Tamr AI-native MDM Demo

We understand that managing critical data across your enterprise can be challenging and often prone to inaccuracies. Traditional Master Data Management (MDM) solutions have attempted to address these challenges by implementing complex rule systems overseen by data teams. However, this approach is prone to human error, hinders business growth, and requires extensive teams to manage. In today's AI-driven era, it's no longer effective to rely on outdated methods.

‍

At Tamr, we believe that leveraging AI to do the heavy lifting, while allowing your data stewards to focus on curation, will help drive revenue and improve customer experience. In fact, 94% of business leaders believe AI is essential for their success over the next five years. Today, I will demonstrate how Tamr's AI data mastering solution provides clean, curated, and accurate data for your key business entities.

‍

Four Simple Steps to Data Mastering

‍

Tamr simplifies data mastering through a streamlined process that involves four key steps: connecting to your data sources, configuring your data product, choosing your enrichment providers, and curating the results.

‍

Step 1: Connecting to Your Data Sources

‍

The first step is connecting Tamr to your data sources. This process is straightforward. Tamr can read and write data from many cloud object stores and data warehouses such as GCS, BigQuery, and Snowflake. To connect a new source, simply visit the sources page.

‍

In this example, I’ve already connected several sources, but let's add a new one. First, give the source a name and description. Then, select the connection from which Tamr will read. The connections to these sources are persistent, not one-time reads, meaning that every time the data product runs, it updates the source data to keep the golden records current.

‍

Step 2: Configuring Your Data Product

‍

Next, we move on to configuring a data product to process the source data. For this demonstration, we’ll use the B2B customer data product to master site-level company data.

‍

First, we’ll select the sources we want to process—two sources are already added, but let’s add another. Each data product comes with a predefined, standardized schema. When you add a source, Tamr automatically maps attributes from each source to the schema, which significantly reduces the time required to onboard new data sources.

‍

As you can see, only a few attributes require manual mapping. Tamr recommends the most important attributes for the AI matching model. For this data product, these include company name, address, and website attributes. Above each source, you can click the view icon to see a sample of the source data, helping you complete the mappings.

‍

For each attribute, Tamr automatically recommends a value for the golden record. You can override this to specify your own logic—for example, by using the most frequent value for company name or setting other conditions. For now, I’ll leave these rules as they are.

‍

Step 3: Choosing Your Enrichment Providers

‍

Once the sources are configured, the next step is choosing how to enrich your data. Tamr’s Verified Match provides a vast corpus of referential data to match against. This helps verify that the entity exists, even if the data is sparse or noisy, making the golden records more trusted and reducing the amount of manual curation.

‍

Tamr curates referential data from many public data sources, including the name, address, website, and operating status of companies and their parent organizations. You can also use data from other enrichment providers for which you have a license. Regardless of whether you're using public or commercial enrichment sources, Tamr ensures consistent matching logic across the board, reducing the need for complex data pipelines.

‍

Additionally, this data product includes built-in data quality and enrichment features for things like countries, phone numbers, addresses, email addresses, and company names. For instance, Tamr automatically validates address data against trusted local postal service data and provides standardized address formats along with supplemental address information. Consumers of the data will have access to cleansed addresses from around the world, with no additional configuration or fixes required.

‍

Step 4: Curating the Results

Finally, after Tamr has created the golden records, it's time to engage your data stewards for curation. Tamr’s AI performs the majority of the work, achieving over 90% accuracy and scalability. Data stewards provide human oversight to ensure any remaining gaps in the data are addressed.

‍

Tamr provides tools for stewards to review and edit the mastered records as needed. Let’s use the company "PepsiCo Beverage Sales" as an example. Tamr identified that this company is a subsidiary of PepsiCo Inc. It automatically linked multiple entities to the account and grouped these source records into a single cluster. Tamr also provides suggestions for similar records that a curator should review.

‍

Curators can merge or unmerge records as necessary, manually override golden record values, or select the correct value from an alternative source. Once the curators finish reviewing the data, they can refresh the workflow to apply any changes. The activity log provides a permanent record of all changes made to the golden record.

‍

After the curation process, you can view the trusted golden record in the 360-degree page, where stakeholders can easily explore and drill down into key business data.

‍

Monitoring and Insights

‍

To help maintain golden records over time, Tamr provides an insights dashboard that tracks metrics such as data changes, the number of records read from a source, and updates to entities. The dashboard includes graphs comparing recent flow results and trends over time. These insights allow quick access to key metrics, eliminating the need to rely on external dashboards.

‍

Tamr’s virtual Chief Data Officer (CDO) is another valuable feature. This tool enables users to ask questions about the data via an intuitive prompt-based interface. Users can access Tamr-managed internal datasets in real time to enrich their own data and make it more complete. For example, the virtual CDO can add missing information to company records, helping ensure data quality.

‍

API and Downstream Integration

‍

Tamr also offers the ability to publish results to multiple downstream consumption endpoints. This process can be automated via our API, which continuously provides clean, curated, and accurate data at scale. Tamr’s APIs also allow for the searching, creation, and updating of golden records directly within enterprise applications in real time.

‍

Conclusion

‍

As you saw today, Tamr provides a turnkey solution for data mastering. With our pre-trained models and out-of-the-box enrichment, you can achieve trusted golden records in just four simple steps. To learn more about Tamr, please visit our website.