AI Code Assistant Company Tabnin today launched a code review agent that aims to help developers adhere to their organization’s best practices and standards. It allows organizations to codify these rules either by providing documentation to the agent or by pointing them to a set of “gold code repositories”.

The agent will passively review the code while the developer works in the IDE, reporting issues as it finds them, and proposing fixes where possible.

As President and Chief Marketing Officer of Tabnine Peter Guagenti told me, many companies – and the developers they employ – don’t even know the rules and best practices contained in their own documentation.

While one of Tabnine’s main pillars has always been customizing agents based on an organization’s needs, the company has also partnered with companies like Redis to collect some of their best practices and pre-train its models on this subject. Starting today, Tabnine is opening it up to other providers who also want to provide their rules.

“Database companies are a great example,” he said. “Everyone has their own model. Everyone has their own ways of working that have good and bad results – and when the result is bad, they blame the product, not the code, right? So we think this is a great opportunity for AI to correct behavior and make products perform better.

Developers can turn these pre-trained rules on and off as needed.

As for the accuracy of the suggestions, Tabnine says his review agents read the code like a human would. This also means that if the code is really obscure, it might miss a problem. But unlike human reviewers, it will read every line of code and not just skim.

It’s surely no coincidence that Tabnine is making today’s announcement on the first day of GitHub’s Universe conference. After all, GitHub’s Copilot surely holds the majority of brand awareness among AI coding tools. But Guagenti doesn’t seem too worried about competition with GitHub.

“We think activity will stratify,” he said. “We think people like Cursor and others are going to gobble up the bottom end of the market because a lot of people don’t want to write code. We think Copilot should target the big middle – it should be 80% for as many people as possible. However, this was not the vision of Tabine co-founders Dror (Weiss) and Eran (Yahav). Their vision was: how can we make the world’s top 1,000 engineering teams more productive, more efficient, and write better applications? This vision is now starting to really come to fruition.