AIW compares ordinary LLM outputagainst governed County execution.
Model KB, document upload, KB-grounded prompt response, voice input, SEEKER Loop routing, human approval gates, RACI, PESTLE, and full audit trail — live in the browser.
Demo Rule
AI advises. Humans decide. Records remain.
County Priorities
Procurement Speed + Vendor Readiness
Show me how Prince George's County can reduce procurement cycle time, improve vendor readiness, and identify bottlenecks without bypassing required review.
Operational pain points
Scenario tags
The County can improve results by using AI tools, automation, dashboards, data analysis, staff training, and better coordination across departments.
What's missing
Decision ID
AIW-PG-20260907001731-6JRGW
Grounded by
Say "AIW" followed by a command. Voice commands are parsed server-side against a fixed intent table — never routed through a free-form model. Persona switches and approval-adjacent actions always require confirmation.
Example commands
No commands yet. Enable wake word or click an example above.
Naive-Bayes log-odds classifier scores every open PG County opportunity against the registered business profile. Evidence-cited posterior probability — no auto-award, no auto-exclusion. Every medium/high match is hitl_required=true.
Click "Run MBE Match" to score 5 active PG County opportunities against the Kemit Group profile.
JARVIS generates a 30/60/90 action plan with RACI ownership, HITL gates, SEEKER Loop trace, PESTLE analysis, and Zoho backend workflow — grounded in the selected KB documents. Select any provider below; falls back to a structured demo packet if unavailable.
Click "Run JARVIS" to generate a governed 30/60/90 execution packet for the active scenario.
Captured typed or voice prompt, persona, selected KB, and operating mode.
Mapped to Procurement, County priorities, and KB refs: PG-ANNUAL-REPORT-2024, PG-VENDOR-REG, PG-EXPO-OVERVIEW, AIW-SEEKER-LOOP-OS, NIST-AI-RMF-PLAYBOOK, AIW-ZOHO-BACKEND.
Assigned High risk with score 75/100. Human approval is required before execution, award, publication, or external release.
Compared operational pain points against procurement, resident service, workforce, economic, governance, and audit constraints.
Prepared governed recommendation, RACI, decision list, evidence register, and next-action packet.
Routed high-risk items to accountable human authority and preserved prompt/source/output record.
Captured missing data, unresolved assumptions, and lessons learned for pilot hardening.
Recommendations
Create one governed intake lane for each request: department, owner, deadline, risk, authority, and required approval.
Use KB-grounded answers only when a trusted source is selected; otherwise mark the output as assumption-based.
Score vendors and opportunities for readiness without making award decisions inside the AI layer.
Connect procurement categories to local workforce pathways, training partners, and employer demand.
Write every prompt, source, risk score, response, reviewer, and final human action to the audit trail.
Decision list
Approve demo/pilot scope
Owner: Procurement Director / Executive Office
HITL approval requiredAuthorize access to sanitized procurement, 311, and workforce data
Owner: Data owner + Privacy/Security
HITL approval requiredSelect first three department workflows
Owner: Executive sponsor
HITL approval requiredLoad official SOPs and public source documents into the KB
Owner: Program office
Demo actionRun Expo comparison: ordinary LLM vs AIW governed model
Owner: Demo lead
Demo actionOffice of Procurement / department buyer
Procurement Director / Executive Office
Economic Development, Compliance, Finance, department users, Employ PG
County leadership, vendors, local business partners
Aligns with County priorities, public trust, agency visibility, and accountable decision-making.
Supports local spend, vendor capacity, service continuity, and better use of public dollars.
Protects equity, resident trust, small business access, and practical community impact.
Uses governed AI, KB grounding, role controls, dashboards, workflow integration, and cloud/local/hybrid deployment.
Requires procurement, privacy, records retention, accessibility, labor, public communication, and compliance review.
Considers infrastructure resilience, transportation, energy, field operations, and degraded-network support.
Agencies, contacts, vendors, opportunities
P(sync) = 1.00Intake, scoring, approvals, HITL workflows
P(workflow) = 1.00RACI tasks, milestones, owners
P(tracked) = 1.00RFPs, contracts, briefs, audit packets
P(stored) = 1.00KPIs, dashboards, cycle time, spend visibility
P(visible) = 1.00311 queue, resident service, escalation
P(routed) = 1.00POST /ask
POST /analyze
POST /jarvis
POST /voice
POST /upload
POST /extract
GET /knowledge
POST /approve
POST /vendor-score
POST /zoho/sync
GET /audit/{decision_id}Bayesian audit baseline · Cohort 1 outcomes
Cohort 2 Curriculum — 6 Modules
All Zoho OAuth2 secrets are configured. Every Builder page intake submission now pushes a Lead record to Zoho CRM in real time via the refresh-token flow. Token is cached in-process and auto-refreshed before expiry.
P(lead captured | secrets present) = 1.00Every AIW URL is scored by the Bayesian router against the active use case. Posterior confidence = P(page relevant | active scenario + KB + persona). High-confidence pages are primary segues. All links are live.
