# Clean data was never the real problem _IFS Copperleaf put IONA on raw data and let staff teach it definitions, claiming 70% activation._ **Published:** 2026-11-16 **Source:** https://www.startuphub.ai/ai-news/artificial-intelligence/2026/clean-data-was-never-the-real-problem --- IFS Copperleaf gave every employee plain English queries across nearly 30 systems with IONA, built on [Databricks](https://www.databricks.com/blog/clean-data-was-never-enough-how-ifs-copperleaf-taught-ai-its-business-databricks-genie-one-0) Genie One. The obvious reading is that raw data plus chat finally fixed analytics. I do not buy that. Finance and the marketing to sales funnel are still fully modeled, and [Databricks](https://www.startuphub.ai/startups/databricks) says IONA answers everywhere else on raw data. Misagh Jebeli, Enterprise Data Architect at IFS Copperleaf, is blunt about the gamble. "We let AI work on the raw data, and have the business teach it what the data means," he says. That teaching happens in Skills on top of Genie Ontology, with definitions owned by the business and approved by an AI Council before they apply company wide. This is where the claim gets tested. Lisa Richter, Director of Technology and AI at IFS Copperleaf, admits she would have killed the idea five years ago because business users now write logic with no ticket and no release cycle. Unity Catalog enforces access, Databricks says, and each definition has a named owner. But promotion still requires council review, so self serve has a gate. To be fair, IFS Copperleaf names a real pain. Ask three leaders for active customers and get three answers. IONA shows its work, points to the system owner, and lets users correct it in chat. A seller can now ask what a customer owns, where whitespace sits and if the contract allows a SaaS move, pulling license records and contract language together. What is missing is proof. Databricks says IONA hit 70% activation since its [July 2026](https://www.startuphub.ai/ai-news/prediction-markets/2026/polymarket-bets-on-bitcoin-s-july-2026-trajectory) launch and covered running costs within four months. Neither activation nor cost is defined. There is no error rate, no query volume, no comparison to the old reports. IFS acquired Copperleaf in 2024, per Techzine Global, so this is also a post merger integration story dressed as an AI story. That [context beats bigger models](https://www.startuphub.ai/ai-news/ai-research/2026/databricks-says-context-beats-bigger-models) argument only holds if the business keeps teaching. If you want this, budget for definitions, not just pipelines. --- 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.