Reading Time • 2 min read
B2B AI writing software serves 16,000 customers using CI
B2B AI writing software serves 16,000 customers using CI

Draft is a platform that pairs artificial intelligence with a marketplace of vetted writers to make creating content seamless and easy. Learn what the process of incorporating the CI process was like, and tune into the thoughts of Draft's full stack engineer.

What is

Draft is a YC-backed platform that pairs artificial intelligence with a marketplace of vetted writers to create high quality content, fast. Draft serves over 16,000 customers, including Opencare, Zapier, and FreshBooks.

Screenshot of's landing page.

We sat down with Mateus, a full stack software engineer at Draft who was responsible for implementing for CI that scales.

Watch the full video testimonial below:

Implementation Process:

Mateus found incredibly easy to navigate as testing infrastructure. He stated the following: "I didn't need to learn a new pattern of this; the implementation method was something I was already familiar with!"

Within a few short hours, Mateus saw improvements in testing speed and was amazed by how each feature could be stored in a single layerfile. This enabled his teammates to work more productively without being hindered by traditionally complex and long YAML files.

Tech Stack:


  • Github Actions for basic CI/CD automation


The following are some issues Draft encountered with their current solution:

  1. Making sustainable code and balancing tradeoffs with quality and speed.
  2. Scaling their development staff at the desired rate, preventing queuing or wait times from hindering their progress.
  3. Sharing ongoing projects with various stakeholders—often those without technical backgrounds—can be difficult.


  • Tech stack remained the same, no changes needed. solutions:

  1. The caching feature of enables setup steps to be skipped based on the files read, preventing time-wasting test runs and resulting to quicker builds.
  2. An advanced CI experience for the same stack, with faster testing throughout the pipeline.
  3. creates per-commit preview environments, which allows team members to avoid the difficult and lengthy process that was previously in place to share work.

Why did they choose

From the perspective of Mateus, a full stack engineer at Draft on implementing CI:

“It was nice because a lot of things we had to write, to Layerfiles, things like that, were very similar to Linux and Docker. We didn't have to learn a completely different standard to configure something in CI, because it was something we already knew.”

- Mateus, Full stack software engineer

Want to supercharge your engineering team too? Try for free here

Last Updated • Nov 02, 2022