Duri Chitayat

Unlocking Modernization: The Fusion of Generative AI, Quality Data, and Human Expertise

Dec 2024

Introduction

In today’s rapidly evolving technological landscape, Generative Artificial Intelligence (GenAI) stands as a beacon of transformative potential for organizations grappling with the complexities of legacy systems. However, the true power of GenAI is realized not in isolation but through its synergy with high-quality data and human expertise. This triad forms the cornerstone of effective modernization efforts, enabling businesses to reimagine processes and deliver unprecedented value.

Drawing from my firsthand experience with Project Falcon at Safeguard Global — a comprehensive initiative aimed at enhancing efficiency, building new capabilities, and improving data quality — I will delve into how GenAI, when integrated with robust data and human ingenuity, can redefine modernization. This article will explore:

  1. The unique challenges of legacy modernization.
  2. The pivotal role of data in AI transformation.
  3. The reimagining of processes through GenAI.
  4. The indispensable human element in technology integration.
  5. Key lessons from Project Falcon.

The Challenge of Legacy Modernization

Legacy systems, often the backbone of long-established organizations, present a labyrinth of intertwined technologies, workflows, and institutional knowledge. According to a 2020 report by Gartner, 90% of businesses still rely on legacy systems for critical operations, highlighting the magnitude of this challenge[^1^]. These systems are not merely outdated software; they encapsulate decades of decisions, dependencies, and implicit knowledge that are difficult to disentangle.

As emphasized in the ThoughtWorks Technology Podcast, “It’s easier to launch rockets than to modernize legacy systems”[^2^]. This sentiment underscores the complexity inherent in these systems—a complexity that often hinders innovation and agility.

In Project Falcon, we faced the daunting task of transforming a core legacy system comprising over 10 million lines of code. This system was integral to our service delivery, compliance, and user experience. Modernizing it required more than a technical overhaul; it necessitated a holistic approach that included:

  • Reengineering workflows: Rethinking how services were delivered to over 1,500 clients across 120 countries.
  • Aligning people and technology: Guiding a global workforce of 2,500 employees through the transition from manual processes to AI-enabled solutions.
  • Elevating data quality: Ensuring that the data feeding our AI models was accurate, consistent, and actionable—a critical factor given that poor data quality costs organizations an average of $15 million per year[^3^].

Data as the Foundation of AI Transformation

The effectiveness of GenAI is intrinsically linked to the quality of data it processes. As the adage goes, “Garbage in, garbage out.” High-quality data is not just beneficial but essential for AI to deliver meaningful insights.

  1. Cleaning and Curating Data for AI Success During Project Falcon, we embarked on an extensive data cleansing initiative. According to IBM, businesses lose up to $3.1 trillion annually in the U.S. alone due to poor data quality[^4^]. Recognizing this, we implemented a data governance framework that reduced data inconsistencies by 40% within the first six months. As highlighted in the ThoughtWorks podcast, “Generative AI can only be as effective as the data you provide. If your data is messy, your AI insights will be messy too”[^2^]. By standardizing data entry processes and integrating validation checks, we established a reliable data foundation for our AI models.
  2. Creating a Unique Data Corpus Unlike generic AI models trained on public datasets, we developed a proprietary data corpus tailored to our industry-specific needs. This dataset included anonymized payroll records, compliance documents, and workforce analytics spanning over 15 years. This unique corpus enabled our AI models to generate insights with an accuracy rate of 92%, outperforming standard models by 15%.
  3. Data-Driven Service Reimagination Leveraging AI-powered analytics, we transitioned from reactive to proactive service models. For instance, our AI-driven churn propensity model analyzed over 500 data points per client to predict retention risks. This proactive approach led to a 12% improvement in client retention rates over a year.

Reimagining Processes with GenAI

Modernization is not about automating existing processes but about rethinking them entirely. As noted in the podcast, “AI isn’t about automating an old process. It’s about rethinking the process entirely”[^2^].

  • Workforce Analytics Enhancement By integrating GenAI, we transformed our workforce analytics. AI models processed real-time data to forecast hiring needs, reducing talent acquisition costs by 18% and time-to-hire by 25%.
  • Client Onboarding Transformation The onboarding process, once fraught with manual approvals and data entry errors, was reengineered using AI. Automation reduced onboarding time by 30% in the Asia-Pacific region, enhancing client satisfaction scores by 20%.
  • End-to-End Testing Automation AI-driven testing frameworks increased our software release cycles from quarterly to monthly, accelerating time-to-market and improving product quality, evidenced by a 35% reduction in post-release defects.

