Secondary · Workflow Sprint
One workflow. End-to-end. Done in four weeks.
2–4 weeks. Fixed scope. Fixed price.
Operators or teams with one specific recurring workflow that's either consuming time or producing inconsistent output. Not the right shape if you need a full brain install.
We pick a single workflow that’s either breaking or that you want AI to handle entirely. I install AI into it from the inputs to the outputs. Production-ready by the end of the sprint.
When a sprint is the right shape.
The Personal Brain Install is broad — structured corpus and queryable infrastructure across your whole working life. The Workflow Sprint is narrow — one specific recurring workflow rebuilt with AI as a first-class participant, end-to-end.
Examples: a consultancy producing a specific monthly client deliverable. A founder running the same competitive-research sweep before every quarterly board meeting. An ops lead doing a recurring weekly synthesis across multiple sources. A marketer producing a specific long-form artifact at a regular cadence.
The pattern: a workflow you know well, currently mostly by hand, with clear inputs and outputs, recurring frequently enough that the engineering investment pays back. The sprint installs the AI participant into that workflow until it’s production-ready.
How the sprint actually runs
Four weeks, fixed price.
- W1
Discovery + design
You walk me through the workflow. Inputs, steps, outputs, quality criteria mapped. AI participation shape picked. Written sprint plan you sign off on before we build.
- W2–3
Build
Agent orchestration on Workers, Claude project + skills, Cursor pipeline, Custom GPT, n8n / Zapier — whatever fits. Daily progress visible. We iterate as outputs produce.
- W4
Hardening
Edge cases. Quality calibration. Documentation. Handoff. Workflow runs in production against real inputs, calibrated to match what you would have produced by hand.
“Pricing is fixed at the start so timeline variance is on me. The structure is: I take the risk on whether the sprint ships, you take the risk on whether you actually use it.”
Discovery is bounded
Pricing fixed before any build work.
Week 1 discovery is $1,500 standalone if you don't proceed (you keep the design doc). After Week 1 I give you a fixed number for the full sprint. Most discovery weeks turn into sprints.
What you walk away with
In the bundle.
The production workflow
Live. Running against real inputs. Outputs meet the quality criteria we calibrated together.
Source files
Prompts, skills, code, configs — yours outright. Move to different infrastructure anytime.
Documentation
Plain-language docs on how it runs, what to do when it breaks, quality checkpoints, how to update inputs.
30-day support
Bugs, tweaks, edge cases included. After 30 days, support converts to Lifeline rate ($150/hr).
Pricing — why the $5K–$15K range.
The spread reflects complexity, not duration. A $5K sprint is focused, with clean inputs (e.g., a structured weekly synthesis from named sources). A $15K sprint has multiple system integrations, careful quality calibration, and human-in-the-loop checkpoints (e.g., end-to-end client deliverable production with editorial review gates).
Price is fixed before any build work — timeline variance is on me. If a $5K sprint takes 4 weeks instead of 2, that’s still $5K. If a $15K sprint lands in 3, still $15K.
Common questions
Asked, answered.
Why isn't this a brain install?
Different shape, different goal. The Personal Brain Install builds a queryable corpus across your whole working life. The Workflow Sprint builds a production workflow for one specific recurring task. Some sprints turn into Install conversations when we realize the workflow needs a brain underneath; some Install clients add a sprint later to operationalize one workflow on top of the brain.
What if the sprint doesn't work?
I’ve had one sprint in three years that didn’t ship to production — the workflow turned out to need an upstream system change that was outside my scope. The client got a 50% refund and we documented what would need to change before the sprint could run again. The structure is: I take the risk on whether the sprint ships, you take the risk on whether you actually use it afterward. So far that’s held.
Can the sprint use my existing tools?
Yes, and it usually does. The sprint outcome runs on whatever stack makes sense for your workflow — could be Claude projects, ChatGPT custom GPTs, Cursor, Workers, n8n, Zapier, or some combination. I pick the stack during Week 1 discovery based on what you already use and what the workflow actually needs.
What happens after the 30-day post-sprint support window?
Three options. Most clients move to time-and-materials support at $150/hr for occasional tweaks. Some convert to a small monthly retainer ($300-$500/mo) for ongoing care. Some self-manage the workflow because the documentation is enough.
Can multiple workflows run in one sprint?
Usually no — the sprint is designed for vertical depth on one workflow. If you have multiple workflows, they typically each warrant their own sprint, or they’re a sign you should be doing a Personal Brain Install instead. Exception: if two workflows are tightly coupled (same inputs, related outputs), they can share a sprint at the higher end of the price range.
Adjacent paths
Where this connects.
Start with the Diagnostic Call.
The Diagnostic Call commits to nothing. If this install isn’t the right shape for your work, I’ll tell you what is.
As featured in
