In the last few years, many organisations have proven that AI can generate impressive demonstrations. Far fewer have proven that it can deliver measurable business value at scale.
Despite record investment and boardroom urgency, 95% of enterprise Gen AI pilots fail to deliver measurable ROI (MIT, 2025).
At NCS, we are seeing a consistent pattern with clients across public safety, telecommunications, transport, and other critical sectors. The gap isn’t the technology, but the foundations. Organisations chase use cases without addressing the main requirements that impact AI programmes before deployment.
The next phase of enterprise AI will centre on how effectively organisations can industrialise AI. It signals a shift from building individual AI applications to building an AI factory – a reusable enterprise foundation that brings the right infrastructure, governance, and tools to build and operate AI at scale. Instead of building every AI application from scratch, organisations can establish a trusted foundation once and use it repeatedly across different business needs.
NCS’s Sunshine.AI suite is our approach to building that AI factory. It includes multiple solutions, several of which that are covered in this article and leverage NVIDIA's Nemotron models and NeMo framework for their production-ready capabilities, integrated safety mechanisms, and deployment flexibility.
The results are starting to show: AI systems in daily operational use, measurable business impact, and a clearer path from promise to ROI.
Our experience has highlighted three requirements for making this model work.
Three Non-negotiable Requirements for Building Production-Ready AI Factories
Through our work deploying AI in production environments, three requirements have emerged as non-negotiable:
- Safety and trust must be built in: Production AI systems need robust guardrails from day one. In our public safety deployments, AI-generated insights inform decisions that affect people's lives. In telecommunications, AI infrastructure must meet enterprise-grade security and governance requirements.
- Measurable business value: The question that determines whether AI moves from pilot to production is not "how sophisticated is the model?" but "what measurable business value does this deliver?" We focus relentlessly on ROI - faster response times, operational efficiency, cost savings, improved safety outcomes.
- Production readiness: The biggest gap we see is between proof-of-concept and “production ready”, the point where AI becomes embedded in daily operations. Production AI needs to handle actual data volumes, work across varied infrastructure, and operate reliably under real-world conditions.
How NCS approaches enterprise AI through Sunshine.AI
By bringing together our expertise, capabilities, and technologies from leading AI ecosystem partners, our Sunshine.AI suite allows organisations to deploy enterprise-grade AI within their own environments while meeting rigorous governance, compliance, and safety requirements.
The AI factory is ultimately about creating a reusable foundation. Once that is in place, organisations can develop and deploy multiple AI applications without having to reinvent the underlying technology and governance model each time.
Sunshine.core: The Foundation for Trusted Enterprise AI
Before organisations can deploy AI at scale, they need a trusted foundation for building and operating production-grade AI agents. Sunshine.core is NCS' AI foundational platform that enables organisations to build, deploy, govern and operate AI agents at scale, while addressing critical requirements for performance, governance and safety.
Built on open standards, Sunshine.core provides a library of reusable, mix-and-match components that reduces the need to rebuild common AI infrastructure for every deployment. These components form a central control plane for governing and monitoring AI agents in production, while helping organisations maintain control over their data and models without vendor lock-in.
The platform runs NVIDIA Nemotron reasoning models to power AI use cases such as network troubleshooting and root cause analysis — surfacing probable causes across complex, multi-layer network estates and shortening time to resolution for operations teams.
Governance is built into the core architecture rather than added later. Sunshine.core applies a layered set of safety controls across the request lifecycle, with NVIDIA Nemotron content safety models forming one component of that stack.
Sunshine.core has been deployed with a major telecommunications service provider, powering enterprise AI services in a regulated environment.
Sunshine.kaisense: From Video Data to Actionable Insights
Organisations generate massive amounts of video data – from cameras to operational monitoring systems. The challenge has been to extract actionable insights from these data at scale and speed.
Sunshine.kaisense is built with NVIDIA Video Search and Summarisation Blueprint, Cosmos, Nemotron and NeMo frameworks to enhance capabilities in video search, summarisation, and alert verification. Already in operation for public safety agencies, Sunshine.kaisense aims to further help authorities respond faster to incidents through intelligent video analysis – delivering 80% faster real-time searching compared to traditional methods and 30% faster deployment times. When seconds matter in emergency situations, AI-powered search delivers measurable impact on response times in critical situations.
