Preparing the foundations for Generative AI

Governments and public services agencies are keen to push forwards with generative AI. Recent research by McGuire Research Services for Avanade found 91% of organisations in the sector believe they need to shift to an AI-first operating model within the next 12 months, while 87% of employees feel generative AI tools will make them more efficient, and more innovative.

Yet making this shift isn’t simply a matter of adopting generative AI tools and hoping this alone will drive success. Without a modern data platform and an effective supporting cloud infrastructure, organisations could be scuppering their progress from the start.

Data modernisation is crucial. Generative AI tools are only as accurate and effective as the data they can access, and while government and public service agencies will have invested in their data infrastructure, it still may not be optimised for the demands of generative AI. There may still be work to do in developing a cloud-first infrastructure, secure by design, where data that is unified, protected, and integrated with both existing business applications and AI systems, can be accessed through composable APIs. This requires skillsets that firms may not have in-house. Data also needs to be sorted, annotated and labelled in order to meet the requirements of generative AI. No wonder CIO’s 2023 AI Priorities study found that data integration was the number one concern for IT leaders around generative AI integration, above security and privacy and the user experience.

This seems a formidable task. Applications and infrastructure may need further modernisation and legacy workloads moved from on-premise hardware to private or hybrid cloud. Microsoft Fabric could be one way to deal with it. It provides organisations with the platform and the tools required to support their drive into generative AI. It unifies all data on a single platform, including data integration, engineering, and warehousing, where it can be used for data science, real-time analytics, and business intelligence – and accessed with natural language queries and the power of generative AI. All data is held in a lake-centric hub, and protected by a strong, universal security model, with data loss prevention and protection for sensitive data, and features for auditing and forensic investigation already built-in.

Optimising for generative AI

Microsoft Fabric is usually hosted on an SaaS solution, based on the same solutions and clouds most governments and public service agencies will be working with already. Yet building an AI-ready infrastructure may require organisations to go deeper, strengthening and optimising their existing clouds. Poor performance, cloud sprawl, a lack of integration and unpredictable cloud costs can also affect the success of generative AI initiatives, so organisations need to inspect and optimise their cloud provisions before jumping headfirst into deploying AI tools.

“We see tremendous potential for public sector services from Microsoft Fabric” says Darren Brown, Government & Public Services Lead for Avanade. “We know that despite agencies investing well in data warehouses and data lakes, there is still much more work to do. Fabric could be the answer for agencies awash in data, but without the time or people to truly make sense of it.”

Beyond the data platform and the cloud infrastructure, agencies and organisations will need to support their workforce to adapt to new tools and processes. Since without the right training, management and integration, they’ll be unable to realise the benefits of AI investments.

If this all seems challenging, Avanade can help. Avanade has significant expertise and experience with Microsoft Fabric and Microsoft Copilot, and the tools and capabilities to help organisations modernise their data platforms and integrate AI into processes and workflows. Avanade can also help agencies prioritise and modernise their data platforms and their applications, while Avanade Cloud Impact can use AI and machine learning to understand cloud usage, identify issues and find opportunities to optimise and drive costs down.

To find out more about preparing for generative AI and the next steps on your journey, visit: Government & Public Services | Avanade.