How SaaS solutions can pave the way for AI adoption

At the STACK Developer Conference 2024, GovTech Singapore’s Chief Technology Officer, Chang Sau Sheong shared during his keynote address that Singapore has successfully migrated 80 per cent of eligible systems to the Government on Commercial Cloud, with more than 63 agencies onboarded.

Next, the Singapore government plans to continue modernising the government’s cloud infrastructure, with more emphasis on common platforms, cloud-native development, and adoption of software-as-a-service (SaaS) tools.

As public sector speakers noted during GovInsider Live – SaaS Day, SaaS solutions enable agencies to procure common services like human resources (HR) and communication tools without having to reinvent the wheel.

Beyond common services, SaaS solutions can go so much further in helping agencies achieve productivity gains. When adopted wisely, SaaS services can also support agencies in “accessing advanced AI capabilities without rebuilding these systems from scratch,” said HCLTech’s Solutioning and Pre-Sales Lead, Public Sector, Salim Aslam.

Simplifying AI adoption

SaaS vendors are increasingly incorporating advanced AI applications within their platforms and agencies can tap on such solutions to fast-track their adoption of these applications, which they might not be able to develop in-house in a controlled and responsible manner otherwise, he said.

“SaaS platforms often include AI-powered data analytics tools that help derive valuable insights, AI-driven automation for improving operational efficiency, and AI-based content generation and summarisation tools,” Aslam noted.

These include generative AI (GenAI) tools that support content generation, semantic search, and code generation, such as GPT-3.5 Turbo and GPT-4 models, large language models (LLMs) that are available via Azure OpenAI Service.

For instance, Pair, the Singapore government’s AI-based assistant, is built on the same LLMs that power ChatGPT, customised to the compliance and security requirements of the Singapore government.

This can support agencies in improving their productivity, efficiency and service delivery.

Tapping on out-of-the-box AI capabilities via SaaS enables agencies to realise the benefits of AI without lengthy development cycles and allows public officers to focus on their core missions, Aslam added.

SaaS – already becoming the norm

In the past, government agencies were more likely to build custom applications to fit their unique requirements, ranging from citizen-facing applications to internal operations.

“However, these came with investments in infrastructure, security and maintenance that were costly and required regular refreshes over time,” said Aslam. Creating custom applications also increased technical debt over time and meant that agencies were duplicating platforms for common needs, such as identity management and payment gateways, he added.

The shift to the commercial cloud means that it has become easier for the Singapore government to adopt and build citizen-wide SaaS services that agencies can use across the board, noted Aslam. These include Singpass, a single sign-on service for many government portals.

“SaaS adoption today is more and more about subscribing to established enterprise SaaS products for HR, accounting and finance, and customer relationships,” he said.

In 2023, GovTech Singapore’s former Chief Executive, Kok Ping Soon, announced that the agency would be allocating a larger proportion of ICT spending on SaaS solutions, in part because many operational needs are common across agencies, such as case management software.

AI-powered SaaS solutions form a core component of the HCLTech Digital Foundation approach, which helps agencies modernise their IT infrastructure, improve operational efficiency, and enhance service delivery, said Aslam. Through this approach, agencies can also ensure these new platforms are seamlessly integrated with existing IT infrastructure and systems, avoiding disruptions.

Currently, HCLTech is actively developing solution accelerators for clients, helping them explore and develop GenAI solutions on top of platforms offered by large cloud providers, such as Amazon Bedrock and Amazon CodeWhisperer.

SaaS solutions also offer many of the benefits associated with cloud adoption, such as increased cost-effectiveness, scalability and flexibility, continuous updates, and enhanced collaboration across agencies.

Securing AI adoption

When adopting AI via SaaS, agencies should bear in mind security and compliance concerns.

“Securing data is paramount. There should be no compromises to ensure all data is safely used and stored,” said Aslam. Through HCLTech’s Dynamic Cybersecurity framework, the company works with agencies to protect against cyber threats and ensure compliance with regulatory requirements when adopting AI solutions via SaaS.

This includes thoroughly evaluating SaaS providers to ensure that they are compliant with government security standards and certifications, including data sovereignty and residency requirements, as well as their overall reliability and support services.

User role-based access controls, real-time monitoring, and periodic security audits are critical to ensure data remains secure as well, he shared. Adopting ethical guidelines for AI use, such as addressing potential biases in AI models and ensuring transparency in decision-making processes, can also ensure the responsible use of AI.

Finally, it is important to train staff adequately to ensure successful adoption and regularly monitor and evaluate the performance of such models.

“Innovation can a long take a long and messy path at times. SaaS shortens the path to using customer-tested, flexible AI solutions,” said Aslam.