Everyone says "we ship fast." What does that actually mean?
The phrase "ship fast" has lost meaning. Every agency and every consultancy claims it. What matters is what fast actually looks like when the constraint is real, and what the trade-offs are that nobody talks about.
Three weeks is the shortest cycle that reliably produces something a real user can touch. Faster than that and you're either shipping a prototype dressed up as a product, or you're skipping steps that will cost you later.
The question is never "can you ship in three weeks?" It's "what should be in scope so that three weeks produces something real?"
What fits in three weeks
Here's the honest breakdown of what we can deliver in a three-week engagement, given a focused scope:
- 1
A single-purpose AI or automation tool
One clear function - document extraction, chat interface, workflow automation. Deployed, authenticated, with basic error handling and monitoring.
- 2
A landing page with a working backend
Marketing site, waitlist capture, payment integration, and email flows. Enough to run a real launch, not a mock-up.
- 3
An internal dashboard on top of existing data
Connects to your systems, presents a focused view, includes basic access controls. Solves a real internal problem, not a hypothetical one.
- 4
A hardened prototype
If you've already built something with AI tools, we can take it from local prototype to production-ready in three weeks. Security, deployment, tests, monitoring.
What doesn't fit in three weeks
Just as important - the honest list of what we can't reasonably do:
Beyond three weeks
Multi-tenant SaaS platforms, complex data migrations, mobile apps with native features, enterprise integrations across five systems, anything that needs regulatory approval. These are real projects, but they need real timelines.
Trying to compress those into three weeks either produces something fragile that you'll rebuild in six months, or something that only technically ships but doesn't actually work. Neither is a good outcome.
How the three weeks actually break down
The rough shape of a three-week engagement, though every project shifts these boundaries:
- W1
Discovery, scope lock, and core build
Days 1-2 are for understanding what you actually need. Days 3-7 build the core loop - the one thing the product must do, working end to end. By Friday, something runs.
- W2
Hardening, integration, and iteration
The rough version is now workable. Week two is where we add authentication, connect real data sources, handle the error paths, and iterate based on your feedback from seeing week one's output.
- W3
Polish, deploy, and hand off
Deployment, monitoring, documentation. A short internal test phase. Handoff, or continued engagement into a longer sprint if there's more to do.
The trade-offs nobody mentions
Working in three-week cycles requires two things most agencies won't sign up for: real focus, and real conversations.
Real focus: we say no to scope creep. If new requirements appear mid-sprint, they go in the next cycle - not this one. That's how the timeline holds.
Real conversations: we need to talk to you directly, weekly, without account manager buffering. Ambiguity is what kills three-week cycles. Direct access is what saves them.
What we ask of clients
Be available for a 30-minute call each week. Reply to Slack within a day. Make decisions when we ask you to, or empower someone on your side to. If you can't do those three things, three weeks becomes five - and that's on both of us.
Why not just do it in six weeks?
You could. Sometimes you should. But three weeks is where a very specific kind of magic happens: everyone stays focused, decisions get made quickly, and you have a working thing in front of real people before the initial idea has time to drift.
Six-week cycles are where scope grows, priorities shift, and the original idea gets muddied by everything you learned along the way. Three-week cycles force a version of the answer to exist before the question moves.
Got an idea that would fit in a three-week sprint?
Tell us what you're trying to build. We'll give you a written scope and a fixed price - the plan for what fits, and what would need to wait for a second cycle.
Start the conversation →Closing thought
Three weeks isn't a marketing timeline. It's the shortest realistic cycle where you get something worth showing to a customer, an investor, or a team. Anything shorter is a demo. Anything longer starts drifting.
If you're clear on what you want and willing to focus, three weeks is genuinely enough. If you're not sure yet, one honest conversation will tell us both.