AIW is a sovereign decision browser that turns problems into governed action.
Run the County challenge through ASK, ANALYZE, JARVIS, HITL, RACI, audit, and Zoho dashboards.
AIW is built around governed human decision support, not AI replacement of officials.
AIW sits above County systems as an integration, routing, governance, audit, and knowledge layer.
AIW routes County challenges into agency-specific workflows.
Each County role sees the same enterprise truth through the right decision lens.
AIW analysis depends on approved documents, workflows, policies, and data sources.
AIW turns demo findings into pilot milestones and annual plan outputs.
AIW maps every prompt to the right office, RACI owner, approval path, and knowledge collection.
AIW connects workforce displacement, AI literacy, youth/CTE, and vendor capacity building.
AIW can run as a standalone or edge decision server when network conditions degrade.
AIW keeps high-risk AI outputs advisory until human authority approves action.
User settings, saved sessions, approvals, knowledge packs, and dashboards live here.
Each agency has a Bayesian weight, default prompt, RACI, and workflow. The router scores agency fit against the active prompt and routes accordingly.
Enterprise risk, public trust, jobs, safety, budget, and 30/60/90 execution.
"Prepare my County Executive daily brief with procurement, workforce, public safety, 311, economic development, and AI governance risks."
RFP intake, vendor discovery, bid scoring, lifecycle transparency.
"Analyze this RFP and identify vendor fit, compliance gaps, local/minority vendor opportunity, and evaluation risks."
Fiscal controls, audit readiness, revenue, fund balance, ACFR-linked KPIs.
"Analyze the fiscal and audit impact of the AIW 90-day procurement modernization pilot."
Budget submissions, capital planning, fiscal impact, annual plan roll-up.
"Rank proposed AIW pilot actions by budget impact, operational value, and implementation risk."
AI governance, cybersecurity, integrations, identity, APIs, data classification.
"Evaluate AIW deployment risk across cloud, standalone, hybrid, and Starlink/edge modes."
Vacancies, AI training, federal workforce transition, redeployment, youth/CTE.
"Identify federal workforce displacement risks and map affected workers to County roles, vendor support roles, and AI training pathways."
Permits, inspections, project delays, applicant transparency, business impact.
"Which permitting delays are slowing business growth, housing, or redevelopment investment?"
Wait times, service requests, multilingual outreach, resident communication.
"Identify 311 categories with rising volume, long wait times, repeat complaints, and delayed resolution."
Police, Fire/EMS, Homeland Security, Courts, Corrections, crisis coordination.
"Show safety risks requiring coordination across police, fire, homeland security, courts, corrections, and community services."
Vulnerable populations, behavioral health, case coordination, referrals.
"Identify vulnerable residents needing coordinated health, behavioral health, housing, food, workforce, or social service support."
Use this script at the Expo. Each step maps to a mode, a prompt, and expected screen elements. P(demo success | script followed) = 0.94.
"AIW is a sovereign decision browser that helps County leaders move from problem to action with human approval, RACI ownership, and audit trails."
Prompt: "Show how Prince George's County can improve procurement speed, vendor discovery, and contract transparency while increasing local and minority business participation."
Prompt: "Run vendor readiness, procurement bottleneck, and workforce-economic sync scoring."
Prompt: "Generate a 30/60/90 implementation plan with RACI, HITL approval, and Zoho backend workflow."
Prompt: "Compare cloud, standalone sovereign workstation, hybrid, and Starlink/edge deployment modes."
"AIW does not replace County officials. It helps them see the issue, assign responsibility, approve action, and defend decisions with evidence."
Demo close line (use verbatim):
"AIW does not replace County officials. It helps them see the issue, assign responsibility, approve action, and defend decisions with evidence."
Live RSS aggregator across 7 channels. Filter by category, search by keyword, toggle grid or list view.
No stories match your current filters.
Sources: PG County Procurement · Employ PG · PG EDC · GovTech · FedScoop · Diversity Inc · Data Center Dynamics
Live RSS · 5-min cache · No API key required
End-to-end production and dissemination hub. Send reports via Zoho Mail, post to Cliq channels with video meeting links, create signature requests via Zoho Sign, manage invoices and estimates in Zoho Books, and dispatch to any platform simultaneously via the unified Share dispatcher.
Unified dispatcher — send any AIW OS report, analysis, or document to multiple platforms simultaneously. Select channels, add recipients, and optionally embed a video meeting link. All dispatches run in parallel.
Search the full AIW persona and use-case catalog. Static preview is always available. Sign in to run live governed LLM queries with Bayesian inference and citation-backed outputs.
Route prompts across OpenAI, Anthropic Claude, Perplexity Sonar, DeepSeek, and Local/Ollama with Bayesian confidence scoring. Auto mode tries providers in priority order. Persona selector adapts the system prompt. All inference is governed — human review required for consequential decisions.