Boston, USA / Tel Aviv, Israel, April, 2026 — Converge Bio, the Generative AI Lab for the Life Sciences, shares findings from a new study showing that ConvergeAB™, their antibody design and optimization solution, was able to improve cetuximab, a widely used cancer therapy, generating a sequence that was later filed as a provisional patent. Originally conducted as an internal test of the platform’s capabilities, the experiment was successfully completed in just eight hours without any task-specific training, manual tuning, or dedicated R&D campaign. The Converge-designed antibody showed binding to epidermal growth factor receptor (EGFR) at double the strength of both cetuximab and the competitor’s antibody, suggesting it could more effectively block cancer-driving signals.

Cetuximab is commonly used to treat solid tumors such as colorectal, head, and neck cancer by targeting the extracellular domain of EGFR. By binding to the receptor with high specificity, cetuximab prevents receptor activation and downstream signaling, slowing the cancer cells growth. Its effectiveness depends on how well it binds, and stronger binding can lead to better patient outcomes. 

To do that, Converge applied ConvergeAB™ in a zero-shot setting, using nothing but cetuximab’s antibody and the target receptor sequence as inputs. From a single prompt, the platform generated 100,000 antibody candidates and identified a design that consistently outperformed them in experimental validation, including SPR (surface plasmon resonance) and kinetic characterization for precise KD determination, achieving an average binding affinity 2.1x stronger than cetuximab and 4.4x stronger than a competing antibody source. Converge’s sequence contained six edits from the original sequence, distributed across both framework and CDR regions, highlighting the platform’s ability to make targeted, high-impact changes in both sections. 

The findings suggest that AI can play a critical role in improving existing therapies. In this case, AI-powered ConvergeAB produced an upgraded version of an established cancer therapy without a human in the loop, and only 10 top sequences had to be tested in the lab, offering a more efficient path to refining existing drug offerings, with less reliance on traditional trial-and-error methods. The work reflects Converge Bio’s broader vision of building a “computational lab of the future,” where AI can simulate, test, and refine ideas digitally before they reach the lab. This would significantly reduce the time, cost, and experimental burden traditionally associated with drug discovery. 

“This work began as an internal experiment to see how far our platform could go, and our results reinforce what is possible in therapeutic development through AI,” says Dov Gertz, CEO and Co-Founder of Converge Bio. “What stands out is both the quality of the results and the pace at which we were able to reach them. In just eight hours, we moved from a single prompt to a set of candidates that demonstrated meaningful improvements, giving a solid example of how AI can compress what has traditionally been a long and intensive process into a far shorter timeframe.”

Read the article on Forbes