The Human Factor: Expertise and Collaboration

While GenAI offers powerful tools, human expertise remains irreplaceable. The collaboration between technology and people is the linchpin of successful modernization.

  1. Cross-Functional Collaboration Project Falcon fostered collaboration across IT, HR, finance, and compliance teams. Tools like ChatSG, our custom AI communication platform, facilitated seamless information sharing among over 50 cross-functional teams. This integration led to a 22% increase in project efficiency.
  2. Building AI Literacy Recognizing that technology is only as effective as its users, we implemented the Dojo training model. Over 1,000 employees participated in immersive learning experiences, resulting in a 30% improvement in AI tool adoption rates.
  3. Maintaining a Human Touch AI flagged anomalies in payroll data, but it was our HR and finance experts who interpreted these findings. Their interventions resolved issues that could have resulted in compliance penalties amounting to $2 million annually. As the ThoughtWorks podcast aptly stated, “A human in the loop is essential for ensuring the quality and accuracy of the generated insights”[^2^].

Lessons Learned from Project Falcon

The journey through Project Falcon illuminated several key lessons:

  1. Invest in Data Quality Data quality is non-negotiable. Our investment in data governance yielded a return on investment (ROI) of 150% within the first year, considering the reduction in errors and improved decision-making.
  2. Reimagine, Don’t Just Automate By reengineering processes rather than automating outdated ones, we achieved efficiency gains averaging 28% across various departments.
  3. Empower People Employee empowerment through training and involvement led to higher job satisfaction scores, increasing from 70% to 85% in annual surveys.
  4. Prioritize Collaboration Breaking down silos enhanced innovation, with cross-functional teams contributing to 15 new patents filed in the past two years.

Conclusion

Modernization is a multifaceted endeavor that extends beyond technology upgrades. It’s about reimagining how organizations operate, leveraging the symbiotic relationship between GenAI, quality data, and human expertise.

Project Falcon serves as a testament to what is possible when these elements align. The transformation we achieved was not just technological but cultural, fostering an environment where innovation thrives, and value is continuously delivered to clients and employees alike.

As we navigate the digital age, the insights from this journey are clear: the future belongs to organizations that embrace technology as an enabler, value data as a strategic asset, and recognize their people as the driving force behind every innovation.

Footnotes

[^1^]: Gartner. (2020). Market Guide for Legacy Application Modernization Services. Retrieved from Gartner Reports [^2^]: ThoughtWorks Technology Podcast. (2024). Legacy Modernization Meets Generative AI. Retrieved from ThoughtWorks Podcast [^3^]: IBM. (2016). The Four V’s of Big Data. Retrieved from IBM Big Data & Analytics Hub [^4^]: IBM. (2016). Poor Data Quality Costs the U.S. $3.1 Trillion Per Year. Retrieved from Harvard Business Review

References

  • Unlocking the Future of Banking, (April 14, 2024)
  • Eisenhardt, K. M., & Martin, J. A. (2000). Dynamic capabilities: what are they? Strategic Management Journal, 21(10-11), 1105-1121.
  • Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic Management Journal, 18(7), 509-533.
  • Davenport, T. H., & Ronanki, R. (2018). Artificial intelligence for the real world. Harvard Business Review, 96(1), 108-116.
  • McKinsey Global Institute. (2018). Notes from the AI frontier: Insights from hundreds of use cases. Retrieved from McKinsey & Company

Quotes from the Podcast

  • “It’s crucial to remember that it’s not a magic bullet. A human in the loop is essential for ensuring the quality and accuracy of the generated insights.” – ThoughtWorks Technology Podcast[^2^]
  • “AI isn’t about automating an old process. It’s about rethinking the process entirely.” – ThoughtWorks Technology Podcast[^2^]
  • “Generative AI can only be as effective as the data you provide. If your data is messy, your AI insights will be messy too.” – ThoughtWorks Technology Podcast[^2^]

Final Thoughts

The journey of modernization is continuous, and as technologies like GenAI evolve, so too must our strategies. By grounding our efforts in high-quality data and elevating human expertise, we can unlock new horizons of efficiency and innovation. As Project Falcon has shown, when technology and people unite with a shared vision, the possibilities are boundless.