Managing vast enterprise data across multiple, fragmented systems while complying with strict regulations is a major challenge for many organisations. Artificial intelligence (AI) and machine learning (ML) offer powerful tools for addressing these issues, with the global AI market potentially exceeding $1.8 trillion by 2030 and enterprise AI projected to surpass $155 billion.
Over 80% of organisations are already exploring or implementing AI, making experienced leadership essential. Ashitosh Chitnis, with more than 16 years in data engineering and analytics at companies like Google, Apple, IBM, and Deloitte, shares practical insights on applying AI/ML to solve core business problems, especially in finance and procurement.
Taming Data Complexity for AI Success
Data fragmentation creates inconsistencies and limits the insights AI can deliver. This isn’t just inconvenient—data silos cost businesses trillions annually. At Google, finance and procurement teams struggled with outdated, rules-based systems. By unifying data and applying AI tools like Vertex AI and BigQuery ML, they achieved over $50 million in annual cost savings and cut manual audits by 90%.
Success depends on strong governance. Implementing Master Data Management (MDM) helped create a single source of truth, ensuring consistency and standardisation across teams. This foundation is critical to scaling AI effectively.
AI-driven Transformation in Finance and Procurement
Finance and procurement are prime candidates for AI optimisation due to their high transaction volumes and accuracy demands. AI can automate many finance functions, reducing costs by more than 30% in specific areas, with adoption expected to reach 58% by 2024.
Building solid data foundations is key. Integrating SAP and non-SAP data into BigQuery-based data lakes, then modelling it into canonical reporting data marts, ensures data consistency and usability. Continuous collaboration with business stakeholders helps refine AI models to align with real workflows.
Procurement benefits significantly as well, with AI-powered techniques delivering up to three times the savings of traditional approaches, targeting 10-15% reductions in sourcing costs. Clear cost-benefit analysis, explainable AI (XAI), regular drift detection, and bias audits are essential to maintain trust and compliance.
Compliance with regulations like SOX and GDPR requires enforcing data lineage tracking and data immutability. Real-time monitoring and anomaly detection provide necessary guardrails to protect data integrity.
Scaling AI Effectively: Technology, Strategy, and Outlook
Deploying and scaling ML models faces many hurdles. Gartner reports that only about half of AI projects reach production, with failure rates up to 85%, often due to data or scaling issues. Preventing this requires strong engineering and operational practices, including automated data quality pipelines with validation, cleansing, and anomaly detection.
Tools such as Datadog and Grafana can support monitoring for data drift and anomalies, helping maintain AI model reliability over time.
Focusing on business needs instead of technology hype is vital. AI projects succeed when they start with clear problems—like revenue optimisation, risk reduction, or process efficiency. Data readiness and governance must be priorities from the outset, as AI’s effectiveness depends on the quality of its data.
Enterprises should invest early in data quality, governance frameworks, and scalable architectures. Implementing AI-driven data pipelines, self-healing systems, and federated governance models lays the groundwork for sustainable, trustworthy AI growth. Organisations taking this approach will be better prepared for future innovation.
The integration of AI and ML into enterprise data management has strong potential to break down data silos, optimise finance and procurement processes, and boost compliance. Achieving this requires more than technology—it demands a solid data foundation, clear governance, cross-team collaboration, and scalable, explainable AI models aligned with business goals.
```Your membership also unlocks:
