Bring Your Own Bot
What building the bot in Klau taught me, why the Basecamp approach is obvious in hindsight, and how Klau and Strata will work with the AI agent each user already has.
Notes on Klau, Strata, and the work behind them, including the weekly Building Klau in the Open series. Newest first.
What building the bot in Klau taught me, why the Basecamp approach is obvious in hindsight, and how Klau and Strata will work with the AI agent each user already has.
I plan and curate the work during working hours. A fleet of AI coding agents does most of the coding, around the clock. My job is to confirm, send back, and refine until the result matches the spec.
A different kind of weekly post. My son turned 13, and I wrote about the entrepreneurs I grew up around and the principles I hope he picks up.
The week went into a project built on Klau that is already running part of a real hauling operation. I am keeping it quiet for now, and looking for haulers who want to adopt a system as it is built.
For years I filled out a forced-ranking spreadsheet called Org Alignment. The people were good. The structure was built to extract, and it is why I build the way I do now.
For legacy vendors without integrations, I am building a Chrome extension that brings Klau tools into their web screens. Also, a week of honest posting about the past.
Years ago I was in leadership during a layoff called a transformation, and flew in days later to work with the team it hit. It is why I plan to wrap my projects in a public benefit corporation.
I spent the week thinking about the public benefit corporation structure, and why I want its charter and duties to protect Klau's customers long after I stop running it.
A camping trip in Big Sur, a podcast with Eric Ries, and a settled answer to how Klau will be structured: as a public benefit corporation.
I barely worked on Klau this week, and the pilots kept running. That is the kind of business I want to build.
I spent Earth Day doing trail work instead of shopping. On the Klau side, I fixed how dispatch keeps one driver on a live load from start to finish.
A prospect asked what happens to the founding member deal if Klau is acquired. My answer: acquisition is not the goal. I want Klau to be a company that can be passed on.
A hauler needed Klau to dispatch jobs from files in an SFTP folder. That one customer need became the Dispatch Exchange, available to every Klau user.
Many haulers run legacy systems with no modern API. The Klau CLI lets them point Klau at a folder and start dispatching.
Klau was meant to sit on top of what haulers already run. After months of hearing that their driver tools fall short, I built a thin driver app that learns each driver's routine.
An enterprise pilot surfaced hardcoded assumptions and time zone issues. I fixed those, merged the job creation paths into one service, and stopped a runaway AI cost.
A calm week with two paying customers, bug reports from real users, and a change to the dispatch optimizer that lets it handle large fleets.
New haulers now get a checklist that walks them through setup. Conversations this week showed where Klau fits for haulers who already take orders online.
Eight weeks in, I can put the idea behind Klau into words: you onboard it like a person. I recorded an 18-minute overview of everything built so far.
Dispatching was the groundwork. The product is a self-service ordering storefront for roll-off companies, with pricing, payments, and dispatch handled behind it.
Klau Coaching is open for sign-ups at $49 a month. Upload your dispatched work and Klau scores each day and shows where time was lost.
The dispatch scorecard is now a coaching tool inside Klau. It scores human-dispatched days and drivers and shows what Klau would have done instead.
A short week. I took my son Grant to King of the Hammers, the off-road race in the desert, and came home refreshed.
Two more demos, a simpler pricing model in progress, and a new scorecard tool that measures what a day of dispatched jobs could save.
The first real demo, what Klau is and is not, pricing based on jobs optimized, and an offer for 10 founding members.
One week after announcing Klau, two conversations led to disposal costs in routing and a new way to think about onboarding dispatchers.
After years of solving logistics problems for others, I am building Klau, dispatch software for roll-off haulers.
Klau is a dispatch and operations platform for haulers. It works alongside the systems they already use. This is how we built its foundation.
A simple guide to share with older parents this Thanksgiving. Covers phone scams, fake virus warnings, and how to find hidden subscriptions.
What a decade in waste management taught me about building a small, profitable software business in an overlooked industry.
Finding the right account when a customer calls can be slow for waste hauler CSRs. HaulerPhone shows the caller's account when the call comes in, with AI call transcripts planned for a later release.
HaulerPhone shows waste hauler customer service reps who is calling and which accounts they have. It starts at $49/month, works alongside existing systems, and is planned to launch in September 2025.
Three days ago, Weighven was an idea. Today the frontend runs on desktop, tablet, and phone. The backend is next. Here is what I built and what I learned.
After nearly two decades in the waste industry, I'm restarting Edison Code. I want to hear about the problems you face and build small, focused tools for them.
Why building for independent haulers means looking past easy metrics and working within their real constraints.
Waste touches every household and business, but the software that runs it is still behind. Here is why I care about closing that gap.