Welcome to the AI Agent & Copilot Podcast, analyzing the latest AI Copilot and agent developments from Microsoft and its partners, delving into customer use cases, and exploring how AI plus the Cloud helps customers reimagine their business.
In this episode, Asad Mahmood, vice president of data and AI at HSO, speaks with Tom Smith about his upcoming sessions at AI Agent & Copilot Summit. Mahmood is speaking in Masterclasses on:
- Maximizing Retention and ROI with Fabric AI for Churn Prediction
- Accelerating Data Migration with Gen AI for ERP
Highlights
Background (02:07)
Mahmood’s focus at HSO is on Microsoft technology and AI services; he has worked in the tech industry for 25 years in the industry and he notes that HSO made in early investment in AI — all the way back in 2015.
Preview: Maximizing Retention and ROI with Fabric AI for Churn Prediction (05:02)
Mahmood explains the importance of trend prediction and customer retention, particularly in the financial services industry. He describes a customized churn model combining conventional AI and LLMs, achieving high accuracy (over 80%) for predicting customer churn. The session will focus on creating focused campaigns to retain customers, highlighting the cost-effectiveness of retention over acquisition. Microsoft Fabric is recommended for most customers, with AI Foundry and Fabric AI creating models on top of customer data. The session will provide strategies for customer engagement based on predictive analytics. He emphasizes the comprehensive insights provided by the models, beyond just churn probability.
AI Agent & Copilot Summit is an AI-first event to define opportunities, impact, and outcomes with Microsoft Copilot and agents. Building on its 2025 success, the 2026 event takes place March 17-19 in San Diego. Get more details.
Preview: Accelerating Data Migration with Gen AI for ERP (10:14)
Data migration in ERP settings poses many challenges, often leading to project delays and cost overruns. He highlights the importance of clean and quality data, noting that data migration issues often outweigh ERP configuration challenges. HSO researches and develops automated AI-driven data mappings, significantly reducing the complexity and cost of data migration, using LLMs to standardize values across multiple source systems. LLMs learn from user feedback, improving the accuracy and efficiency of data mappings.
The automated Fabric pipeline creates templates for future data migrations, allowing business users to interact with and adjust the mappings. Significant cost and time savings are achieved through AI-driven data mappings, particularly for customers with multiple legal entities or undergoing mergers and acquisitions.
Ask Cloud Wars AI Agent about this analysis
