Monthly close and variance reporting
Pulls the month's actuals from your property management system, builds the variance tables against budget, prepares draft commentary on what moved, and formats the package to your template.
More room in the week for the work your team was hired to do.
Automated workflows for Canadian real estate operators that handle the assembly behind your team's reporting packages, so the hours go to analysis and decisions. Built by an engineer designing agentic AI systems for real estate analytical work since 2022.
For principals, asset managers, and controllers whose teams lose days each month turning system exports and spreadsheets into reports.
What it automates
Four workflows, installed independently inside your systems. Every one runs the same way: the workflow handles the assembly; your team reviews and decides.
Budget season and quarterly reforecasts
Loads the actuals baseline into your budget model, drafts each property's budget on your template with your team's assumptions, rolls the portfolio together, and re-runs the year at each reforecast — actuals in, assumptions updated.
Acquisition diligence
Prepares the data and modeling behind an acquisition: submarket research, comp pulls, pro forma scenarios run through your model — every input documented.
Capital partner reporting
Assembles the quarterly package for LPs and co-investors: performance, capital account statements, fund-level summary — built to your partner-reporting templates, not a vendor's.
These four are the most common starting points, not the full list. If your team's heaviest recurring package is something else — lender reporting, loan draws, anything with a deadline and a template — it likely fits the same pattern.
How it works
- Define A first conversation about where your team's hours actually go: monthly close, partner package, diligence work. Then a written plan for the workflow worth building first. (No cost to start.)
- Build The workflow is built inside the systems you already use: Yardi, Argus, Excel. It handles the assembly: pulling the data, filling in your model, preparing the draft package.
- Refine It's tested against packages your team has already produced, and adjusted until the output meets your standard.
- Handover Your team is trained to run it, and everything built is yours. No ongoing dependency.
The package lands as a finished draft — tables built, commentary written, every figure traced — and your team's week starts at review.
How an installed workflow runs
Your systems
AUTOMATED ASSEMBLY
feeds in
01
Capture
every figure tagged to its source
02 · BUILD
Set steps
same every run
AI
reads & writes, within the steps
03
Draft
a draft package, ready for review
Your model
HUMAN JUDGMENT
Review
experience · interpretation · decisions
Finished package
in your format, ready to send
every figure traces back
Why work with me
You know who you're hiring
You deal directly with the person who does the work: no account team, no rotating juniors, no consultancy markup.
Built by an engineer who's done the work
Most AI for this work comes from engineers without the domain or consultancies without the engineering. This practice has both — four years building agentic AI systems, plus the finance background to produce analysis to institutional standard.
Every number traces to a model
Behind every report is a real Excel model. Every figure ties back to the cell or source it came from.
Owned by you
It runs on your own machines, and what gets installed is yours to keep — an asset, not a subscription.
Proven on institutional REIT diligence
A diligence workflow for publicly traded REITs, installed for a senior buy-side reviewer in 2025. Built over six weeks, calibrated to the reviewer's standard. The workflow produces deep diligence analysis on any target REIT, and runs on subsequent REITs without rebuilding.
I was skeptical AI output could meet the standard this work requires. The workflow surprised me: research and assembly at a depth I could rely on, and when I questioned a figure, I could trace it to its source. We've kept using it on names since.
Pricing
No cost
First conversation
30–45 minutes to walk through the workflow you're considering and whether it's a fit.
$5,000
Installation scoping
A written plan for what gets built, how it connects to your systems, and a firm price for your installation. Two to three weeks. Credited toward installation if you proceed.
$30,000–$60,000
Single workflow installation
Build and install one of the workflows above. Six to ten weeks. Scoping sets your firm price. The final payment is tied to your reviewer's acceptance.
FAQ
Why not just use ChatGPT, Copilot, or Claude?
A conversation isn't a process. Those tools are assistants your team steers, task by task — great for the everyday. A monthly close or partner report has to come out the same way every cycle, built to your standards, every figure traceable. That's a process you own, not a conversation you run.
What stops the AI from making numbers up?
It never writes a number. The AI reads your documents and drafts the words; every figure is filled in automatically from your model or a source document, and checked against that exact cell or document before the package reaches your team.
What happens to my data?
Most of the workflow runs on your own systems — your data and your Excel or Argus model never leave your environment. Only the steps that need AI call Anthropic's models, and only with the specific data that step requires. Those calls run under terms that prohibit training on your data or retaining it, in a Canadian region where required.
Will this work with my existing systems?
Yes. It installs over your current systems — Yardi, Buildium, AppFolio, MRI, custom, or any mix — and platform AI features (Yardi Virtuoso, etc.) keep working alongside it. No migration or replacement.
How much of my team's time does installation require?
Roughly 15–25 hours over the 6–10 weeks: an initial deep-dive on your current workflow, calibration reviews at milestones, and training at handover. Most of the work happens without your team.
What happens after the engagement ends?
Your team runs the workflow through an internal AI champion trained during installation. The systems, models, and processes built are yours to extend — no ongoing dependency. The practice takes installations through 2026; after that I'm taking this same work in-house at a global equity hedge fund. Engagements are built so the end date doesn't matter: the workflow, and the person who runs it, are already yours.
Background
I built the first version of what runs this practice in 2022, to automate the analytical work real estate operators produce — diligence, performance analysis, partner reporting. That first system took a property address and produced a post-acquisition value analysis.
The output fell short of the standard the work requires; the four years since have been spent reaching it.
Based in Vancouver, working with operators across Canada.
Get in touch
The first conversation costs nothing — bring the report or process that takes the most of your team's month, and we'll see whether a workflow fits. One or two sentences is plenty.