AI systems for real estate funds, operators and enterprise teams
I build the AI systems your executives can actually trust.
Governed data platforms, MCP servers and AI agents that answer real business questions from your own systems, with the source behind every answer.
Systems built for and with
- Triten (opens in a new tab)Real estate investment
- Black Mountain Fund (opens in a new tab)Real estate fund
- Anchor Mill Group (opens in a new tab)Real estate operator
- Back2SleepNow (opens in a new tab)Consumer health
- Prestyj (opens in a new tab)AI voice agents
- EZ CoderAI coding agents
Selected work
Systems in production
Every outcome on this site links to a build record or public evidence.
Triten
Triten OS
Triten's executives ask ChatGPT or Claude plain questions about their portfolio and get ranked, cited answers drawn from governed Yardi, Monday.com and SharePoint data, without anyone assembling a report first.
- Data platforms
- MCP servers
- AI agents
- Analytics
Back2SleepNow
Back2SleepNow autonomous content engine
Back2SleepNow publishes researched, fact-checked articles without a writer or editor in the loop. An automated gate checks every claim against its source and only merges what passes.
- Content engines
- Automation
- AI agents
Black Mountain Fund
Easefolio fund management platform
Black Mountain Fund runs fund accounting, investor reporting and an AI assistant on Easefolio, while a modern replacement is rebuilt behind evidence-checked migration gates.
- Internal tools
- AI agents
- Analytics
- Automation
Capabilities
What I build
AI agents
Agents that do real work inside your systems, with guardrails and an audit trail.
MCP servers
Your governed data exposed to ChatGPT and Claude through the Model Context Protocol.
Data platforms
One trusted model of assets, people and money that every tool reads from.
Automation
Recurring manual work replaced by pipelines that run, retry and report.
Content engines
Evidence-backed publishing systems that turn real work into public proof.
Voice AI
Voice agents that answer, qualify and book, around the clock.
How I work
Small team, direct access, production from the start
Start from the decision
We pick one question an executive needs answered, or one job a team repeats, and define what a trustworthy answer looks like.
Govern the data first
Your systems of record stay in place. I build a governed layer that keeps evidence, permissions and lineage on every fact.
Put agents on top
Agents and ChatGPT or Claude connections read from that layer, read-only by default, with sources cited in every answer.
Ship, measure, hand over
Production from the first weeks, checks in CI, and documentation your team can own. No black boxes.
Writing
Latest case studies and notes
Article
Running Marketing With a Gate Instead of a Team
Back2SleepNow's content engine researches, writes, fact-checks and publishes on its own. What changes when an automated gate replaces the editorial team.
· 4 min read
Case study
How a Full Day of Portfolio Risk Review Became 10 Minutes
An asset manager asks ChatGPT for the top 20 assets to worry about. A read-only MCP server answers from governed Yardi, Monday.com and SharePoint data.
· 5 min read
Field note
Why I Let AI Agents Build Client Systems, and What I Never Let Them Skip
How Nolan Grout ships enterprise systems with AI coding agents. A second agent reviews every job, gates block bad changes, and owners keep the decisions.
· 4 min read
Have a system in mind? Let’s scope it.
A focused 30-minute call with Nolan. You leave with a clear view of what to build first and what it will take.