How we work

A dedicated engineering team you can think of as yours.

A team that works with one operator-founder at a time and ships production AI. Start with a 2–4 month pilot or go straight to a long-term partnership.

Process

From first call to production, in five honest steps.

  1. 01

    Discovery call

    45 min

    We talk about what you're building, who it serves, what you've tried, and where engineering is bottlenecked. No deck. No pitch. We ask a lot of operator-specific questions.

  2. 02

    Proposal & alignment

    Within 1 week

    We send a written proposal: scope, team composition, monthly retainer, success criteria for the first quarter. If the fit isn't right, we say so plainly and point you elsewhere.

  3. 03

    Team setup

    1 week

    We assemble the team from people already at BroutonLab. You meet them, see their work, sign off. They become your team — not a rotating pool.

  4. 04

    Sprint 0 — the hardest problem first

    2 to 4 weeks

    We sit down with you — and a couple of your customers — to understand the domain. Then we build a working slice of the hardest technical question, so everything after is de-risked.

  5. 05

    Build & ship

    Ongoing

    Bi-weekly demos. Monthly strategy review with you. Quarterly retros. The team ships in production, runs the system in production, iterates from real-user signal.

Start here

The pilot.

A long-term commitment to a team you just met is a hard ask. So don't make it yet.

Start with a scoped pilot: 2 to 4 months of custom AI development, scope and budget agreed upfront, no obligation to continue. We pick one hard, valuable slice of your product — an agent, a matching engine, a voice flow — and ship it to production.

If it works out, the same team stays and the pilot becomes the first quarter of a longer partnership. If it doesn't, you keep everything: code, infrastructure, documentation, a clean handover. Most of our long-term engagements started exactly this way.

Engagement options

Pick what your stage needs.

Pilot

$30–50K fixed · 2–4 mo
2–3 engineers
1 tech lead + 1–2 backend/ML engineers

One hard slice of your product, shipped to production against a written scope of work. You pay exactly what was agreed — no surprises, no obligation to continue.

Compact

$15 – 25K/mo
3 engineers
1 tech lead + 2 senior backend/ML engineers

Early product, validating with 5–20 design-partner customers. You need shipping muscle, not headcount.

Standard

$35 – 45K/mo
5 engineers
1 tech lead + 2 backend engineers + 1 ML engineer + 1 MLOps engineer

Product is taking shape, you have 20–100 users, the AI surface is real (agents, ranking, ML).

Full team

$55 – 65K/mo
7–8 engineers
1 tech lead + 3 backend + 2 ML + 1 MLOps + 1 PM/designer

Product is mid-stage, you're scaling, the AI is core to the offering and needs continuous improvement.

Questions you'll probably have

Plain answers.

Why dedicated team, not fixed-bid?
Vertical AI products are not specifiable upfront. The interesting questions only emerge once real users touch the system. Fixed-bid forces premature certainty. Dedicated team lets us learn together, in production, every two weeks. When you do want a bounded start, that's what the pilot is for — one slice small enough to scope honestly.
Can we start smaller before committing long-term?
Yes. A scoped pilot — 2 to 4 months, scope and budget agreed upfront — is often the right first step. You get a production-grade slice of your product either way, with a clean handover if you stop there. No obligation to continue. Partnerships that continue do so because the work earned it.
How quickly can we start?
Typically 1–2 weeks from signed proposal to team kickoff. We don't keep a bench, so timing depends on what's wrapping up. We tell you the realistic start date in the proposal — never sell capacity we don't have.
Who builds the UI?
The same engineers, with AI-assisted frontend tooling. In 2026 a dedicated frontend engineer is rarely where a vertical AI product wins or loses — so we staff backend, ML, and MLOps depth instead, and the UI gets built along the way.
Do you work with non-technical founders?
Most of our clients are operator-founders without an engineering background. You bring industry knowledge and the product vision. We bring engineering. Together we make sure the technical choices serve the business choices, not the other way around.
What happens if it's not the right fit?
We end the engagement cleanly. There's a 30-day notice clause both ways. Code, data, infrastructure, documentation — all yours. We hand off properly. We'd rather end early and stay friendly than waste your money.
Do you sign NDAs? Do we own the IP?
Yes to both. Standard mutual NDA at the start. All code, models, data, and IP belong to you from day one — we work as an extension of your team, not as a licensor.
Do you support what you build after launch?
Yes — that's the model. Engagements are 12–24+ months because production AI needs continuous iteration. If you ever want to bring it in-house, we help you hire and transfer cleanly.
What time zones does the team work in?
The core team overlaps with US and EU business hours. We default to bi-weekly synchronous demos and async-first day-to-day, so your team isn't blocked waiting for ours.

Ready to talk about what you're building?

Tell us about your product