Benchling cuts migration time by 75% with cloud agents

Hi, I'm Emily. I am a software engineer on the developer productivity team at Benchling. We are really trying to

Benchling. We are really trying to advance the pace of life sciences development. We really have this culture

development. We really have this culture of experimentation especially with AI.

Cursor really stood out right away because of its customizability and extensibility as compared to some of the other tools. We have a very complex

other tools. We have a very complex developer environment. Think lots of

developer environment. Think lots of requirements, custom Docker files, that's all running locally, but cloud agents in particular, this is what we're most excited about. We're able to lift

that entire local development up into the cloud. We have thousands of old

the cloud. We have thousands of old enzyme tests in our codebase. We need to migrate them to RTL and it's blocking a major React upgrade for us. It was

moving very slowly. We were quoting about a year to get this entire migration done. But what we've tried out

migration done. But what we've tried out in the past quarter is we've rolled out cloud agents that's orchestrate hundreds of agents that we spin off and do

migrations for us. Now the scale and the rate at which we're converting our RTL tests has increased significantly to the

point we are quoting about one quarter to finish this rather than an entire year. We've really sort of built up a

year. We've really sort of built up a whole developer experience around cloud agents. 30% of our merged PRs are

agents. 30% of our merged PRs are actually from cloud agents. Now we're

able to support that with skills in context in order to match the quality we still expect. Cursor is really helping

still expect. Cursor is really helping advance a lot of the speed and the quality piece and the composer models have been very costefficient in that

regard and still giving us a great output. We actually give developers the

output. We actually give developers the option to choose out of a set of AI tools and to choose out of most

available models. We have seen that

available models. We have seen that people have gravitated towards cursor because of its cloud agent capability.

Each quarter we run a development survey. Last quarter we rolled out

survey. Last quarter we rolled out cursor cloud agents. What we saw from last quarter to this quarter was a 20% jump in AI developer satisfaction. We

also saw that 95% of our developers are satisfied with our AI tooling right now.