3 labs share how Benchling AI is saving scientists time and toil

At Benchling, we’re on a mission to put AI in the hands of every scientist. That’s why we designed Benchling AI to be directly embedded into your daily work. Throughout the process of building, we’ve collected feedback from the scientists using these tools daily.
Here’s what they shared so far: Scientists at Corteva are cutting data entry time by two-thirds; researchers at Elephas are building complex dashboards in minutes instead of days; and teams at Mammoth are reclaiming hours of mental bandwidth for the work that actually matters.
Read on to learn how they’ve implemented AI into their daily workflows and where they’re rolling out AI next.
Mammoth Biosciences cut hours to just minutes in preclinical data workflows — with no custom code or mental overhead
"It has saved many hours, but more importantly, many brain cells," says Clarissa Scholes, a data scientist at Mammoth Biosciences, a biotech company harnessing CRISPR systems for disease detection and therapeutics. Their data science team handles complex preclinical data from CROs, often delivered in inconsistent Excel files. Previously, processing these hefty files required writing and maintaining custom Jupyter notebooks just to ingest incoming data.
Now, Data Entry in Benchling automates the registration of samples, lots, and box maps using just a single uploaded manifest and a few guided prompts. It eliminates the need for custom code, even when CRO file formatting varies significantly.
Time to import data dropped from several hours to about 10 minutes, removing the taxing mental load that came with it. “This is work that I would do late at night because it doesn't feel like it should take up time in your day, and I lose brain cells over it because it's so much checking and correcting,” said Scholes. “Now I can do it in the background while I'm listening to a meeting, and it will take 10 minutes, and half of that time is just letting the model run. I don't even need to think about it."
Mammoth is also exploring more intuitive ways to support experimental design. “One of our biggest challenges is capturing early-stage screening metadata in a structured way,” said Scholes. “If researchers could just say, ‘everything is held constant, but I changed the RNA,’ and have the system infer the rest — that’s where we’re headed.”
“If there's a way to accommodate the fact that people think in sentences, and then want structured data out of the end of it, that’s the dream. The biggest problem is training biologists to work with complex digital tools. If AI can bridge that gap, it’s a game changer.”
Elephas created SQL dashboard superhumans and detected an oversight early before impact
Elephas Biosciences, a Madison, WI, based company, has developed an oncology platform that uses live fragments from a tumor biopsy to predict immunotherapy response. They needed fast, accurate ways to track lab activity and validate experimental workflows, and they needed extensive SQL experience to get the job done. Their business analyst, Justine Newhouse, handled multiple IT, data, and lab-adjacent responsibilities but had limited bandwidth and no formal SQL training.
“One of our data scientists looked at my dashboards and said, ‘Did you take superhuman juice?’”
With Benchling’s SQL Writer, Justine went from basic SQL knowledge to building complex dashboards solo — no engineering help required. “Now I can write all of my own dashboards myself, and I don’t need to refer to a SQL expert,” said Justine. “It’s been phenomenal.”
Newhouse built over 20 dashboards to answer critical questions that had been on her team's backlog. “Before, dashboard building took a couple of days, multiple hours a day across multiple SQL gurus. Now, I can build one in 15 minutes to two hours, on my own.”
Elephas also started using Benchling’s Notebook Check to support data validation for highly hands-on workflows involving human tumor samples. “It gives reviewers a checklist. Now they don’t need every brain cell firing to catch errors. Notebook Check guides them.” It flagged incorrect parent material usage that had gone unnoticed in human review, catching the mistakes before it became a widespread problem.
Corteva cuts data entry time by two-thirds and improves accuracy on critical, high-volume assays
Corteva Agriscience is a global agriculture company that develops advanced seeds and crop protection technologies. Their scientists process a high volume of fluorescence assay plates from unstructured Word documents. Each experiment requires manual formatting, entry, and error-checking across 30+ individual result tables. The time cost was substantial: it took roughly 30 minutes per run, repeated multiple times per month, plus the added risk of human error while copy-pasting data.
Corteva research associates Abbi Snell and Ayla Verschoor saw this as an opportunity for efficiency gains and tried Data Entry to expedite the process. “It was almost instantaneous… it did exactly what we needed it to do,” said Stephanie Lex, senior research associate. “For those simple but time-consuming tasks of getting data from here to here, it’s very easy to use and easy to start using.”
Through this workflow, the team can reserve their focus for the important work of validating experiment results, instead of mindlessly chasing formatting issues. Their team was able to:
Replace 30 separate result tables with a single, auto-formatted table populated via drag-and-drop of a Word file.
Reduce entry time for 30 plates from 30 minutes to 10 minutes, which adds up to significant time savings when plates are run many times each week.
Improve data accuracy by eliminating mismatched plate labels and manual data transposition.
They expect Data Entry to benefit many other teams across Corteva. “I think for our team, it's finding more ways to keep continuously being more efficient… after we've improved one thing, we should keep moving on to the next thing,” said Verschoor. They’re beginning to use other Benchling AI Features across more workflows:
NGS workflows: Automating structured uploads of dual analysis results
Excel-based templates: Removing manual transformation steps for condition tracking
Multi-schema entry: Enabling AI to route genomic, CDS, and protein sequences to the appropriate tables, replacing workaround schemas that compromise validation
Benchling AI earns its place at the bench
Fewer wasted brain cells. Saved time. More confident decision-making. Teams using AI in the lab are opening up space for creativity, curiosity, and faster breakthroughs.
Now generally available in Benchling: Notebook Check and SQL Writer are now available for all Benchling users. And for teams working in regulated environments, SQL Writer is now available in Validated Cloud, making it safe for even the most compliance-sensitive workflows.
Want to see what Benchling AI can do for your team? Request a demo.
Powering breakthroughs for over 1,300 biotechnology companies, from startups to Fortune 500s