Work samples

Work you can inspect.

Three things we built and shipped, then four worked examples that show how we would approach a problem like yours.

Built and shipped

Software and guides that exist today.

Real products, real users, described by their actual status. Each one is the kind of work we would do for you, done first for ourselves.

Private beta · Construction

Construct Buddy

The problem. Estimators spend hours on every plan set counting doors, windows, fixtures and square footage by hand, and the drawings often disagree with the schedules. Missed counts and unresolved discrepancies become RFIs and rework after the bid.

What we built. A web application that reads construction drawing PDFs, identifies building elements with vision models, cross-references the door and window schedules, and produces a quantity takeoff by CSI division with discrepancies flagged and draft RFI language, all behind a human review step.

What it shows. Document-heavy workflow automation with review built in, cloud infrastructure we operate, and the discipline to say plainly what the tool does not do.

See Construct Buddy

Public, free to use · Education

RateMyCollege

The problem. College rankings say nothing about whether a graduate can afford to live after graduation. Families compare prestige when the useful comparison is what is left after rent, taxes and loan payments.

What we built. A free public tool that turns published earnings, cost-of-living and debt figures into one number a family can use: the monthly budget a graduate of that program can expect where they live. Trade schools and community colleges included. No account required.

What it shows. A data product built on original research into college outcomes and shipped at no cost for the families who need it most.

Try RateMyCollege

Open source · Guide

AI Strategy Field Guide

The problem. Most organizations start their AI conversation with a vendor demo instead of a question about their own work, and end up with a pilot nobody owns.

What we built. A free, open-source field guide for leaders who are not engineers: identifying use cases, deciding build versus buy, data readiness, governance and implementation roadmaps.

What it shows. The thinking behind our Technology & AI practice, published for anyone to use before they ever talk to us.

Read the guide on GitHub

Worked examples

How we would work a problem like yours.

Four executive briefs, one from each practice area, each walking from the challenge to the decision, the approach, the deliverable and how success would be judged.

Illustrative work sample. Built on a synthetic organization with invented figures to show how we think and what a deliverable looks like. It is not client work and describes no achieved result.

Strategy & Operations

Standardize intake before the new contract starts.

A three-site workforce nonprofit compares three operating models on compliance, continuity, staff capacity and cost. The chosen model comes with a 90-day plan and built-in decision points.

Option ASite-led intake, shared reporting

Option BShared standards, one data owner

Option CCentral intake team

Open the brief

Technology & AI Transformation

Which AI request should a city pilot first?

Three AI requests, one pilot budget, and no framework for choosing among them. The decision record sets dependencies, stop conditions and acceptance checks before anything is bought or built.

Internal updateBounded pilot

Policy finderResolve ownership first

Resident answersDefer

Open the decision record

Custom Solutions & Automation

The report is ready. The inputs are not.

Two program teams use different definitions for the same report, and nobody noticed until the funder asked. The fix is a workflow that holds the run on any exception, flags it and keeps a person as the release gate.

Held for review

FindingDefinition mismatch between sources

Next actionConfirm definitions before aggregation

Open the workflow design

Workforce & Change

A polished answer can still invent the policy.

A change plan for introducing AI-assisted drafting into student-services responses, built around the fears staff actually raised. It includes the practice exercise staff use to learn exactly where the source ends and invention begins.

“Walk in Friday before 4 p.m.; the help desk can grant a late-arrival exception.”

Which of those claims does the source support?

Open the adoption plan

Also on this site

The longer brief and the writing behind the work.

The worked examples show judgment. The articles show how we think about technology decisions and working with AI. Echo, our memory engine for AI agents, is in early access.

Printable reporting decision brief
The long-form version of the reporting example, formatted as a client-ready document.
Echo
A memory engine for AI agents and teams. Early access, in development.
Insights
Guides on build-versus-buy, specifications, deployment and working with AI.

Next step

What would a useful deliverable look like for you?

Start with the decision, the process or the change your organization needs to get right.

Discuss your priorities Email us