Free for open source projects and schools. AI Council costs nothing for OSS maintainers and educational institutions. Apply →
Get started
Menu
How it worksFor engineering leadersCompareCase studiesPricingDocsBlogFree for open source & education Get started

Home / Use cases / Onboarding

New developers are productive in a legacy codebase from day one

A new hire usually spends 2–3 weeks learning the project and its environment, then another 2–3 weeks at half speed, asking colleagues all the time. With AI Council, they and their AI agent get the team's decisions and rules while working on a specific file.

Get startedRun your own numbers

Why getting into a large project takes a month

  • Knowledge lives in people's headsOnly one or two people know why a module is built the way it is. The docs lag behind the code.
  • The agent is new tooAn AI assistant reads the project from scratch every time and doesn't know the team's conventions. So code from the new hire and their agent keeps coming back from review.
  • Senior developers get pulled awaySenior engineers spend hours answering a new hire's questions, and it starts over with every new person.

How AI Council solves it

  1. The project is analyzed before the new hire arrives

    The code and knowledge graph is built from the repository and tickets. Decisions are tied to modules and functions.

  2. Context at the moment of work

    When they open a file, the new hire and the agent get related decisions, mandatory rules and similar past tasks.

  3. Questions go to the project chat

    The answer is built from the code and the knowledge base, then saved. The next new hire gets it right away.

  4. Checks before review

    The diff is checked against project rules, so first MRs don't come back because someone didn't know the conventions.

From a customer project

1 day

instead of 2–3 weeks of ramp-up and another 2–3 weeks at half speed

On customer projects, a new developer picks up tasks on day one. When several developers and agents work on the same project, they save 40–60% in time and tokens: knowledge gathered in one task helps with the next.

All case studies

What you see in the first weeks

  • A new hire's first tasksWhen the new person starts taking tasks, and how often their MRs come back from review.
  • Agent sessionsWhich knowledge and rules the new hire's agent got while working with files.
  • Project chat questionsWhat people ask, and which answers are already saved as knowledge.

Try it on your hardest project

Record a baseline before you connect, then compare after 4–6 weeks.

Get startedSee pricing
Made on
Tilda