name, which carries no recorded divergence.

Gaps

Whether Identifiers apply per-feature or once for all of Invoca’s agentic surfaces is undecided — this exact question is flagged on the overview, which points to this page for it.
Whether Signal AI’s name extends to a future assistant or copilot panel, or whether each gets its own, is undecided.
Whether any Invoca agentic surface speaks in first person is undecided for every feature, Signal AI included.

Identifiers: Avatar · Identifiers: Color · Identifiers: Iconography · Identifiers: Personality · AI Experience overview · Content: Voice and tone

Why it works this way

A name is the cheapest identity decision available and the easiest to make by accident — copy in a spec document, a Slack thread, or a prototype settles into a nickname long before anyone decides it should be one. Treating naming as deliberate, with its own constraint against silent adoption, is what keeps a placeholder from calcifying into a brand. Separating “does it have a name” from “does it say I” matters because they fail in opposite directions. An unnamed feature under-discloses — a user cannot tell which capability they are using or hold accountable. A named feature that also speaks in first person without deciding to can over-disclose the wrong thing: it can read as a person, which is a bigger claim than Invoca has made about what the feature is.

Overview

Wayfinders

Inputs

Tuners

Governors

Trust builders

Identifiers

Actions

Identifiers: Name

What it governs

Vocabulary

Choosing a value

Reference

Relationships

Accessibility

Constraints

Divergences

Gaps

Why it works this way

​What it governs

​Vocabulary

​Choosing a value

​Reference

​Relationships

​Accessibility

​Constraints

​Divergences

​Gaps

​Related

​Why it works this way

What it governs

Vocabulary

Choosing a value

Reference

Relationships

Accessibility

Constraints

Divergences

Gaps

Related

Why it works this way