How Google uses AI

The Chrome security team uses Gemini based agents across the life of a security bug. An automated triage pipeline filters spam and duplicates, reproduces bugs, adds severity and routes them to the owner; fixing agents propose candidate patches that a critic agent reviews, and test writing agents add tests before a developer evaluates the fix. Chrome fixed 1,072 security bugs in milestones 149 and 150, more than the prior 23 milestones combined, and Google says LLMs now generate candidate fixes for most vulnerabilities. It estimates the triage automation saves hundreds of hours of developer time a month.

Google used an internal LLM based system to help engineers with large scale code migrations, such as changing identifier types from int32 to int64. Engineers doing the migrations estimated the total time spent was reduced by about 50%, and reported that 80% of the code changes in landed changelists were AI authored, with the rest written by humans.

Google site reliability engineers use an agent in the Gemini CLI across an outage: reading the page, investigating, proposing mitigations, finding the root cause and drafting the postmortem. Every proposed change passes a policy layer (for example rules that need two person approval) and a forced human confirmation, and every proposal and approval is logged. No outcome figures are published.

Google's security engineering team built a pipeline that prompts Gemini to generate code fixes for bugs that sanitizers find during unit tests in C and C++, Java and Go code, such as uninitialised values, data races and buffer overflows. Every generated fix goes to a human reviewer before it lands. Google reports that the pipeline fixed 15% of these bugs, hundreds in total, and expects the rate to improve.

Google ships large volumes of high value, time critical hardware to strategic locations across hundreds of facilities worldwide. It uses Everstream Analytics' AI powered platform to collect real time signals on incidents that could affect its transportation network and to highlight which events could have critical outcomes for inventory in transit, giving logistics staff, carriers and internal customers faster and more accurate information. Everstream reports that with the system in place, Google managed three major incidents over two months without disrupting wider operations.

Mike Chen, Director of Product Strategy, leads sales for Google's Ads team serving small and medium businesses, with thousands of sellers worldwide. The team piloted Gong's Scorecards to have managers rate call effectiveness, then trained Gong's Smart Tracker, an AI model, on that baseline to spot best in class calls at a scale managers could not review by hand, building a Call Library used to train new hires and refresh the rest of the team.

Automating tasks that sellers used to do by hand is estimated to give thousands of sellers back about 12% of their time each year, and the team also reports a higher new hire graduation rate and a weekly revenue increase in one sales segment, without a figure disclosed for either.

Google and DeepMind applied a neural network, trained on historical turbine data and widely available weather forecasts, to 700 megawatts of wind power capacity across a group of wind farms in the central United States. The model predicts wind power output 36 hours ahead of actual generation, and recommends the optimal hourly delivery commitments to make to the power grid a full day in advance, so the wind fleet can be scheduled like a conventional generator instead of treated as unplannable.