Services

Fixed price, fixed duration, one agreed outcome metric.

Start with a free half hour. Everything after it has a price and a duration.

00Free · 30 minutes

Office Hours

A live 30-minute session working through your actual problem. No pitch, no deliverable.

01 / Offers

Four offers.

01

Discovery

Agent Readiness Teardown

Two weeks in your repository. You get a working prototype agent on your code and a 90-day plan.

from $4,5002–3 weeks · fixed price
  • A written assessment of where agents will and will not work in your codebase
  • An evaluation harness set up for one workflow
  • A working prototype agent on your code
  • A 90-day plan

Half the fee credits toward a sprint signed within 30 days. If the Teardown does not find work worth more than the fee, you do not pay.

026–8 weeks · fixed price

“Our team is slow on an old codebase and the AI tools didn’t stick.”

Harness Sprint

We install the evaluation rig, context pipeline, review gates, and CI integration inside one team, and measure one outcome metric agreed before signing, using your own instrumentation.

from $28,000
  • Evaluation rig running against your real workflows
  • Context pipeline: repo maps, retrieval, tool access
  • Review gates before anything reaches main
  • CI integration and one measured outcome metric

You own all of it. No license, no lock-in.

026–10 weeks · fixed price

“We need one AI capability shipped into a system we already run.”

Build Sprint

One AI capability shipped to production inside an existing system: search over your documents, an agent that uses your internal APIs, or a workflow with evaluations and a human review step.

from $30,000
  • The capability, in production, inside your codebase and CI
  • Evaluations so it can be changed without being broken
  • A human review step where the output matters
  • Runbook and handoff

Regulated systems (audit trail, retention policy, SSO and role-based access) are a priced tier.

03

After a sprint

Embedded

Ongoing forward-deployed engineering, one to three days a week: embedded in your team, shipping to production, responsible for an outcome.

$9,000–24,000/mo3-month minimum · rolling quarterly
  • Advisory, about one day a week: office hours, design review, escalation
  • Build, about two days a week: shipping against an agreed backlog
  • Both founders, about three days a week combined

02 / Process

How an engagement runs.

  1. 0030 minutes

    Office hours

    Your actual problem, and a straight answer on whether we are the right call.

  2. 01Weeks 1–2

    Teardown

    We work in your repository: map the codebase, choose one workflow, set up an evaluation harness for it, and get a prototype agent running on your code. You get a written assessment and a 90-day plan, and we state the sprint price out loud before any proposal is written.

  3. 02Weeks 3–10

    Sprint

    The outcome metric is agreed before signing. The evaluation rig goes in first, so everything after it is measured. We build inside your CI, behind your review gates, and hand off with a runbook.

  4. 03Monthly

    Embedded

    The engineers who built it stay on: advisory at about a day a week, or hands-on shipping at two. Three-month minimum, then rolling quarterly.

03 / Fit

Who this is for.

A prospect fits when the buyer has an established codebase, a mandate to act, and the technical background to evaluate our work directly.

A fit

  • 50–500 employees, roughly 15–150 engineers
  • A codebase five or more years old: monoliths, legacy systems, migrations in progress
  • A software company, or an operations-heavy business with a real internal engineering team
  • AI seats bought or a pilot run, adoption stalled on the main codebase, leadership asking for results
  • A VP Engineering or CTO who can approve a $4,500 engagement without a procurement process

Not a fit

  • Fewer than ten engineers, or a new product with no existing system
  • Consulting bought through a formal RFP process
  • A strategy deck, roadmap, or training program without a build attached
  • Already committed to a model vendor's own deployment program

04 / Prices

Price list.

Tiers scale with team size, repo complexity, and how regulated the work is.

  • Office HoursFree

    Live technical session, no deliverable.

    30 min

  • Teardown · Standard$4,500

    Single repo, up to 50 engineers. Assessment, evaluation scaffold, working prototype, 90-day plan.

    2 weeks

  • Teardown · Extended$7,500

    Multiple repos or 50–200 engineers. Adds a cross-team adoption plan.

    3 weeks

  • Harness Sprint · Scoped$28,000

    One team, one repo. Evaluation rig, context pipeline, CI integration.

    6 weeks

  • Harness Sprint · Standard$32,000

    Adds review gates, team enablement, and a measured outcome metric.

    6 weeks

  • Harness Sprint · Multi-team$40,000

    Two to three teams, shared foundation, training for internal champions.

    8 weeks

  • Build Sprint · Scoped$30,000

    One workflow, one integration, evaluation harness and human review step.

    6 weeks

  • Build Sprint · Standard$38,000

    Two to three integrations, production hardening, runbook and handoff.

    8 weeks

  • Build Sprint · Regulated$45,000+

    Adds audit trail, retention policy, SSO and role-based access, subprocessor documentation.

    8–10 weeks

  • Embedded · Advisory$9,000/mo

    About one day a week. Office hours, design review, escalation.

    3-month min.

  • Embedded · Build$16,000/mo

    About two days a week. Shipping against an agreed backlog.

    3-month min.

  • Embedded · Both founders$24,000/mo

    About three days a week combined.

    3-month min.

  • Scale modifierRetainers for organizations above 150 engineers.+$2,500/mo
  • Compressed timelineA client-imposed deadline that shortens a sprint.+25%
  • Additional team cohortEnablement for a second team of up to 12 people.$6,500
  • Architecture reviewHalf-day, standalone.$2,500
  • Post-engagement supportA 30-day support period after a sprint without a retainer.$4,000/mo
  • Out-of-scope workChange requests. Everything else is a fixed price.$225/hr · $1,800/day
  • Outcome bonus (optional)A portion of the fee at risk against the agreed metric, offered when we set the metric.15–25% of fee

05 / Terms

Terms.

  • Fixed price, fixed duration, one agreed outcome metric.
  • You own everything we build. No license, no lock-in, no dependence on a single model vendor.
  • Half the Teardown fee credits toward a sprint. If the Teardown does not find work worth more than the fee, you do not pay.
  • Hourly rates apply only to change requests.
  • A Harness Sprint costs about the same as five to eight days of a large consultancy’s senior architect.
  • We do not sell audits, roadmaps, or training without a build attached.

06 / Questions

Common questions.

Which model do you use?
Whichever scores best on your evaluation suite. We are independent of every model vendor, and we swap the model when a better one ships. The evaluations make that a measured decision rather than a preference.
Who does the work?
Both founders, who are practicing senior engineers. The people on the call are the people in your repository. Nothing is handed off.
How is pricing structured?
Fixed price and fixed duration, with one outcome metric agreed before signing and measured in your own instrumentation. Hourly rates apply only to change requests.
What do we own at the end?
Everything: the evaluation suite, the context pipeline, the review gates, the prototype, the runbook. It lives in your repository under your license, with no runtime of ours to keep paying for.
We already bought AI coding seats and usage dropped off. Is that a problem?
It is the usual starting point. Seats without an evaluation suite, a context pipeline, and review gates rarely hold up on a large codebase. The Harness Sprint exists for exactly that.
Do you need access to our code?
Yes. The Teardown happens inside your repository, under your access controls and your NDA.
Where are you?
Utah. The work is remote, across US time zones.

Start a conversation

Book office hours.

30 minutes, no pitch. We look at your actual problem.