Tamr has introduced Curator Hub, a new module aimed at helping organisations address persistent data quality issues that can impede analytics and generative artificial intelligence projects. The new tool pairs AI-based agents with human supervision, focusing on complex data inconsistencies and gaps that often require specialist intervention.
Data quality risks
Poor data quality continues to present significant challenges and costs for enterprises pursuing AI initiatives. Research by Gartner has found that up to 30% of generative AI projects could be discontinued after initial trials due to issues with data quality, risk management, costs, or unclear value. These barriers emphasise the need for systems able to remediate complex or ambiguous data conflicts.
AI and human collaboration
Curator Hub aims to streamline the process of resolving 'last mile' data curation issues. While machine learning can automate the bulk of unifying and cleaning enterprise data, Tamr estimates that around 5% of cases remain difficult to resolve without human expertise. These typically involve ambiguous records-such as deciding whether two similar business names refer to the same entity-that previously meant manual review, taking up time and resources.
The Curator Hub module uses large language model-based agents to flag potentially problematic data. Human overseers can then review these flagged items, see side-by-side record comparisons, understand the agent's rationale, and preview proposed changes before acceptance.
Customer uptake
CHG Healthcare, which uses Tamr to unify fragmented provider information and reduce record duplication, expects the new capability to speed up data handling.
"I'm excited about agentic data curation," said Taner Maia, Senior Product Manager, Provider Services Platform, CHG Healthcare. "There's great value in focusing on the curator experience, and Curator Hub and AI agents have real potential to help. As we scale provider onboarding and streamline placement, getting the right data into the hands of the right people even more quickly would be a big win for CHG."
Functional overview
The module features a queue that ranks potential duplicates, incomplete values, or anomalies by urgency. Data stewards can see detailed comparisons between records and receive explanations for each suggested update. An audit trail logs edits to maintain governance oversight, while a dashboard offers insights into overall data quality and issue resolution rates.
Workflows within Curator Hub can be customised, allowing organisations to determine when cases need human review and how data problems are routed. The included library of prebuilt AI agents supports industry-specific tasks, and customers can add their own agents using a low-code interface.
User perspective
Consultants working with enterprise clients have highlighted the importance of transparency and usability in such systems.
"Curation workflows only work if stewards actually want to use them," said Paul Balas, Tamr user and Data Strategy Consultant, 303Computing. "Curator Hub is both intuitive and powerful - surfacing high-impact issues, showing what's changing, and keeping experts in control without overwhelming them. That's how you build trust, accelerate adoption, and get more value from your data."
Product rollout
Curator Hub is now available to all Tamr Cloud customers as part of the company's AI-driven master data management platform. The company plans to expand the module's agent library to address more industry challenges and to enable further customisation for complex data sets.
"Data projects lose momentum and value when results take too long to materialise," said Anthony Deighton, CEO, Tamr. "Curator Hub changes that, giving stewards an intuitive way to see, make and explain data improvements. It also provides an important vehicle for leveraging AI agents to automate and improve data curation efficiency - representing a pivotal shift in how organisations can build trust in their data and realise greater returns from downstream AI applications."
