No specs or technical briefs.
Most clients don't arrive with a specification. They arrive with a problem, such as reporting takes too long, data lives in spreadsheets, teams re-type the same information twice. That's enough to start.
The process
Onboarding
You share access to what matters: existing site/app, repos, analytics, brand docs. I define success metrics and set up your backlog. This takes days, not weeks.
Tell me what's needed
Add requests anytime, in your own words. "Our shift reports take hours" or "clients keep emailing for updates" is enough. I'll turn the problem into a proposed solution and confirm it with you before building.
I ship in a queue
I work on the highest priority item in your backlog. You get progress updates and shipped increments, typically every few days.
Review + iterate
You review on staging or via Loom walkthrough. Revisions are part of the model. I refine until you're satisfied.
Keep shipping
When one item is done, I move to the next. Add as many requests as you want, built one at a time, in your priority order.
Billing runs on a 30-day cycle with pause-anytime flexibility. See pricing for the full breakdown.
As many requests as you want, one at a time
You can add as many requests to your backlog as you like, there's no cap. I work through them sequentially, completing one before starting the next.
Why one at a time? Focus produces better work. Rather than juggling multiple half-finished tasks, I give each request full attention until it's done and approved.
Need more throughput? Get in touch and I'll figure out what works for your backlog.
Big projects, delivered in pieces
You might say "we need a way to see what's happening across all our sites". That's a valid request, even without knowing it should be a dashboard. I break large asks into logical chunks and deliver progress every few days.
Example: A field inspection app
- Core form interface and data structure
- Offline functionality and sync
- Admin dashboard and reporting
- Final polish, testing, deployment
You review each piece as it's delivered. Course-correct early. Never wait months to see something tangible.
Rolling out AI without the guesswork
AI requests run through the same queue, with more time spent on the people who will use the thing. That might be search over your documents, or assistants and agents that do work in the background.
A typical first rollout
- Agree the one question it has to answer well
- Gather and sort the docs / knowledge it will read
- Ship a pilot two or three people can try
- Walkthrough video and a one-page guide for them
- Fix what it gets wrong, then open it up to the team
Async-first, no endless meetings
We don't do daily standups. I don't do "quick calls" that turn into hour-long rambles. I ship work and show you the results.
| Backlog | Shared backlog and changelog (email + Loom) |
| Walkthroughs | Weekly Loom video + changelog |
| Overlap | One scheduled call per week if needed |
| Documentation | Written decisions so nothing gets lost |
Turnaround expectations
| Small changes | Often within 1-3 business days |
| Medium features | Shipped in increments over 1-2 weeks |
| Larger builds | Broken into milestones with weekly visible progress |