- Forbes Technology Council argues that AI-vetted procurement for legacy companies is regulated by issues with siloed architecture in their data.
- ZDNET outlines PRCA’s approach: tests with gen AI models before any code is submitted, including what happens after the results are generated, before the code is forwarded to Zencoder for documentation to support business-logic modernization.
- PRCA plans to complete its migration from AS/400 to the end of 2026, and then proceed with the next step of moving to ASP.NET Web Forms, implementing AI-assisted process automation.
In materials from Forbes Technology Council and ZDNET, it is explained that companies can modernize their systems only if they address AI deployment issues in their legacy architecture. The focus of the publication is on problems of legacy architecture with data, technical debt, and the need to test AI-based processes on the basis of existing AI.
In materials from Forbes Technology Council, it is noted that companies are “stuck” with siloed legacy architecture, where data cannot be processed across the organization. The author of the material also points out that, although the models are trained on new data, they can still be limited by the lack or incorrectness of data that is needed to complete the downstream “downstream exposure” process.
ZDNET reports that it has reviewed the Professional Rodeo Cowboys Association (PRCA) case, in which, according to the materials, the organization has used AS/400-based systems for nearly 100 years and data from backend systems for about 40 years. In a statement, CTO Jeff Love said that, rather than starting from scratch, he and his team used gen AI models to test the possibility of using new tools, and then to evaluate the results. ZDNET also notes that Jeff Love and his team used Zencoder as an agentic platform, which analyzes business logic in plain English. The materials also state that the company provided an AS/400 codebase to Zencoder for documentation, describing the existing files and business logic flows in the code. The author also says that the first results were “far from ideal,” and that the company then refined the process.
ZDNET adds that Jeff Love’s team built agent instructions, guidelines, diagrams, and a workflow to help PRCA’s business analysts. In addition, the organization created wireframes based on the existing business logic, and also ran unit tests, which, according to the materials, helped catch issues before the business logic modernization system went live.
Overall, ZDNET notes, Zencoder reduces the time needed for code review by 50%, which the IT company says helps it support critical business services and instructions for users. The publication also states that PRCA plans to complete its migration from AS/400 by the end of 2026, and then implement the next step—moving to the legacy platform, ASP.NET Web Forms. The materials also note that, with the help of Jeff Love’s approach, the company is building a pathway to integrate agentic processes into its existing workflows.
Forbes Technology Council concludes that AI modernization platforms can help organizations modernize how AI is used for code writing, and that modernization services and platforms can speed up the process of implementing new workflows. The materials also describe how, by combining culture, data, and AI instructions, the organization can improve the quality of AI-generated outputs.