Sparks & Honey
A division of Omnicom, the large public marketing services company, S&H focuses on creating AI driven insights to its large universe of advertising partners
Sparks and Honey, a renowned cultural intelligence consultancy, empowers organizations to navigate the ever-evolving landscape of cultural shifts and trends. As part of the Omnicom Precision Marketing Group, Sparks and Honey leverages their AI-powered intelligence platform, Q™, to quantify and predict cultural insights, igniting transformation for brands and consumers alike.
The Challenge
Sparks and Honey embarked on a mission to enhance their Q™ application to better serve their clients' evolving needs. The challenges included:
- Legacy Application Support: The existing Q™ application required maintenance and improvements to ensure it continued to function efficiently. Azumo was tasked with providing break/fix support and adding new features to the legacy Python application.
- Data Enhancement: To improve data ingestion and enhance the accuracy of cultural insights, Azumo developed custom scrapers for data collection.
- Architectural Redesign: The monolithic design of the application hindered agility and rapid feature development. Sparks and Honey needed an architectural overhaul to support client-specific requirements and enable faster updates.
Behind every huge business win is a team win, a coalition win, and a technology win. So it is worth pointing out the teams, stacks, and packages we've been using to achieve low-latency and real-time GenAI on our 24/7 platform and live in our studios … and it all came together with a fantastic set of developers from Azumo.
The Solution
Azumo collaborated closely with Sparks and Honey to deliver a comprehensive solution that addressed these challenges:
Phase 1: Discovery and Legacy Support
- Full Assessment: Azumo conducted a thorough assessment of Sparks and Honey's existing code, infrastructure, and specific needs.
- Legacy Application Support: Azumo provided break/fix support for the legacy Q™ application, ensuring its continued functionality and stability.
- Frontend Enhancement: Azumo added new features to the legacy application's frontend, improving its usability and client experience.
- Data Scrapers: Custom data scrapers were developed to enhance data ingestion, allowing Sparks & Honey to gather valuable cultural insights more effectively.
Phase 2: Architectural Redesign and Application Rebuild
- Architectural Redesign: Azumo led a redesign effort, transitioning from the monolithic architecture to a modular one that facilitated rapid iteration and feature additions.
- Q™ Application Rebuild: The Q™ application was rebuilt from the ground up, ensuring it was aligned with Sparks and Honey's evolving needs and future scalability.
- Frontend-Backend Separation: By separating the frontend and backend components, Azumo enabled faster updates and improvements without disrupting the entire system.
- Prediction Charts: Azumo added a new section to the Q™ application, featuring prediction charts that provided valuable insights into cultural trends and shifts.
Phase 3: Leveraging Generative AI for Public Briefings
- Generative AI Integration: Azumo harnessed ChatGPT and other tools to support Sparks and Honey's public briefings.
- Pre-Briefing Support: Azumo assisted in creating question cheat sheets and scripts for higher management and opinion leaders, enhancing the effectiveness of the briefings.
- During Briefing: Azumo generated live transcripts, word clouds, takeaways, and proposed follow-up questions, facilitating real-time engagement during briefings.
- Post-Briefing Insights: Azumo's generative AI generated transcripts, summaries, and key takeaways for publishing on social media, expanding the reach of Sparks and Honey's insights.
Phase 4: Tailored Reports and Frontend Enhancement
- Co-Pilot Briefing Support: Azumo provided assistance in creating pitch decks for specific Omnicom clients, ensuring industry-specific topics were effectively addressed.
- Flash Reports: Azumo developed questions for follow-up discussions, leveraging the input of news articles and baseline questions.
- Audience-Focused Reports: Using Omnicom data, Azumo crafted industry-focused reports tailored to specific target audiences, enhancing the relevance and impact of Sparks and Honey's insights.
- Open Source Model Comparison: Azumo conducted a thorough comparison of open-source models to improve performance and efficiency.
- Frontend Styling: The frontend of the Q™ application was aligned with Omnicom's design and style, ensuring a cohesive and branded user experience.
Results
The collaboration between Sparks and Honey and Azumo yielded significant outcomes:
- Enhanced Application: The Q™ application was transformed, becoming more agile, feature-rich, and aligned with client needs, all while maintaining its design integrity.
- Data Accuracy: Custom data scrapers and enhancements led to improved data ingestion, resulting in more accurate cultural insights.
- Efficiency: The separation of frontend and backend components allowed for faster updates and feature additions, enhancing overall efficiency.
- Insightful Briefings: Generative AI-powered support during briefings improved the quality and depth of insights shared with clients.
- Tailored Reports: Sparks and Honey could provide highly customized reports to Omnicom clients, increasing the relevance and impact of their cultural intelligence.
The collaboration between Sparks and Honey and Azumo exemplifies a successful partnership in leveraging AI and software development to empower cultural intelligence. Azumo's support enabled Sparks and Honey to enhance their Q™ platform, providing clients with more accurate insights and facilitating more engaging briefings. This case study demonstrates the power of technology to transform cultural understanding and drive business growth.
