ComAct: New Paradigm for AI Software Agents Using Component Object Model
Researchers introduced ComAct, a new approach for AI agents to interact with professional software by using the Component Object Model (COM) as executable actions rather than visual GUI manipulation. The method addresses limitations of existing GUI-based and API-based agents, particularly for complex industrial software like CAD programs. This work demonstrates significant performance improvements and could enable more reliable AI automation of professional software tasks.
A new research paper proposes ComAct, a paradigm shift in how AI agents interact with professional software applications. Rather than relying on fragile visual grounding (GUI-based approaches) or struggling with heterogeneous APIs, the method treats the Component Object Model (COM)—a Windows technology for inter-process communication—as a unified executable abstraction for software manipulation. The researchers validated this approach by creating ComCADBench, the first benchmark for AI agents operating real industrial CAD software, and developed ComActor, a self-correcting agent trained through a progressive three-stage framework. Experiments showed that while frontier proprietary models achieved near-zero success with GUI-based interaction, COM-based execution yielded substantial immediate gains. ComActor achieved state-of-the-art performance on the benchmark with strong resilience in long-horizon tasks and demonstrated generalization to external CAD benchmarks.
What's missing
The paper does not discuss potential limitations regarding applicability beyond Windows environments, compatibility with non-COM-based software, or practical deployment challenges in enterprise settings. Additionally, the scope of evaluation appears limited to CAD software; generalization to other professional software categories remains unclear.
What different sources said
- arXiv cs.AICenter
ComAct: Reframing Professional Software Manipulation via COM-as-Action Paradigm
Related
Topology-Aware Thermodynamics Improves DNA Probe Specificity Design
Researchers developed a new framework for designing DNA probes that accounts for the spatial organization of matched sequences, not just overall thermodynamic stability. Traditional methods rely on scalar measures like melting temperature and free energy, which miss how mismatches are distributed along the probe. The approach could improve diagnostic accuracy in applications like HPV detection and gene expression profiling.
Study Identifies Optimal Thermal Dose for Combining Focused Ultrasound with Immunotherapy in Tumors
Researchers used multimodal PET imaging to identify an optimal thermal dose range for focused ultrasound ablation that destroys tumor tissue while preserving conditions for immunotherapy delivery. The study found that excessive heating collapses blood vessels needed for antibody access, while insufficient heating fails to adequately reduce tumor burden. The findings could guide clinical design of combination treatments pairing thermal ablation with immunotherapies.
Plant MSH1 Protein Functions as Mismatch-Directed Nuclease for Organelle Genome Maintenance
Researchers have identified the precise mechanism by which the AtMSH1 protein in Arabidopsis plants recognizes and cleaves DNA mismatches and lesions, preventing mutations in organellar genomes. The protein combines a DNA mismatch recognition module with a nuclease domain that makes staggered cuts at specific positions relative to DNA damage. This discovery explains how plants maintain unusually low mutation rates in their mitochondrial and chloroplast DNA compared to other eukaryotes.