GLP embraces AI-driven data analytics on Microsoft Azure

How we helped

Presented with 20 potential use cases by GLP for its envisioned AI-driven data analytics solution, the Foundry for AI by Rackspace (FAIR™) team embarked on a series of discovery sessions to meticulously analyze GLP’s needs and identify the most pressing use case. Amid a fast-approaching visit by a global team to Singapore, a demo was created to showcase the potential of addressing GLP's core issue: the transformation of its daily reports through AI-enhanced data-querying capabilities.

“We did a few rounds of our sprints. And from sprint to sprint,  it was getting better and better. The use case become much more usable and real for our organization”, explains Song. “That was a key differentiator for Rackspace. I think the company has a different approach compared to traditional consulting firms. They are much simpler, faster and practical.”

Nathan Cheng, Data Analytics and Machine Learning Systems Engineer at Rackspace Technology, explained the wish list's two main themes: a user-friendly product interface and an intelligent system capable of deep-diving into data dimensions to answer complex business inquiries. This wish list evolved into concrete test cases, marking the roadmap for the project's technical progression.

Building a centralized hub for open-ended data insights was an ambitious endeavor that began with a focus on fund management and could extended to ESG metrics and beyond. The goal was to create a dynamic interface where users could easily access their conversation histories, particularly when reviewing the fund management KPIs via the app.

The interface design focused on a cohesive user experience and easy conversation-based functionality. Interactions needed to be logged, as documentation was crucial for identifying and analyzing any existing logic gaps, which would, in turn, inform iterative improvements to the system. Then conversations turned toward complex queries, like comparing the occupancy rate metrics of properties across different quarters. Such requests required the system to sift through multiple data dimensions, pinpoint the relevant data, and identify the factors that drove changes in the metrics.

Transitioning to the technical execution, Srinivas Chakravarthy, a Senior Director of Business Development at Rackspace, highlighted the innovative approach of crafting on-the-fly SQL queries based on users' inquiries. Initially working with CSP files, the goal was to transition into a production-ready phase where data lakes could be utilized for real-time data access and analysis. The process entailed running an SQL query, interpreting the data and furnishing the user with insightful responses. Achieving this required a precise SQL phrasing, aligning the technology to comprehend and respond to the contextual nuances of each inquiry.