About

I build the AI systems I wish smaller companies could buy off the shelf

Most teams do not need another generic chatbot. They need a reliable workflow that saves time, catches what gets missed, and fits how the business already works.

Headshot coming soon

Practical automation from someone who has run it in the real world

I started S2 Analytics for companies that can see the value in AI but do not have the budget, staff, or patience to manage a pile of enterprise subscriptions. The opportunity is real. So is the waste.

My work is built around a simple idea: use normal code for the parts that should be predictable, and use AI for the parts that need language, judgment, search, or synthesis. That is how you keep the system cheaper and easier to trust.

I have built and supported systems across industrial operations, furniture retail, mortgage investing, insurance distribution, financial data, and internal analytics teams. The patterns repeat: messy inputs, manual handoffs, reporting gaps, and decisions that should be easier to make.

The services on this site come from work I have actually done: ETL modernization, survey sentiment analysis, lead follow-up, competitor newsletter digests, internal GPTs, data pipelines, territory planning, and AI-assisted workflows with approvals and audit trails. I care less about demos and more about whether the workflow still works three months later.

10+
Years in data and automation
50+
Projects delivered
Lean
AI spend kept intentional
Solo
Senior builder, no agency bloat
Philosophy

How I think about AI work

Start with the workflow

A good AI project starts with a painful manual process, not a model choice. If the workflow is fuzzy, the system will be fuzzy too.

Keep the boring parts boring

Calculations, routing rules, lookups, and scheduled jobs should be deterministic whenever possible. AI should not guess when a script can know.

Build for ownership

You should know what the system does, what it costs, where it can fail, and how to update it. The handoff matters as much as the build.

Toolkit

Tools I use when they fit

The stack depends on your existing systems, data sensitivity, budget, and team skills. I am not trying to force every client into one platform.

AI orchestration

OpenAI Claude LiteLLM RAG pgvector

Automation

n8n Windmill Python TypeScript

Data

SQL PostgreSQL Snowflake dbt

Apps and delivery

Vercel Astro APIs Dashboards

Business systems

Email CRM SharePoint Slack Teams

BI and reporting

Power BI Tableau Metabase Scheduled reports
Have a workflow in mind?

Start with one useful AI system

If your team is manually chasing leads, reading competitor emails, answering the same internal questions, or copying data between systems, that is a good place to start.

Want examples? Drop your email and I will send practical AI workflow ideas for smaller companies.