YYUKU
YUKU AI Center architecture

AI-ready workflows for field operations, time, safety, estimating, reports, and marketplace products.

The AI Center models future modules, guardrails, workflows, provider boundaries, and marketplace readiness. No AI provider, model, API key, prompt runtime, or backend workflow is connected.

AI modules

11

Provider adapters

7

Connected providers

0

Vision

Photo Analysis

Future vision workflow for reading site photos, identifying progress signals, and surfacing missing context.

Inputs

Site photos, Project context, Task metadata

Outputs

Detected conditions, Progress hints, Follow-up questions

Language

Daily Report Generator

Future report drafting from tasks, notes, photos, weather, crew activity, and time entries.

Inputs

Field notes, Tasks, Photos, Weather, Time entries

Outputs

Draft daily report, Missing-data warnings, Review checklist

Planning

Task Suggestions

Future task recommendations from project stage, incomplete work, site observations, and prior templates.

Inputs

Project template, Milestones, Open tasks, Notes

Outputs

Suggested tasks, Priority hints, Crew assignment prompts

Planning

Schedule Optimization

Future scheduling assistant for sequencing work, crew availability, site constraints, and weather conditions.

Inputs

Milestones, Worker availability, Weather, Dependencies

Outputs

Schedule options, Conflict warnings, Crew load balance

Estimation

Material Estimation

Future material estimation from project details, takeoff inputs, calculators, and marketplace templates.

Inputs

Measurements, Drawings placeholder, Calculator outputs, Scope notes

Outputs

Material estimate, Assumptions, Variance warnings

Safety

Risk Detection

Future risk signals for delays, missing reports, safety patterns, incomplete photos, and schedule conflicts.

Inputs

Activity feed, Reports, Time entries, Safety forms

Outputs

Risk flags, Recommended review, Escalation prompts

Voice

Voice Notes

Future voice capture for field notes, task updates, report comments, and hands-free documentation.

Inputs

Audio note, Project context, Worker profile

Outputs

Transcript, Suggested note, Action items

Language

Report Summaries

Future executive and project summaries across daily reports, activity, labor time, photos, and issues.

Inputs

Daily reports, Activity timeline, Timesheets, Project metrics

Outputs

Summary, Highlights, Open issues

Safety

Safety Assistant

Future assistant for safety forms, toolbox meetings, site risks, weather warnings, and compliance prompts.

Inputs

Safety forms, Site conditions, Weather, Training content

Outputs

Safety prompts, Checklist suggestions, Training links

Vision

Equipment Analysis

Future equipment insight from photos, usage notes, maintenance records, and project activity.

Inputs

Equipment photos, Maintenance notes, Usage history

Outputs

Condition notes, Maintenance prompts, Utilization hints

Marketplace

Future AI Marketplace

Future marketplace category for AI workflows, assistants, templates, estimators, plugins, and subscriptions.

Inputs

Product metadata, Licenses, Compatibility, Subscription tier

Outputs

AI product listing, Access rules, Install placeholder

Future AI Workflows

Workflow placeholders show how modules can connect across Field, Time, Marketplace, Academy, and future Estimating.

Field day to reviewed report

Photos, notes, tasks, weather, and time entries become a reviewable report draft.

photo-analysisvoice-notesdaily-report-generatorreport-summaries
draft-placeholder

Risk signal to safety review

Risk detection can point to safety assistant prompts, checklist recommendations, and training placeholders.

risk-detectionsafety-assistantschedule-optimization
future-placeholder

Estimate to marketplace tooling

Material estimation can connect future calculators, takeoff tools, templates, and AI marketplace products.

material-estimationtask-suggestionsai-marketplace
future-placeholder

Provider-neutral safety boundary

No provider SDKs, API keys, model names, prompts, embeddings, or background jobs are connected.

Future modules should require explicit product permissions, source traceability, human review, and audit logs.

Safety, estimating, and risk features need policy controls before production use.

Governance Placeholders

These items define the controls needed before live AI features exist.

Human review

AI output should be treated as a draft or recommendation until a permitted user confirms it.

productavailable-placeholder

Source traceability

Reports, summaries, and risk flags should link back to source notes, photos, time entries, or forms.

operationsdraft-placeholder

Data boundaries

Future provider integrations must define retention, training use, redaction, region, and tenant isolation.

securityfuture-placeholder

Safety and legal review

Safety, estimating, and risk modules need disclaimers, policy controls, and review workflows before launch.

legalfuture-placeholder

AI Provider Adapter Placeholders

Provider connections are intentionally absent. These boundaries keep future AI integrations replaceable and auditable.

not-connected

Model router

Future provider-neutral routing for vision, language, planning, voice, and estimation models.

not-connected

Vision provider

Future image analysis boundary for site photos, equipment photos, and evidence review.

not-connected

Language provider

Future text generation, summarization, task suggestions, and report drafting boundary.

not-connected

Voice provider

Future transcription and voice note processing boundary with privacy controls.

not-connected

Embeddings and retrieval

Future search, knowledge retrieval, documentation, and project context boundary.

not-connected

Evaluation pipeline

Future quality checks, regression tests, prompt versioning, and output scoring.

not-connected

AI audit log

Future source, prompt, output, approval, edit, and provider event tracking.

Integration rule

Future AI work should connect through adapters only. Product UI should never call providers directly or store secrets in frontend code.