How it works
How an engagement runs.
A conversation first. Then, once engaged, three phases in the same order: understand the work and design it on sample data, build it in your AWS account, hand it over. Nothing real is connected until you have seen it working.
The phases
Thirty minutes, no pitch
You describe what is slow, manual or untrustworthy, and which systems are involved. We ask who touches the information, where it goes stale, and what decision it feeds. You hear whether Birdrock can help and what an engagement would look like.
You leave with: a straight answer.
Understand, then design
The first phase of every engagement. We sit with the people who do the work, map the systems and the handoffs between them, and find where information goes missing or goes stale. Then we draft the screens on sample data and put them in front of the people who will use them, until they are right.
You own: the designs, the system map and a written scope for the build.
In your AWS account, from day one
The build happens in AWS, in an account that belongs to you. You see progress on the real screens as it goes. Live systems are connected only after you have approved the design and the safeguards are in place.
You own: everything, as it is built: code, data, account, documentation.
Handed over, or looked after
Monitoring, updates and changes as the business moves, under a care plan you can end any month. Or take the documentation and run it yourself. Birdrock has no standing access unless you grant it.
You keep: the choice.
Questions people ask before the call
Straight answers
Who owns the code and the data?
You do, from the first line of code. It lives in a repository and an AWS account in your company's name. Birdrock keeps only the right to reuse its general-purpose tooling, nothing specific to you, and there is nothing to buy out later.
What happens if Birdrock is not available?
Everything keeps running, because it runs in your account, not ours. The documentation is written for the next engineer and the setup is written as code, so any competent team can take it over. That risk is real with any vendor; the answer is design, not reassurance.
Can this handle regulated data, such as patient records?
No software is compliant by itself; the setup is. The work runs in your own AWS account under your own AWS agreements, builds are designed so that access to sensitive records is logged, and the design phase uses no sensitive data at all. Your compliance lead gets the architecture in writing before anything live is connected.
Where does the AI run and who sees our data?
Through Amazon Bedrock, inside AWS in your region, under AWS terms that prohibit training on your data or retaining your prompts. It is never passed to a third-party AI company. Drafts are reviewed by a person before they are sent or filed.
Do you use AI to write the code?
Yes, as a tool. The code is reviewed and tested by an engineer before it ships. You get the speed without an unreviewed pile of generated code. Tests and documentation are part of the deliverable.
What if we want to stop partway?
You keep everything built so far, in your account, with the documentation to that point. The roadmap stands on its own, and the care plan ends whenever you say.
Are we too small for this?
Birdrock works with businesses of roughly ten to two hundred people. If the problem turns out to have a ready-made answer, you will hear that on the first call.
Start with the call
Thirty minutes. Describe the problem and the systems involved, and hear whether an engagement makes sense.