Practical AI & Automated Workflows for Growing Businesses
We help boutique firms, regional operators, and mid-market teams deploy custom AI agents, automated workflows, and secure internal knowledge bases—cutting operational overhead without adding headcount.
- 60%+ Administrative Cuts
- CIO-Level Tech Oversight
- 100% Secure Data Practices
Led by Proven AI Architecture, Not Generalist Consultants
Joseph Van Harken is the Founder and Principal at MindRiot Labs, blending direct technical architecture with IT systems strategy. In addition to guiding private clients, Joseph serves as the IT Innovation & Research Innovator-in-Residence at Grand Valley State University (GVSU).
At GVSU, Joseph spearheaded the architecture for ProfAI—an advanced AI-powered educational system that was honored with the prestigious 2026 National CIO 100 Award for breakthrough innovation.
Through MindRiot Labs, Joseph operates a direct principal partnership model. Rather than working with distant agency account managers, you partner directly with Joseph acting as your Fractional Chief AI Officer, establishing data governance frameworks and secure workflows. Execution is completed rapidly via his node of elite specialized developers.
The Principal Advantage
We do not produce theoretical slide decks. We write production-grade code, configure secure pipelines in your cloud environment, and train your staff on custom tools.
Joseph Van Harken
CIO 100 Winner • Innovator-in-Residence at GVSU • Systems Developer
Productized SMB AI Offer
Flat-rate integrations designed to deliver immediate utility. Zero scope creep, zero billing surprises.
The 2-Week SMB AI Sprint
A high-velocity, structured implementation to isolate your highest-overhead operational bottleneck, build a secure prototype, and deploy it inside your stack. Designed for $2M–$20M revenue operators seeking immediate operational ROI without multi-month agency commitments or open-ended hourly billing.
-
Week 1: Operational Bottleneck & Security AuditProcess mapping + data privacy assessment.
-
Week 2: Functional AI Build & DeploymentCustom internal RAG engine, automated agent workflow, or reporting loop.
-
Hand-off & SupportLive team training session, 1-page technical architecture blueprint, and 30 days of post-launch tuning.
- 1 Custom Automated Pipeline
- Secure Data Model Isolation
- 1-Page Technical Blueprint
- 30-Day Tuning Support
Real-World AI Portfolios
We replace manual administration with secure, custom pipelines. Review the Problem, Solution, and ROI of our active deployments.
PoolzHQ / Bilz HQ
Field Service Route OS & Autonomous AI Marketing
Independent field operators lost 15+ hours/week managing manual routing schedules, customer dispatch pipelines, and executing local marketing campaigns.
Built a custom white-label Route OS paired with a secure multi-agentic AI marketing assistant that plans, drafts, and schedules local campaigns automatically.
Reduced administrative overhead by over 60% and established hands-off marketing generation loops.
RadianceSIS
Role-Gated RAG & Secure Knowledge Engine
Complex institutions struggle with scattered SOP document silos, strict role-based viewer permissions, and data leaks to public LLM systems.
Designed a secure, role-gated RAG engine that queries internal docs using vector retrieval while strictly respecting employee classification tiers.
Achieved 100% data governance compliance with instant, source-backed internal document search capabilities.
Member Portal Platform
AI-First Org Management & Agentic Workflows
Legacy member organization directories rely on manual spreadsheets, paper records, and manual roster updates that consume officer hours.
Built an AI-first member portal featuring automated check-in tracking, a Chapter Health Score dashboard, and an AI Rush Strategist agent (piloted with UMich SAE).
Replaced fragmented administrative overhead with autonomous check-ins and self-executing AI agent loops.
Lax Evals / FieldIQ
Digital Tryouts & AI Score Synthesis
Youth sports programs managed tryouts using physical clipboards and chaotic spreadsheets, leading to multi-evaluator scoring bottlenecks.
Deployed a digital tryout evaluation platform featuring automated AI score synthesis and instant player-rating summaries emailed to parents.
Streamlined multi-evaluator coordination from hours to seconds; live-validated with Spring Lake Lacrosse.
InstaStatsAI
Sports Video Analytics & Scoreboard OCR
Commercial sports analytics software costs $50K–$500K/year, placing pro-grade analytics out of reach for regional academies and schools.
Built a Firebase & Vertex AI pipeline utilizing scoreboard OCR models and Path B consensus algorithms to process raw gameplay footage automatically.
Delivers pro-grade analytics at SaaS price points while completely eliminating typical computer vision over-counting errors.
Curated SMB AI Tools Directory
We test and vet new artificial intelligence tools every week. Filter our direct recommendations by category and data security profile.
Custom Dedicated RAG
🟢 Enterprise SafeOpen-source models hosted in your dedicated AWS or Azure VPC. Zero data sharing or model training outside company boundaries.
NotebookLM (Google)
🟡 Exercise CautionExcellent for synthesizing source documents. Requires a corporate Workspace account and policy setup to opt-out of model tuning.
Self-Hosted n8n
🟢 Enterprise SafeA node-based workflow editor hosted inside your cloud account. API keys, database credentials, and customer data never leave your servers.
Make.com / Zapier
🟡 Exercise CautionFast integrations. Requires enterprise-level compliance plans and active endpoint encryption to keep data secure during transit.
Local Whisper OCR Pipeline
🟢 Enterprise SafeProcesses voicemail transcripts and scanned documents on local nodes or secured private APIs, keeping intake data private.
Public API Voice Agents
🔴 Public Model RiskHighly interactive, but audio streams are processed on public server arrays. Recommended only for public-facing marketing or screening.