ARTICLES · DUE DILIGENCE, ON US

Read the guides before you commit.

Decision guides and engineering notes on choosing, scoping, building and running AI systems that keep working in your daily operations, long after the launch.

  • Read guides to choosing and building useful AI systems.
  • Explore the engineering work behind reliable operation.
  • Test a proposal before you commit budget or people to it.
A woman and a man reviewing a document across a desk by office windows

What To Demand From Whoever Owns Your System

Handing a live system to one accountable person outside your payroll works only if you hold them to a demanding bar; otherwise you have bought a single point of failure. Here is that bar, and how to check it before you sign.

Product lead shows a one-page target to an engineer at a small kitchen table

3 Tests Before Anyone Starts Building

Readiness is not enthusiasm, and it is not feasibility. It is whether a stranger can grade the target, the systems can be reached today, and one person can change the target tomorrow, all checked before any budget moves.

A finance controller and a warehouse operations manager check a printed exception against a clipboard beside sorting trays, the warehouse floor behind glass

The Work Your ERP Was Never Built To Do

The enterprise system covers the standard 80% of the work. The delay, cost and eventual control finding sit in the 20% around it, now run on a spreadsheet and a weekly meeting. Build the software around it, not instead of it.

A finance controller reviewing audit records with a remote audit committee on a large video screen

Your Material Weakness Disclosure Is A Software Deadline

A disclosed control weakness is read in the boardroom as a governance sentence. It is also a dated commitment to produce evidence in a volume an auditor can sample, and records at that volume are produced by systems.

A head of data walks an operations vice president through a sample of records plotted on a wall screen

Is Your Data Good Enough To Start Building?

You have a workflow worth fixing and no way to tell whether your data can carry it. Build-ready data is accessible, representative and stable enough for one workflow. This is the lowest bar it has to clear before a real build starts.

Two colleagues comparing papers at a desk near office windows

A Spreadsheet Is A Control Until Someone Asks

Remediation fails on evidence, not process. The control usually exists and someone really performs it; what is missing is proof that it ran, in order, by the right person, before the period closed. Make the step record itself.

A controller marks a row of calendar pages while the CFO and the audit committee chair review the restatement timeline in a high-floor boardroom

A Restatement Is A Deadline Nobody Can Move

A financial restatement adds a second period of urgent work to the process that produced the error, on dates set by regulation. Build the evidence and controls needed to meet those dates without repeating the same error.

A telecom director on a phone call in a glass corridor above the city, holding a closed laptop

Your Routing Decisions Run On Last Week's Numbers

Nothing looks broken. Your team still produces defensible answers from last week's numbers, and margin leaks out of the routes nobody had the time to re-price. Here is when routing becomes an AI problem, and where to start.

A service coordinator and field supervisor checking a job packet in a depot dispatch room

Which Service Workflow Deserves Your First AI Investment

Compare operational consequence, system access and adoption before funding an AI workflow, with the evidence written beside each answer. The most useful first investment solves a problem your team can act on today.