Questions, answered
Everything we get asked on the first call, written down so you can skip to the part that matters.
What does it cost?
Projects are fixed price, agreed in the proposal before any work starts. The embedded AI expert is a monthly retainer from one day a week to full time. Most first projects land between 15,000 and 80,000 USD; retainers start around 6,000 USD a month.
Will you train on or expose our data?
No. Your data stays in your accounts. We use model providers under zero-retention terms, can deploy inside your own cloud, and sign your NDA before the first call if you prefer.
Who owns what you build?
You do. Code, models, prompts, documentation and accounts are yours from day one. We work inside your repositories so there is nothing to hand over at the end.
How long does it take?
A 30-minute call, then a proposal in 5 days. A first working AI agent in 3 to 5 weeks. A mobile or web product v1 in 8 to 12 weeks. A data platform migration in 6 to 12 weeks. Timelines are in the proposal and we hold ourselves to them.
What happens after launch?
Every project includes a handover and 30 days of fixes. After that you can run it yourself, keep us on a light retainer for monitoring and improvements, or bring the embedded expert in to keep shipping.
Which AI models do you use?
Whatever fits the task and your constraints. We work with Claude, OpenAI and Gemini models, and open models when data must stay in your environment. We measure quality, cost and latency on your real data before choosing.
Do you work with our existing engineers?
Yes, and we prefer it. We pair, review and document so your team can carry the work forward. The embedded expert offer exists precisely for teams who want to learn by shipping.
Snowflake or Databricks?
Snowflake when analytics and simplicity come first. Databricks when you have heavy Spark, machine learning or streaming workloads. We build on both and will tell you which fits, including when the answer is to stay where you are.
Where are you based?
We work globally and overlap with European and US working hours. Calls are on Google Meet or Zoom; work happens in your Slack, Linear or Jira and your repositories.
What if we only need advice?
Book the call anyway. If a two-week audit is the right first step, we will say so and price it as one.