AI Experience overview

Exemplar section — first pass, entirely proposal, with one narrow exception. No pattern in this section is built as an experience, which means none of them carry code fact or evaluation data — every statement here is design intent, offered for review, not established policy. The one exception is Signal AI’s transcript-summary capability, which is shipped and named — but even there, only its external API contract is verified; the experience around it (disclosure, outcome states, correction) is proposal like everything else. See Coverage, stated honestly.

What this section governs

Anything in Invoca’s product where the output is generated by a model rather than rendered from state, or where a feature acts on the user’s behalf rather than only responding to input. A Signal AI call summary, a natural-language search over call data, an assistant that drafts a follow-up or updates a record, a panel that recommends a routing or budget change — all of it lives here, not in Patterns. It does not govern the visual materials an AI surface is built from. A summary still renders in a Card; a streaming answer still uses Foundations motion and color tokens. This section governs what is different about the experience once the thing filling those components is not deterministic.

Why this isn’t a Pattern

The rest of this design language resolves top-down through four tiers — Foundations, Components, Patterns, Views — and each tier is defined by a test: is it a token, a package export, a composition with no export, or a page archetype. An AI experience fails all four tests the way a Pattern would pass them, because four assumptions those tiers depend on do not hold here:

A Pattern can assume	An AI experience cannot
The same input produces the same output	Output varies, and sometimes it is wrong
Every value is derivable from source	A behavior like “cites its sources” has no emittable value
The failure case is one state among several	The failure case is the design surface
Documented behavior stays true	Model capability moves underneath the page

So an AI experience page carries every section a Pattern page carries — constraints, content rules, accessibility, gaps — plus four sections nothing else in this design language has: agency tier, outcome states, disclosure & recourse, and evaluation. Skipping straight to Components or Patterns for an AI feature means improvising exactly those four.

Vocabulary

Seven categories, adopted from the public pattern catalogue Shape of AI. The names are taken deliberately: two of the seven — Governors and Trust builders — exist purely for oversight and correction, which is the right weighting for a house system and worth keeping as emphasis rather than reinventing.

Category	Answers	First page
Wayfinders	How does the user find or start using the AI feature at all?	Initial CTA
Inputs	How does the user tell the AI what they want?	Open input
Tuners	How does the user adjust what the AI is doing before or during a run?	Controls
Governors	How does the user see and check what the AI is doing?	Stream of thought
Trust builders	How does the interface admit what it doesn’t know?	Caveat
Identifiers	How does the AI present itself — visually and verbally?	Avatar
Actions	What is the AI actually being asked to do?	Search

Actions is Invoca’s own category, not Shape of AI’s. Shape of AI catalogues the surrounding chrome — how you prompt, watch, and correct a model — and deliberately stops short of naming the jobs themselves. Invoca’s agentic surfaces already do specific jobs today — Signal AI summarizing a call transcript is shipped, named, and real — so this section names the seven jobs its agentic features are known or expected to do: search, summarize, draft, update, disambiguate, recommend, and the welcome/empty state every one of them needs on day one.

Identifiers reads differently from the rest of this section. Avatar, Color, Iconography, Name, and Personality describe how the AI presents itself, not an interaction — there is no run to interrupt and no output to be confidently wrong about. Those five pages use the lighter foundation shape (what it governs, choosing a value, constraints) instead of the full pattern shape below. Everything else in this section — Wayfinders, Inputs, Tuners, Governors, Trust builders, Actions — uses the full shape.

The four sections nothing else in this design language has

Agency tier

Every interactive page in this section (everything except Identifiers) states, plainly, how much authority the feature has. Naming the tier is what makes the confirmation-versus-undo question askable at all.

Tier	Authority	Requirement
Suggests	Proposes; the user acts	Refusable without friction
Drafts	Produces reviewable output	Obviously a draft; never auto-submits
Acts, reversible	Does it; the user can undo	Undo is real, discoverable, bounded in time
Acts, irreversible	Cannot be undone	Explicit confirmation naming the specific consequence
Acts unsupervised	No per-action review	An audit trail, a kill switch, a stated blast radius

Escalating a feature’s tier is a decision, never a default. A feature that starts suggesting and ends up acting unsupervised has crossed a line that needs recording, the same way a token change does — see Constraints.

Outcome states

A component has interaction states — hover, pressed, disabled. An AI experience has outcome states, and every pattern page in this section carries a treatment for each one that applies: working, streaming/partial, confident and right, confident and wrong, uncertain, refused, empty, interrupted, degraded, rate-limited, stale. Confident and wrong is the primary case to design for, not the edge case. A wrong answer delivered with no visible seam is the characteristic failure of this whole section, and every pattern page states how a user notices and corrects one. “The model is usually right” does not count as an answer.

