How to Test Your Backups: I Opened All 23 Copies, Not a Sample
To test your backups, check four things: that the job ran, that it copies everything that matters, that each copy opens, and that you can put one back in a known time. Here is how I run all four on my own server, with the counts, including the one the usual advice skips.
Read it →
Website Monitoring False Alarms: Why My 2:30 AM Alert Was Not an Outage
A website monitoring false alarm comes from one of two kinds of alert, and each kind needs a different fix. I dug into eight of mine, seven from outside the server and one from inside it, and here is what each turned out to be and the rule that has held for both.
Read it →
Can You Trust AI-Generated Code? Seven of My Seven Scripts Had a Fault Before They Ran
You can trust AI-generated code as far as someone who did not write it has checked it. An AI wrote or rewrote seven scripts for my business that each change something on a running system. An independent reviewer read every one before it was run for real, five reviewers and fourteen reports in all, and all seven scripts had a fault.
Read it →
MCP Server for Small Business: I Built Two, and the Building Was the Easy Part
An MCP server lets an AI assistant ask your business a question directly instead of reading your pages and guessing. I have built two, the second in one afternoon, and neither is in an app directory yet, so this is the list of what stands between a working server and a person using it.
Read it →
AI Vendor Lock-In for Small Business: I Ran the Exit Audit on My Own 57 Systems
When an AI engagement ends, a small business keeps what sits in files it can open and loses what was rented. I audited my own 57 systems as if a client were leaving tomorrow, and the files were ready to go while my own contracts were not.
Read it →
AI That Reads Your Own Documents: I Asked Mine 12 Questions and It Missed 3
Pointing AI at your own files is cheap to build and nearly free to keep running. I graded mine this morning on twelve questions I already knew the answers to, and the failures all traced back to what I let into the pile rather than to the software.
Read it →
The one-page AI policy a small business actually needs
Two thirds of small businesses use AI and just over a third have written a rule for it. Here is the whole policy on one page, plus the part the templates leave out: which lines software can enforce and which ones only hold if somebody remembers them.
Read it →
How to Measure AI ROI in a Small Business: 5,671 Runs, One Honest Grade
My own AI agents ran 5,671 times in two weeks and 99% of those runs succeeded, and exactly one of the 36 can tell me whether its work was any good. Here is what that gap means for measuring AI ROI, and the four steps I use instead.
Read it →
AI Maintenance Cost: $1.61 in Model Spend and 50 Things to Keep Alive
Two months of receipts from my own AI systems. The model spend was $1.61 across 508 runs. In the same two months the number of things that have to keep working went from 13 to 50. The bill is not the cost.
Read it →
AI and Data Case Study: The Database That Refuses to Guess
I built a database of all 978 NCAA softball schools that joins federal college data to coaching contacts read from each school's own website, with an AI that answers from it. This case study is about the part worth copying, the list of things it refuses to do.
Read it →
Why AI Writing Sounds Like AI, and the Four Checks I Run Before It Goes Out
AI writing sounds like AI for reasons you can fix in an afternoon. The thing that will actually embarrass you is different, and no detector catches it. Here are the four checks I run on everything that goes out under my name, and what each one has actually caught across nine posts.
Read it →
Why Your AI Gives Generic Answers: It Doesn't Know Your Business
Your AI gives generic answers because it has never seen your business, and neither a better prompt nor the newest models with their bigger context windows and built-in memory close that gap. Here is what does, what the work actually takes, and the honest limits of it.
Read it →
Private AI for Small Business: What Your Own Hardware Can and Cannot Do Yet
Data security is now the top barrier to small business AI, and running models on hardware you own answers it completely. I ran mine for two weeks and graded every page. The narrow jobs finished clean, the writing job passed 12% of its graded pages, and the split tells you which of your own jobs can move.
Read it →
AI Email Management for Small Business: The Rules I Wrote First
Most guides to AI email management pick a tool. The harder part is the list of things the tool may never do on its own. These are the five rules running on my inbox, and the numbers that say whether they held.
Read it →
How to Know Your Website Is Broken Before a Customer Does
Most website failures never take the site offline. The page loads, the form submits, the button says thank you, and nothing on the other end happened. Six ways a website breaks quietly, four of them ours, and the check that catches each one.
Read it →
Central Ohio Housing Market, August 2026: More Homes for Sale, Prices Holding
Central Ohio had 10.5% more homes for sale in August 2026 than a year earlier, a gap that has held all year, while August's new listings barely moved. Homes are staying on the market longer and prices are holding. County-by-county numbers, and how our forecast graded itself.
Read it →
The Data Your Business Already Owns
Most small businesses think a data project starts with collecting something new. It starts with reading the records you are already required to keep, plus the itemized history your bank has been storing for you. Here is what you already own, what it can answer, and how to read one folder of it this week.
Read it →
What Your Website Needs to Get Found by AI Assistants
AI assistants find a business by reading its website, and most of what they need is old, cheap, and easy to check. Here is the short list, what Google says you can skip, and what we found when we tested our own two sites.
Read it →
Does Your Business Need a Chatbot? Four Questions to Answer First
A chatbot is worth it when the same questions arrive while nobody is around to answer them and the answers are already written down. Here are the four questions that decide it, and what broke when we tested our own.
Read it →
AI Data Center Claims: The Fact Is True, the Cause Is the Propaganda
Electricity, water, and the AI bubble. Each narrative sits on a real number, and each one gets bent at the same seam. Here is how much of each actually holds up, with the sources.
Read it →
Human in the Loop Doesn't Mean a Person Checks Everything
Every AI vendor says they keep a human in the loop. Here is what the phrase actually means, why a review queue for everything makes people rubber-stamp, and the rule I use across 24 agents to decide which calls a person has to make.
Read it →
We Ran Our Grocery Receipts Through AI. A Third of the Bill Was Snacks.
We ran our own grocery receipts through AI to find out where the money actually goes, and about 36% of two trips turned out to be snacks, desserts, and drinks. Here is the full breakdown and how to run the same check on your receipts or your business spend this week.
Read it →
How Much Does AI Cost a Small Business? Real Numbers, Including Ours
The median small business now spends $2,200 a year on AI. Here is what that money actually buys, what we charge, and how to tell which way of buying AI fits your shop.
Read it →
I turned the two-hour Seth Godin episode into 18 clips for business owners
Seth Godin's Diary of a CEO episode is two hours long and most of it is aimed straight at small business owners. I built a flipbook companion: 18 timestamped clips, what each one means for your business and for you, and a flashcard deck.
Read it →
I graded my own forecasts for five months. The simple model won.
I run a monthly forecast on 15 Central Ohio counties and score every prediction against what actually happened. The plain baseline beat the sophisticated model on every measure. Here is why that matters if you are buying AI.
Read it →