01 / HITO · PRODUCT DESIGN & INTERACTION SYSTEM
AI proposes.
Humans
decide.
An evidence-aware AI workspace
for complex decisions.
A deadline is
not an approval.
Project note #A4Checked · 09:24
Mail discussionInconclusive · 09:25
HISTORICAL SNAPSHOT / HUMAN REVIEW REQUIREDMY ROLEProduct concept, information architecture, UX flows, visual system and interactive prototype.
DELIVERABLEFour connected working views and a reusable design system, explored through a fictional launch decision.
Approval still open.Checked 09:24 · historical
unconfirmed.Checked 09:25 · inconclusive
A deadline is not proof of approval.
Are we ready to launch?
Not yet.Design sign-off remains open.
Supplier timing is unconfirmed.
Design sign-off remains open.
Supplier timing is unconfirmed.
2 historical sources · Not sentFull brief remains visible behind this detail.
Start with the decision that needs attention.
01 / EXPERIENCE THE DECISION
A deadline isn’t
an approval.
Atlas Launch has a recorded sign-off deadline. But is the team ready to launch? Follow the evidence from workspace to decision.
A fictional 15 October snapshot. Scripted responses, local data, no connected services. Nothing is sent or saved outside this visit.
- 01 Open Atlas Launch
- 02 Ask the question
- 03 Inspect both sources
- 04 Review the decision
Start with Atlas Launch. Find the approval that still needs a human decision.
PROJECTS · CONVERSATIONS · SOURCES · DECISIONSThe answer is only useful
if you can judge it.
An AI assistant can turn incomplete information into a confident sentence. The design challenge is to make the missing certainty visible—without making everyday work slow.
Keep the decision legible.
Atlas Launch brings a project note, a Mail check and an open approval into one task. A recorded deadline tells us when approval is expected. It doesn’t tell us that approval happened.
Keep authority with the person.
The Agent can explain evidence and prepare wording. Read-only access cannot create a provider draft or send a message. Reviewing the text does not grant new permissions.
Independent portfolio design study. The scenario is fictional; no user research, adoption figures or measured business outcomes are claimed.
One decision.
Context that stays with it.
Projects hold the decision. Conversations explore it. Sources let the person inspect it. Artifacts retain the rationale.
- 01 / WORKSPACEFind what needs attention.Atlas sign-off is unresolved.↗
- 02 / PROJECTUnderstand the state.Deadline, owner, access and evidence.↗
- 03 / CONVERSATIONQuestion the evidence.Ask, inspect, preserve uncertainty.↗
- 04 / DECISION BRIEFReview the next step.Human review, without external sending.↗
Explain the checked material.
Flag unresolved claims.
Prepare text for review.
Whether the evidence is sufficient.
Whether the wording is appropriate.
What action to take outside Hito.
Show what is known.
Keep the unknown intact.
Deadline recorded.
Maya’s design sign-off is due Thursday at 14:00. The approval remains open.
09:24 · historical checkSchedule unconfirmed.
The checked discussion does not establish a revised supplier deadline or owner.
09:25 · historical checkSource checks have a time and a limit. A previous check is not a claim of current availability. The inspector exposes that distinction where the answer is read, rather than hiding it in a disclaimer elsewhere.
05 / DESIGN SYSTEM
Precision, repeated.
Dark navy surfaces, controlled violet, two product typefaces and compact data structures. Reusable components carry the same hierarchy from workspace to review.
Read-only information, linked sources and decisions that stay with the human.
#080E1E#101728#111B2E#29364D#A99CFFRhythm4 / 8 base · consistent 24px groups
Structure240px navigation · 72px header
Density64px data rows · 36px controls
Inspect the original Figma foundations
06 / BEYOND THE VERTICAL SLICE
One visual language.
Different working contexts.
Other completed Figma screens extend the same system. These are design artifacts, separate from the four functional demo views.
Useful intelligence.
Visible limits.
The system is deliberately restrained. Dense rows make work scannable. Persistent context avoids re-explaining the task. Evidence stays close to the answer. Human review stays separate from permission to act.
The goal is a credible decision workspace—not a demonstration of autonomous power. This interactive study makes one complete path tangible, while keeping the simulation and its boundaries explicit.