Beyond public safety, the solution's capabilities extend to applications in workplace safety and transport (real-time video analysis for traffic management and incident response).
Sunshine.productivity: Bringing AI into everyday work
Sunshine.productivity provides intelligent assistance for everyday tasks - from document analysis to content generation - while addressing the deployment and security requirements that enterprises demand.
NCS is deploying Sunshine.productivity for a major telecommunications service provider, using NVIDIA Nemotron open models and Alibaba's Qwen models for its on-premises MVP The project combines production-ready AI optimisation with a practical deployment approach, enabling enterprise AI services to be delivered efficiently in a secure on-premises environment while supporting workforce productivity.
This approach addresses the needs of organisations with stringent data residency and security requirements, enabling them to adopt AI capabilities in environments where cloud-only solutions may not be viable.
Sunshine.coder: Accelerating Software Development
Sunshine.coder is an agentic coding application that accelerates the software development lifecycle by helping developers write, test and convert code faster. Unlike generic coding assistants, it leverages NCS' proprietary knowledge base to provide context-aware coding guidance that aligns with organisation-specific architectures, development standards and secure coding practices.
Leveraging NVIDIA Nemotron models, Sunshine.coder combines visual understanding, analysis and code generation capabilities to support software engineering workflows. NVIDIA Nemotron 3 Nano Omni enables the platform to understand architecture diagrams and UI screenshots, while Nemotron 3 Nano 30B provides analysis and code generation capabilities that help accelerate development workflows.
Designed with a sovereign-ready architecture, Sunshine.coder enables organisations to retain full ownership and control of their models and data while meeting stringent security, compliance and data sovereignty requirements. Implemented internally at NCS and scaling towards more than 1,000 engineers, Sunshine.coder has delivered approximately 30% improvements in productivity and quality across a variety of engineering use cases. The platform also enables developers to rapidly create specialised AI assistants tailored to their coding needs, from proposing fixes to security vulnerabilities to converting legacy code.
Sunshine.builder: Democratising AI Development
Sunshine.builder is an AI-powered application that enables business analysts and non-technical users to build prototypes and software applications without writing code. By translating business requirements into working systems, the platform allows users to define what they need, analyse how solutions will work, validate that requirements are met and deploy applications without relying on developer resources.
Powered by NVIDIA Nemotron models, Sunshine.builder incorporates multimodal understanding, analysis and generation capabilities to support the application development process. NVIDIA Nemotron 3 Nano Omni enables the platform to understand architecture diagrams or UI screenshots, while Nemotron 3 Nano 30B provides analysis and generation capabilities that help transform business requirements into deployable solutions.
Implemented internally at NCS, Sunshine.builder enabled the creation of more than 800 prototypes and AI-powered agents in less than 90 days. By empowering business users to participate directly in application development and validation, organisations can reduce bottlenecks, respond more rapidly to changing requirements and improve business agility. The platform is being adopted across enterprise and public-sector environments, including regulated and air-gapped deployments.
Business Imperatives for an AI-ready Organisation
Safety Enables ROI: Integrated safety mechanisms are not obstacles to deployment – they are enablers. Without robust safety, AI remains stuck in pilot phase.
Business Outcomes Matter More Than Technical Sophistication: The value of an AI model is ultimately determined by what it enables an organisation to do better, faster, or more efficiently.
Production Readiness Determines Success: Building for operational deployment from the start makes the difference between pilots that stall and solutions that reach production and deliver ROI.
Looking Ahead
Enterprise AI is maturing. The question is no longer "can AI do this?" but "can we deploy AI that delivers measurable, sustained ROI"?
Our experience building our Sunshine.AI suite demonstrate that the answer is yes - but it requires systematic approaches, proven technology foundations, and relentless focus on business value. Our work with clients and technology ecosystem partners are helping us explore what the future can look like.
And the goal is simple: move from AI promise to ROI.