Disclosure & recourse

Six questions, on every interactive page, answered or explicitly marked not applicable:

Does the user know this is AI, at the moment it matters?
What did it use — call data, prior conversation, other records?
How sure is it, and does that change what the user should do?
How does the user check it?
How does the user correct it, and does the correction persist?
How does the user get out — reach a person, do it manually, turn it off?

Question 6 is the one most often missing and the one a user wants most when the feature has already failed them once.

Evaluation

Nothing in this section is built, so nothing in this section has been evaluated. Every page says exactly that rather than implying otherwise. Once a pattern ships, its page states the eval set, the metric and threshold, named failure classes from real output, and the date of the last run — a capability claim with no evidence behind it is the failure this section exists to prevent.

Coverage, stated honestly

Category	Decided for Invoca?
Wayfinders	No — proposal only, no shipped surface
Inputs	No — proposal only, no shipped surface
Tuners	No — proposal only, no shipped surface
Governors	No — proposal only, no shipped surface
Trust builders	No — proposal only, no shipped surface
Identifiers	No — proposal only; Invoca’s agentic features have no stated visual or verbal identity yet
Actions	Partly — Signal AI’s transcript summary is shipped and named; search, draft, update, disambiguate, recommend, and the welcome/empty state are proposals with no shipped surface

Unlike every other tier in this design language, there is no generated reference layer here and there will not be one until something ships. A page in this section that looks thinner than a Pattern page is not missing content — it is missing a product to observe.

Choosing where to look

Wayfinders

Getting a user to try the feature, and to keep going once they’re in it.

Inputs

Letting the user say what they want in their own words.

Tuners

Adjusting scope, tone, or depth before or during a run.

Governors

Showing and checking what the AI is doing.

Trust builders

Admitting what the system doesn’t know.

Identifiers

How the AI looks, sounds, and refers to itself.

Actions

The seven jobs Invoca’s agentic features are known or expected to do.

Constraints

ID	Constraint	Rationale
TITAN-AI-01	Every interactive page in this section states its agency tier.	Confirmation, undo, and audit requirements all follow from the tier. Without it, each feature improvises its own answer.
TITAN-AI-02	Every interactive page names how a user notices a confident-but-wrong output.	The failure with no visible seam is the one that costs the most, and it is the one silence hides best.
TITAN-AI-03	A confidence signal ships only if it changes what the user should do.	A number that reads the same at 82% and 94% is decoration wearing the shape of information.
TITAN-AI-04	The user learns this is AI at the point of the output, not only in settings or a footer.	Disclosure that requires the user to already suspect something is not disclosure.
TITAN-AI-05	A capability claim on any page in this section names how it was established, or states plainly that it hasn’t been.	”It works well” with no eval set behind it is a hope, not documentation, and it is indistinguishable from a claim that was checked.

Gaps

No model, prompt, or vendor has been named for any pattern in this section except Signal AI’s existing transcript summary. Every Reference section below is a placeholder for content that does not exist yet.
No evaluation framework exists. There is no eval set, no scoring rubric, and no named owner for deciding when an AI feature is good enough to ship.
The escalation path from Suggests to Acts is undecided. Nothing says who approves a feature moving up a tier, or what evidence that decision requires.
Whether Identifiers apply per-feature or once for all of Invoca’s agentic surfaces is undecided. Signal AI, a future assistant, and a future copilot panel could each need their own name and avatar, or could share one identity. See Identifiers: Name.

Patterns overview, for the compositions this section’s outputs still render inside of. Foundations: Accessibility, for the criteria every page here inherits before adding its own. Foundations: Motion, for streaming and loading treatments.

Why it works this way

Borrowing the vocabulary and rejecting the shape. Shape of AI’s own page format is description, design considerations, related patterns, examples — no constraints, no “choose something else when,” no stated risk, no gap. That is the right shape for a public survey of what the industry does, and the wrong one for a house system, where the reader needs to know what Invoca requires and what it must not do. This section keeps the names and adds the parts a public catalogue has no reason to carry. Oversight is not a chapter here — it is two of seven categories. Governors and Trust builders exist to let a user check and correct the system, and giving them equal billing with Wayfinders and Inputs is a stated priority, not an accident of alphabetizing a list.

​What this section governs

​Why this isn’t a Pattern

​Vocabulary

​The four sections nothing else in this design language has

​Agency tier

​Outcome states

​Disclosure & recourse

​Evaluation

​Coverage, stated honestly

​Choosing where to look