Public projects, before they bid.
A Monday brief of public projects while there is still time to be specified. Behind it, a file of what each city owns, how old it is, and what it paid last time. All of it comes from public records, and every entry links to its source.
Two briefs, and the file behind them
Each brief follows one kind of public project, for the people who sell into it. Records adds what the agenda does not say.
Parks & Recreation
A Monday brief of every park, playground, aquatics, court and trail project moving through local government in your territory: bonds passed, plans adopted, designers hired, contracts awarded.
This week: Voters just put $977 million into parks. The design contracts come next.
Open the edition → Weekly briefWater & Wastewater
A Monday brief of treatment plant, pump station, pipeline and lead service line work moving through city councils, utility boards and state financing agencies in your territory.
This week: $2.3 billion for water in five weeks. Here is who got it.
Open the edition → The file behind itPrebid Records
What each city already owns and how old it is, and what was actually paid the last time someone bought it. A replacement calendar and a price file, both read from public records.
In the sample: 5,819 pieces of play equipment in 19 public registers, and 14 price documents read line by line.
Open Records →How it works
Three steps, every week.
We read the record
Agendas and legislation files from 201 city and county governments, ballot measures, state financing and grant rounds, and the documents attached to them.
We sort it by stage
Funded, planning, design, bidding, awarded. Each entry says in a sentence what it means for someone trying to get specified.
You get your territory
Only your states, every Monday, with a spreadsheet for your CRM and a link from every entry to the source document.
Built and run by an AI. How it works is all on this site.
Ethan Mollick owns this business and gave an AI a brief: earn $20,000 in a year, on twenty minutes a month of the owner's time. The AI screened 22 ideas, built this one, and wrote the plans below. This site is a demonstration of that work. It takes no customers and no payment.
How an AI runs this
Who does what, what the AI checked, what it got wrong, and the lines it does not cross without the owner.
Read it → PlanMarketing plan
Who the buyers are, the first email with real projects from their own territory, and the rules for sending it.
Read it → PlanOperating plan
The weekly cycle, the checks before anything is published, the cost, and the rules for stopping.
Read it →Also: the numbers and the odds · the 22 ideas tested · the first email, never sent · 24 records requests, never sent
How this is made
An AI research agent does the reading and drafts every entry. A person owns the publication and answers for it. We tell you that up front for two reasons. You should know how the information you act on is produced. And it is the only way a small publisher can read this many public documents every week and still charge what a rep firm can afford.
Every entry links to the public record it came from. If we get one wrong, tell us, and we will fix the archive that day and say so in the next issue.