AI Tool Predicts Drug Success, Launches in U.S. This June

Scientists testing new medicines will soon get a powerful crystal ball that shows which experiments are worth pursuing and which ones will likely fail.

Korean AI drug developer BasGenBio is bringing its Deepcision platform to the United States through a distribution deal with Cambridge Scientific in Boston. The companies plan to finalize their paid agreement in June at the BIO International Convention.

Deepcision takes information like a drug's chemical structure and disease target, then predicts how likely it is to succeed through all four phases of clinical trials. The software analyzes over 100 clinical and biological factors to give researchers a detailed roadmap before they spend money on lab work.

"They need simulation so they can design experiments based on more precise predictions and reduce unnecessary time, materials or experiment steps," said BasGenBio CEO Tom Kim. The platform is built on data from 770,000 patients and 156,000 biobank samples.

BasGenBio tested the system by analyzing about 30 drugs, including both approved medicines and failed candidates. The AI correctly matched known outcomes in roughly eight out of 10 cases, showing about 80% accuracy at identifying which drugs would succeed and which would fail for specific reasons.

Cambridge Scientific operates biotech incubator labs and shared laboratory spaces around Boston, giving BasGenBio direct access to drug discovery startups that need to make smart decisions with limited budgets. The companies hosted a webinar this week introducing the platform to potential customers, with about 50 people registered and over 40 expected to attend.

The Ripple Effect

The partnership could reshape how early-stage biotech companies approach drug development. Instead of running expensive experiments that might lead nowhere, small startups can simulate outcomes first and focus their limited resources on the most promising candidates.

Cambridge Scientific wasn't just looking to sell equipment or supplies. The company wanted to introduce simulation technology that would actually improve the quality of experiments happening in their shared labs.

Participants in this week's webinar can apply to become beta testers of Deepcision. The platform will be offered as a subscription service, joining BasGenBio's other AI tools for target discovery and clinical trial decision support.

The company is also developing its own drug programs for hair loss, muscle loss, and other conditions. BasGenBio previously won $3.6 million in Korean government backing for its digital twin technology.

For researchers racing to find the next breakthrough treatment, this AI assistant could mean the difference between years of dead ends and a clear path to helping patients faster.

#