Commercial Context
Healthcare has urgent challenges. A way to support large-scale advanced AI imaging services without escalating infrastructure costs or burdening clinical workflows. Hospitals are increasingly relying on AI to speed up diagnosis and reduce error rates, from CT scan analysis to retinal screening, but these benefits often require specialized GPU hardware, high-speed networks and constant tuning. In many hospitals, especially resource-limited settings, this level of investment is unrealistic.
However, CSPs already manage distributed computing and network infrastructure. By intelligently adjusting these resources, AI tools can be delivered as on-demand services. That was the goal of “agents and autonomous AI for business excellence.” Catalyst: To create a scalable, intentional operational environment that efficiently delivers AI workloads across communications networks.
Solution
The Catalyst team has built and demonstrated a cloud edge platform tailored to AI medical imaging. The platform allows hospitals to access advanced imaging models through telecom delivery services and remove the need for in-house computing clusters.
The hospital connects to the service via 5G. The diagnostic models are centrally hosted and deployed to a local edge server where inferences are performed in real time. This ensures rapid and degraded performance without compromising patient data privacy. The entire process from service requests to model delivery is governed by intention-based orchestration.
CSP manages this environment using several TM forum assets, including Business Process Framework (ETOM), Autonomous Network Reference Architecture, and APIs such as TMF921 (Intention Management) and TMF640 (Service Activation and Configuration). These allow for automatic deployment, monitoring and scaling. With minimal manual input.
Operationally, this approach helps CSPs shift to a higher level of autonomy. In the demonstration, the team showed that up to 60% of the operational process could be automated, increasing resource utilization by 30% and lowering operational costs by 50%.
Wide range of applications and value
The value of a healthcare provider is immediate. One AI model used in the project analyzed the 3D CT scan within 30 seconds. The same task usually takes 20-30 minutes to the radiologist. Other models showed a 37% reduction in false positives during breast ultrasound analysis. Faster results, fewer errors and reduced workload means results for both patients and clinicians.
Cost reduction is also important. Hospitals no longer have to buy and maintain costly GPU infrastructure. Industry benchmarks suggest that switching to a shared AI platform can reduce total imaging costs by 50-70%. The service is centrally managed by CSP, so hospitals can be installed within a few days. No custom buildout is required. “We are pleased to announce that Yang Jianjian, director of China Unicom and Catalyst Project Lead,” said:
The project also demonstrated broader benefits for the health sector. A single telecom host platform can serve hundreds of hospitals through the same infrastructure and model pipeline. This “model as a service” approach improves delivery times, increases consistency and expands the customer base without increasing linear costs. As AI Medical Imaging continues to grow, a scalable infrastructure from $1.52 billion in 2023 to an estimated $6.35 billion in 2030 is important.
With CSP, the benefits go beyond healthcare. The same architecture can support AI applications in logistics, gaming, or smart manufacturing. By coordinating calculations and network operations across domains, CSPs can provide value-added services and position themselves as technology leaders. This creates new revenue streams and strengthens customer relationships.
The impact of social benefits is also clear. Faster and more accurate diagnosis leads to early treatment, better outcomes and reduced costs. The WHO emphasizes that AI can close diagnostic gaps in areas that are not being served. In fact, this Catalyst solution will allow local clinics to offer sophisticated screening without local infrastructure. This means that more patients are screened, treated and supported before the conditions escalate. This catalyst provides a clear demonstration and pathway of how CSPs can drive the future of AI in healthcare and bring speed, scale and autonomy to one of the sector's most demanding and important domains.
