Iress, a provider of financial services software, utilized generative AI to streamline its response process to lengthy Requests for Information (RFIs) and Requests for Proposals (RFPs) commonly encountered in procurement procedures.
Utilizing GenAI to Enhance Procurement Responses Jeff Gibson, Head of Business Intelligence at Iress, shared during the Databricks Data Intelligence Day in Melbourne the company’s approach to addressing a significant business challenge with GenAI.
“We were focused on tackling a genuine business issue from the start, rather than engaging in another theoretical experiment. Our aim was to genuinely resolve a real business use case and gain insights into implementing and managing AI,” Gibson explained.
The company identified a specific challenge faced by its commercial and product teams and brought together representatives from various departments, including Business Intelligence, Data Innovation, and Databricks, to collaborate on a two-day project.
“We identified a significant challenge related to managing Requests for Proposals/Requests for Information within our business,” Gibson elaborated. “These requests often comprise numerous questions, ranging from 50 to 950, sourced from various specialists and documentation across the organization. Responding to these requests efficiently is crucial, as it can take up to three-to-six weeks, involving multiple teams and incurring substantial costs.”
With a focused timeline, the team developed an application named ‘Bella’ within the two-day sprint, offering an interface for querying questions and retrieving data-driven responses.
Gibson emphasized that while other methods, such as implementing a search engine over a spreadsheet or intranet site, were feasible, GenAI introduced an additional layer of context to the process.
“GenAI provides essential context surrounding the question, the client, and how the information should be presented,” Gibson stated.
‘Bella’ was developed using Databricks, AWS for storage infrastructure, and leveraged Hugging Face for accessing AI models.
Gibson noted the rapid evolution of Databricks’ platform and highlighted the potential benefits of incorporating additional capabilities into future projects.
Aside from resolving the immediate business challenge, the two-day project provided valuable insights for the team members involved and informed Iress’s roadmap for integrating GenAI into its products.
“We had a diverse team that gained extensive knowledge in a short period, shaping our internal GenAI roadmap and its application in our products for our clients,” Gibson remarked.
Furthermore, the project underscored the importance of effective data management in AI initiatives, prompting Iress to refine its data management and governance practices accordingly.
