China's CAICT and AIIA publish an FDE delivery maturity spec that grades organizations from L1 to L5
On 6 November the AI institute of the China Academy of Information and Communications Technology (CAICT), through the AIIA's FDE delivery research group and close to 60 member organizations, released technical specification AIIA/T 0356-2026. It rates how well a company delivers forward deployed engineering across five capability areas and 20 items, on five maturity levels.
China's forward deployed engineering push now has an organization-level yardstick. According to a CAICT announcement summarized by the FDE digest Zhiqi Yicheng, the CAICT AI institute and the China Artificial Intelligence Industry Alliance (AIIA) released Requirements for Forward Deployed Engineering (FDE) Delivery Capability Maturity (AIIA/T 0356-2026) on 6 November 2026. Contributors named include Lenovo, Zhipu, Digital China Information Service, China Unicom and China Mobile units, China Telecom's digital arm, China Southern Grid's AI company, and VNET.
The spec groups FDE delivery into five capability areas: business advisory (scenario insight, problem definition, solution design, value validation), engineering delivery (architecture, AI engineering, integration, testing and deployment, operations), project management (schedule, requirements and change, quality and acceptance, security and compliance), outcome evolution (lessons learned, knowledge assets, reuse) and organizational support. Across 20 capability items it defines five levels, L1 to L5.
Unlike the individual FDE certificates recently launched by Tencent Cloud, Alibaba Cloud and Lenovo, this one assesses whether a company can deliver FDE work consistently, not whether a person passed an exam. CAICT had already opened a first round of FDE delivery maturity assessments in October.
Why it matters for FDEs: in China, buyers may soon ask vendors for a maturity level, and the five areas make a useful checklist of what employers expect FDE teams to own beyond writing code, including problem framing, acceptance criteria, compliance, and turning field work into reusable product.
Source: www.jiangyou-ai.com