The workflow spans literature analysis, candidate design, experimental planning and industrial decision support, extending AI4S into materials research
SHENZHEN, China, Oct. 31, 2026 /PRNewswire/ — Wenge AI demonstrated an AI-assisted hydrogen-production catalyst design workflow supported by ScienceOra at the 2026 Tencent Global Digital Ecosystem Summit. The demonstration showed how AI for Science (AI4S) can connect literature analysis, catalyst candidate design, experimental planning and industrial decision support, helping materials researchers translate scientific evidence into actionable research plans.
ScienceOra is Wenge AI’s intelligent research decision platform, built on the ScienceOne Scientific Foundation Model. For specialist research tasks such as materials development, the platform brings together scientific models, professional literature, research data and tools to support information access, research-plan development and scientific decision-making.
ScienceOra plans and executes a five-step catalyst design workflow
From literature analysis to catalyst candidate design
The demonstration focused on the design of a high-entropy alloy catalyst for the alkaline hydrogen evolution reaction. To address the need for high activity and durability, ScienceOra searched professional literature and drew on published research methods and elemental properties to support the development of candidate designs. It also cross-checked optimization proposals against relevant literature, providing researchers with references that could be reviewed and verified.
Rather than simply generating material formulations, the workflow connects documentary evidence, materials knowledge and specific design objectives to support more targeted analysis of candidate solutions. Researchers remain responsible for assessing scientific validity and determining subsequent research directions.
From experimental planning to industrial decision support
For experimental planning, ScienceOra can assist in developing research protocols around candidate designs, helping researchers organize experimental variables and producing testing standard operating procedures (SOPs) and related reference information for subsequent validation.
For industrial applications, the platform can also support comparative analysis across dimensions such as the feasibility of scaled hydrogen production, performance at industrially relevant current densities, durability, raw-material availability and compatibility with mass-production synthesis processes. It can then assist in establishing decision priorities, screening criteria and process-improvement recommendations.
Candidate designs, experimental plans, process recommendations and references can be consolidated into a structured research report, helping R&D teams assess scientific evidence alongside engineering constraints. AI provides analytical and decision support, while experimental testing and scientific validation remain essential parts of materials research.
Previously disclosed collaboration reports design efficiency and activity improvements
In a previously disclosed industry–research collaboration on hydrogen-production catalyst design, ScienceOra supported literature analysis and research-plan development. The collaboration reported a 38% improvement in catalyst activity and reduced the design stage from several months to 30 minutes. The case illustrates the potential of domain-specific AI to connect scientific literature, research design and engineering requirements.
ScienceOra was recently made available on Tencent WorkBuddy, with an initial offering of more than 30 research expert agents and skills covering literature analysis, research hypotheses, plan design, modeling and simulation, validation analysis and research communication. The conference demonstration further illustrated how ScienceOra can be applied to specialist materials research, as AI4S moves beyond general information processing toward support for research workflows and industrial R&D decision-making.
For more information, visit https://www.scienceone.ai/portal/.
About Wenge AI
Wenge AI (1956.HK) is a provider of artificial intelligence technology and service platform products for enterprises. With self-developed large models as its core, the company helps enterprises build practical intelligent platforms and applications. Founded by a team of AI scientists from the Chinese Academy of Sciences, Wenge AI has deep technological expertise in large models, cognitive intelligence and complex-system modeling, and is committed to advancing AI from technological capabilities to enterprise-level applications.
