Data+AI World Tour for Government 2025 -Measuring Mission Impact

Learn how to evaluate the quality and deployability of your AI investments for measuring mission success, justifying the mission, and securing future resources. This session moves focus beyond basic reporting to show how operational AI outcomes can be quantitatively measured and linked back to high-level strategic goals. Databricks will describe our latest science for quantifying AI quality. See how a unified data environment enables a continuous feedback loop, constantly monitoring both program performance and mission objectives to ensure constant improvement and measurable mission success. Takeaways: practical approaches for establishing reliable metrics for subjective evaluations, techniques for calibrating LLM judges to enable cost-effective, scalable assessment, and actionable frameworks for evaluation systems that evolve with your AI capabilities. 𝐂𝐡𝐚𝐩𝐭𝐞𝐫𝐬 𝟎𝟎:𝟎𝟎 – 𝐖𝐡𝐲 𝐌𝐞𝐚𝐬𝐮𝐫𝐞𝐦𝐞𝐧𝐭 𝐈𝐬 𝐭𝐡𝐞 𝐑𝐞𝐚𝐥 𝐁𝐨𝐭𝐭𝐥𝐞𝐧𝐞𝐜𝐤 𝐢𝐧 𝐀𝐈 Jonathan Frankle, Chief AI Scientist — Databricks 𝟏𝟖:𝟒𝟓 – 𝐌𝐮𝐬𝐭-𝐇𝐚𝐯𝐞𝐬 𝐯𝐬. 𝐌𝐮𝐬𝐭-𝐍𝐨𝐭-𝐇𝐚𝐯𝐞𝐬 𝐢𝐧 𝐀𝐈 𝐒𝐲𝐬𝐭𝐞𝐦𝐬 Jonathan Frankle, Chief AI Scientist — Databricks 𝟐𝟕:𝟒𝟎 – 𝐌𝐞𝐚𝐬𝐮𝐫𝐢𝐧𝐠 𝐀𝐈 𝐰𝐢𝐭𝐡 𝐂𝐚𝐥𝐢𝐛𝐫𝐚𝐭𝐞𝐝 𝐋𝐋𝐌 𝐉𝐮𝐝𝐠𝐞𝐬 Jonathan Frankle, Chief AI Scientist — Databricks 𝟒𝟕:𝟑𝟎 – 𝐌𝐞𝐚𝐬𝐮𝐫𝐢𝐧𝐠 𝐌𝐢𝐬𝐬𝐢𝐨𝐧 𝐈𝐦𝐩𝐚𝐜𝐭 𝐚𝐭 𝐭𝐡𝐞 𝐖𝐨𝐫𝐥𝐝 𝐁𝐚𝐧𝐤 𝐆𝐫𝐨𝐮𝐩 Arvin Shinasan, Head of Data & Analytics — World Bank Group