# Databricks tells ad algorithms what value means _Databricks launched CustomerLake Conversions to send value-ranked signals to ad platforms via CAPI._ **Published:** 2026-11-19 **Source:** https://www.startuphub.ai/ai-news/artificial-intelligence/2026/databricks-tells-ad-algorithms-what-value-means --- [Databricks](https://www.databricks.com/blog/customerlake-conversions) launched CustomerLake Conversions to push value-based conversion signals from first-party data directly to ad platforms. The capability sits inside Campaign Agents, its Agentic CDP built natively in Databricks, and it will be available to Campaign Agents customers through their account team, the company said. Two purchases stop counting the same. Databricks frames the change around a retailer example. A full-price first purchase from a shopper with high predicted lifetime value and a discount-driven one-off both register as conversions today. CustomerLake Conversions lets the paid media team pick only the first type in a dedicated Conversions tab, using a lifetime value model that already lives in Databricks, then ship it through native Reverse ETL and Conversions API connectors. Campaign insights return the other way through OpenSharing integrations and Lakeflow Connect, so spend sits next to customer value in the same foundation. That replaces manual extracts, platform-specific uploads and spreadsheet reconciliation. Launch destinations include Amazon Ads, [Google](https://www.startuphub.ai/ai-news/ai-stocks-daily/2026/ai-stocks-2026-06-04), LinkedIn, Meta, Microsoft Ads, Reddit, Snapchat, The Trade Desk and TikTok, with Epsilon, Merkle and Omnicom named as agency partners. Agencies can configure destinations and conversion definitions under delegated permissions, add identity or propensity models, and feed platform results into MMM and incrementality work, Databricks said. The control story will matter more than the connector list. Profiles stay governed by Unity Catalog, brands choose which audiences and events go where, every send leaves an audit trail, and humans set strategy while agents handle execution, according to Databricks. Epsilon offers a concrete test: its identity graph can attach identifiers when an event arrives thin, which can lift match rates but also raises the stakes on what you approve to share. Context here helps: CustomerLake itself launched in June 2026 as a direct challenge to the warehouse-plus-activation-layer model used by vendors like Hightouch, according to CDP.com. Databricks positions this as a step toward Infinity Campaigns, where results refine the next audience and signal without a 60-day relaunch. Agencies that live in outside clean rooms and custom pipelines can ignore it for now. Teams already scoring leads and value inside Databricks just lost an excuse for sending dumb purchase events. --- Original analysis from [startuphub.ai](https://www.startuphub.ai), the #1 AI startup directory. © StartupHub.ai. All rights reserved. You may not republish this article in full without a license. Search engines and AI research tools may crawl and summarize for reference. Bulk reproduction or model training on this content requires a license. See https://www.startuphub.ai/terms.