I build AI systems for ops-heavy businesses.
I map real workflows, find the highest-ROI bottlenecks, and build production systems that show what expired, aged, changed, needs approval, and who owns the next step.
Who I help
Ops-heavy businesses with work stuck between people, tools, and spreadsheets.
The best AI projects usually do not start with a model. They start with a workflow that keeps breaking, aging, changing quietly, or waiting for one person to approve the next step.
Family offices & property operators
Local-first systems for back-office workflows, recurring compliance, cash visibility, document tracking, approval queues, and owner-ready reporting.
Construction & real estate teams
Workflow automation for status reporting, field updates, property research, vendor coordination, and project visibility.
Wholesale distribution teams
Automation for inventory checks, reorder decisions, supplier follow-up, spreadsheet updates, and exception reporting across SKUs and vendors.
Insurance & Salesforce-heavy teams
Controlled AI agents for teams already living in Salesforce, CRM queues, policy records, service cases, and approval-sensitive workflows.
Client Work
Real projects. Measurable outcomes.
Proof that the work gets past strategy and into operating systems clients actually use: dashboards, exception queues, audit logs, and approval paths.
Problem
A Midtown hotel construction project required manual reporting and status chasing before client check-ins.
What was built
Construction progress dashboard for a Midtown hotel project, turning field updates into live status visibility and stakeholder reporting.
Problem
A Brooklyn real-estate operator needed recurring StreetEasy market research without spending hours manually searching, copying, and formatting listings every other week.
What was built
Automated StreetEasy research pipeline that pulls matching listings, filters by criteria, and delivers structured results into Google Sheets.
Problem
A NYC family-office and real-estate operator needed automation around sensitive back-office work without moving financial records, tenant data, or documents into another SaaS layer.
What was built
Local-first automation infrastructure on client-owned hardware, starting with insurance expiration tracking and expanding toward delinquency, QuickBooks, cash, loan, and internal communication workflows.
Systems I build
A dense map of the workflows I can turn into production systems.
Not a tool gallery. These are the operating surfaces buyers actually feel: exceptions, approvals, dashboards, follow-ups, and audit trails.
Insurance expiration tracking
Expiring docs, COIs, follow-ups, and daily exception summaries.
Rent delinquency readiness queues
Dirty ledgers separated from outreach-ready rows before automation starts.
Maintenance & vendor triage
Requests routed by urgency, owner, vendor, and next required action.
Owner/operator dashboards
What changed, what is stuck, who owns it, and what needs approval.
n8n workflow automation
Self-hosted workflows with dry runs, logs, alerts, and fallback paths.
Agentforce service agents
CRM-native agents for intake, qualification, service, and routing workflows.
AI research pipelines
Prospect, market, app, and niche research turned into usable briefs.
Content intelligence systems
Signals, drafts, scoring, freshness gates, and publishing queues.
Approval + audit trails
Human approval boundaries for sensitive financial, tenant, and client actions.
Local-first automation infrastructure
Systems designed around existing files, local servers, and office workflows.
App marketing operating systems
Scoreboards, ASO/SEO loops, competitor intel, and test calendars.
AI Ops Teardowns
Workflow maps that show how I think before anything gets built.
AI Operations Diagnostic
Start with the workflow, not the tool.
For teams that know AI matters but do not know what to build first. I map the work, find the exception layer, and scope the first production system with human approval and audit trail.
2 weeks
timeline
3–5
use cases
1
first scope
Workflow map
What actually happens today, where work stalls, which reports are trusted, who owns each handoff, and which tools are involved.
Exception inventory
A scored list of expired, aged, missing, changed, blocked, and approval-needed work that automation should surface first.
Implementation roadmap
The first 3–5 systems to build, sequenced by business impact, data readiness, audit needs, and implementation risk.
First-build scope
A clear proposal for the safest highest-leverage workflow, including human approval design, logging, timeline, dependencies, and acceptance criteria.
AI Ops Teardowns
How I think through the workflow before building the automation.
Teardowns are compact workflow maps: the bottleneck, the data source, the exception, the approval boundary, and the system I would build first.
View automation workInsurance expiration exception layer
Property OpsScan policies and COIs, flag expirations, draft follow-ups, route approvals, and log every status change.
Rent delinquency data readiness queue
Family OfficeValidate reports before outreach, separate clean rows from edge cases, and keep humans in the approval loop.
Field note to owner update workflow
ConstructionTurn messy jobsite notes into punch items, owner-safe updates, PM approvals, and daily project logs.
Work
A balanced view of client outcomes, sandbox capability demos, the operating systems I run myself, and apps. Some projects are anonymized or demo-scoped because client trust matters more than theatrical proof.
