Insights from the Data Masters Forum: Agentic AI Takes Center Stage

Data leaders showed up in a big way for Tamr’s inaugural Data Masters Forum. With stops in New York City and London, each event was filled with insightful conversations and meaningful connections. Here’s a look back at some of the highlights.

‍AI as the Great Unlock for Master Data Management (MDM)

Agentic AI is gaining ground, with LLM-based agents disrupting routine work across a number of functions ranging from customer service and business development to software engineering. Data engineering is also routine work, making it a prime target for AI automation. And it's these shifts that are prompting organizations to rethink their MDM strategies. 

Think about it: AI is driving up the demand for clean data more than ever before. Because as we’ve all experienced, when algorithms learn from bad data, the results are faulty insights, bad conclusions, and even hallucinations. But cleaning and mastering bad data is time-consuming and labor-intensive. Throwing more people at the problem doesn’t work. Nor does adding more processes, especially when they are rules-based and governance-heavy. And the reason these approaches don’t work is because they can’t scale. 

What does work is adding more productivity. That’s what Tamr does. Tamr’s AI-native approach uses advanced AI and machine learning capabilities to address data mastering challenges at scale, for a lower cost and with better outcomes. And with the addition of LLM-based AI agents, Tamr only gets better. 

Today, Tamr does a great job of creating golden records with persistent IDs for key business entities like customers, suppliers, and contacts. But increasingly, organizations want to understand the business context and how these entities relate to each other, which means it is as important to resolve these entities as it is to link them together. By connecting entities and their complex relationships with each other, we can create an enterprise knowledge graph that highlights the relationships among and between business entities. And when you understand these relationships, you see results, like an increase in revenue, decrease in cost, and reduced risk. 

By using Tamr’s AI/ML approach to match and merge records, companies can master about 90% of their data. Combine that with select, company-specific rules and the percentage of records mastered jumps to about 95%. That leaves the “last mile”—the last 5% of data that remains unresolved after the data mastering process. This last mile is notoriously hard to resolve, requiring human insight and expertise. However, with the introduction of LLM-based agents, the last mile becomes easier to manage. Agents can do the grunt work of basic data management—and when they’re unsure, they can escalate to a human for resolution. But more on that in a moment. 

A Shift from Managing Transactions to Mastering Trust

In 2024, generative AI was all the rage. But just one year later, the conversation has already shifted to agentic AI and LLM-based agents. AI and AI agents are driving massive changes in areas related to hyper-personalization, predictive insights, customer needs, and customer behaviors. Which begs the question: Do you trust your customer data?

If you don’t, you’re not alone. According to KPMG, 56% of CEOs are concerned about the quality of their data. But as AI—and now agentic AI—continue to take hold, an important, yet very different, question emerges: Do you trust your data enough to have AI act on it?  

Driving AI innovation is at the top of the agenda for many global organizations. But without a culture of trust and high-quality, trustworthy data, they will struggle to move this agenda forward. After all, there are countless examples of AI and AI agents providing misinformation simply because the data sets they learned from were inaccurate. 

Yet as companies move forward in this new, AI-driven era, trust becomes more important than ever before. You don’t have to trust all of your data, but you need to trust the data that is most important to your organization. And for many organizations, that is their customer data. 

So how do you improve trust? To start, think about data not as a record, but as a relationship. And consider how to move that relationship forward. Build trust into your workflows. Enrich your data so it’s more trustworthy. And shift your mindset—and your culture—from managing transactions to mastering trust. Because the companies that succeed in this AI-driven era are the ones who enrich, unify, and activate their data to truly know their customers.

Introducing Tamr Curator Hub: Mission Control for Human-AI Data Curation

Overcoming the challenges of mastering the “last mile” of data has often felt insurmountable. But with the introduction of Tamr’s new Curator Hub capabilities, that’s all about to change. 

Curator Hub signals Tamr’s entry into agentic data curation, a new concept that uses LLM-based AI agents to automate more of the data curation process by capturing and acting on the contextual insights needed to make confident curation decisions. Designed as the “mission control” for human-AI data curation, Curator Hub gives data stewards a central space to manage issues flagged by AI agents or submitted by users, making it easier and faster to resolve the data that remains after the mastering process. 

Curator Hub will be available as part of Tamr’s AI-native MDM platform to all Tamr Cloud customers and includes a range of features to help teams scale data quality, all within a customizable, user-friendly workspace. As showcased during the event, key features of Curator Hub include:

• Prioritized issue queue: Use Tamr’s agents to surface potential duplicates, missing values, and anomalies, ranked by urgency, so you can address the most pressing issues first.
• Decision-ready views: See side-by-side comparisons, understand why issues were flagged (with labels and confidence scores), and preview changes before applying updates.
• Golden record refinement: Move or reassign source records to ensure the correct grouping and composition of mastered entities, such as a person or organization to improve accuracy.
• Customizable workflows: Define how issues are routed, when agents are triggered, and which cases require human review.
• Transparent curation history: Track who made what changes and when, helping teams audit decisions and maintain data governance.
• System health insights: See resolution progress, monitor issue volume, and visualize data quality trends in real time.
• Built-in (and expanding) agent library: Tap into a starter set of prebuilt AI agents, with more on the way, including industry-specific tools and reusable logic templates.
• Bring Your Own Agent (BYOA): Integrate custom-built agents into Tamr’s workflow using a supported low-code approach.

In Summary

It’s clear we’ve entered an era where AI agents are transforming data management. New capabilities, such as Tamr’s new Curator Hub, will continue to provide new ways for organizations to improve data quality, increase efficiency, and deliver the best version of their data for use in AI, analytics, and operational use cases. And the companies that excel are the ones that embrace this transformation and successfully shift from managing transactions to mastering trust.