Client Outcomes
Production client work from real operating problems. These stay first because the consulting buyer needs proof that the work survives contact with messy workflows.
Sandbox & Capability Demos
Built demos that show how I design, test, and deploy controlled AI workflows before applying the pattern inside a client environment.
What I Run
Internal agent operations and automation infrastructure that runs my own consulting, content, research, and governance workflows.
AgentGuard — AI Governance Layer
The trust layer missing from most AI deployments. Confidence scoring decides what the AI handles autonomously vs. what needs a human. Every decision — automated or overridden — is logged with full reasoning. Live demo: HR candidate screening.
≥70% confidence → auto-execute · <70% → human review · every decision auditedApps
Product work that shows taste, systems thinking, and AI-directed implementation without making the site look like a developer portfolio.
Vista
Native iOS movie rating app — live on the App Store. Taste profiling, AI-powered performer rankings, and 30-day trending analysis built entirely through AI-directed development.
Nash Satoshi — Crypto Game Theory Rankings
4-LLM ensemble that cross-validates crypto market analysis across Claude, GPT, Gemini, and Grok to produce consensus rankings. Eliminates single-model bias.
4-model consensus · 32-node n8n pipeline · ~4 min end-to-endAdversight AI
Automated competitor ad analysis tool that scrapes rival Facebook/Instagram ads, analyzes their creative elements, compares them to your ads, and tells you exactly what to change to match their results.
Traditional analysis: $3k+/mo → automated weekly deliveryServices
Outcome-first implementation: diagnose the workflow, define the exception layer, add human approval where sensitive work requires it, then build into production.
n8n automation consulting →Operations Diagnostic & Roadmap
A focused workflow audit for teams that know AI matters but do not know what to build first. I map the current process, define the exception layer, and scope the safest high-ROI first system.
Best first step for family offices, property operators, finance teams, and ops-heavy businesses with aged, changed, expired, or approval-sensitive work.
Internal Workflow Automation
Production n8n and Python workflows for reporting, recurring research, approvals, inbox routing, spreadsheet updates, vendor follow-up, and local-first workflows where data cannot casually leave the client environment.
Typical result: 10–20 hours/week recovered from manual processes.
AI Dashboards & Reporting Systems
Client-ready dashboards and AI reporting systems that show what changed, what expired, what aged, what is stuck, who owns the next step, and what needs approval before it becomes a bigger problem.
Built for owners and operators who need visibility without another meeting.
Salesforce / CRM AI Agents
Specialized AI agents for teams already living in Salesforce or a CRM. Conversation design, routing, deterministic actions, testing, and production deployment around your actual data model.
Turn existing CRM workflows into controlled AI-assisted operations.
Includes ConversationFirst™ methodology.
How I work
The first job is judgment: figure out what should be automated before anyone starts buying tools or building demos.
Map the real workflow
I start with how the work actually moves: files, inboxes, approvals, spreadsheets, people, edge cases, and the places ownership gets fuzzy.
Rank the bottlenecks
Not every workflow deserves AI. I score opportunities by time saved, visibility, risk, implementation complexity, and how painful the failure mode is.
Scope the first system
You get a clear build recommendation: what to automate first, what data it needs, what success looks like, and what should stay human-owned.
Build it into production
If the scope makes sense, I build the workflow, document it, test the edge cases, and leave your team with a system that actually runs.
About
I spent six years as a Business Systems Analyst at Spectrum Enterprise, working across systems, teams, process gaps, and rollouts where the hardest part was rarely the software itself. It was getting the workflow, ownership, exceptions, and handoffs right.
That is the lens I bring to AI implementation. I map how work actually moves, identify the highest-leverage bottleneck, then build controlled automation around the real operating pattern instead of forcing teams into a generic tool.
My work now spans client dashboards, recurring research pipelines, Salesforce Agentforce agents, n8n workflows, and local-first operations systems. The common thread is practical implementation: visible exceptions, human approval where it matters, and systems that operators can actually use.
Experience
AI Implementation Consulting
Sep 2025 – Present
Spectrum Enterprise / Charter Communications
Business Systems Analyst
Jun 2019 – Aug 2025
Ithaca College
BS Sport Management, Minor Legal Studies
2014 – 2018
Client Results
NYC construction & real-estate operator
Built a live AI dashboard for 3 active job sites, then automated their bi-weekly StreetEasy research. Two projects, same client, both in production.
NYC family-office operations
Building local-first automation infrastructure for sensitive back-office workflows: dedicated workflow machine, audit trail, human-in-loop controls, and recurring compliance automation.
Tools & Systems
AI & Automation Stack
Adjacent Technical Fluency
Let's build something.
Tell me what workflow is slowing the team down. I'll tell you what I would build first.
Response within 24 hours.