Reframing Data Governance: How Autopay Leverages AI and Precision Strategies to Treat Data as a Financial Asset, Achieving 8x Better Marketing Performance and 92.9% Targeting Accuracy.

Data Leaders members can access the full presentation and connect with peers via the Data Leaders Hub.

During Data Leaders’ peer exchange: Mature Approaches to Data Governance, held in Munich on 13th November, Mateusz Jakitowicz, Chief Data Officer at Autopay introduced a new perspective on data governance, proposing that data should be seen not merely as a compliance requirement but as a core financial asset.

In this blog post, we explore Jakitowicz’s approach, which revolves around structuring data as an asset, leveraging AI for scalable governance and monetising data through advanced subscription models.

Data as a Financial Asset: Moving Beyond the “New Oil” Analogy

Jakitowicz reframes Autopay’s approach to data:Data isn’t the new oil; it’s money itself.

With this, he argues that by treating data as a direct asset each data point holds intrinsic value, and every inaccurate or outdated entry represents a financial loss. This positioning redefines data quality from a compliance or operational concern to a direct driver of revenue.

Operationalising Data as a Financial Asset

To put this philosophy into practice, Jakitowicz employs an “atomic model” for data structure, drawing on his experience in asset management. This model breaks down data into manageable, evaluable components, from master data such as B2C customer details to behavioural data that enriches these primary data points. By organising data hierarchically, Autopay can gauge each data point’s revenue potential, ensuring that governance efforts align closely with broader business goals.

Structured Data Classification for Maximum Value Extraction

Jakitowicz explains a multi-layered approach to data classification:

  1. Hierarchical Classification: Information is organised in descending levels of specificity. Broad information classifications are placed at the top (e.g., “critical company data”). Beneath this, business terms are defined from business glossary.
  2. Data Classes: Specific categories such as names or unique identifiers, which are vital for the business, fall under this layer.
    Location-Specific Classification: Each data type’s storage location, such as a column in a database, is defined, improving data accessibility and reliability.
  3. Data Quality Scoring: A quality score is assigned to each data point or class, directly tying data reliability to revenue by identifying errors as financial liabilities.

This classification system enables a granular view of data quality and directly links data integrity to business value, reinforcing Jakitowicz’s view of data as a core financial asset.

Scaling Data Infrastructure with AI-Powered Virtual Team Members

For lean teams, scaling data governance requires innovation. Jakitowicz explains how Autopay leverages AI-driven virtual agents to handle routine data requests, enabling efficient scaling while reducing dependency on manual processes. Far from simple bots, these agents, like “Valerie,” engage directly with Autopay’s data lake, answering queries in real-time or “Claire” “The LLM agent that operates at 2% of human labour costs and is almost 1,000 times faster,” underscoring the cost-effectiveness of AI.

By deploying these agents across departments, Autopay empowers employees at all levels to interact directly with data, fostering a self-service approach to data governance. “If I want to ask about data in our lake, I don’t need to reach a physical person; I just ask Valerie,” Jakitowicz notes. These AI agents facilitate autonomous data access, allowing Autopay’s 250 employees, including senior leadership, to make swift, data-driven decisions.

Autopay’s virtual agents play an instrumental role in scaling and streamlining data governance processes:

  1. Automated Data Querying and Access: Agents like Valerie and Claire allow employees to interact with Autopay’s data lake, querying information in real-time. This self-service approach makes data governance accessible across the organisation.
  2. Real-Time Compliance Verification: Agents continuously monitor and verify data within Autopay’s systems. For example, Claire reviews merchant information to confirm alignment with compliance regulations, automating risk management in a regulated environment.
  3. Enhanced Data Quality and Integrity: By identifying and cleaning data inconsistencies, these agents uphold data quality standards, which are critical to reliable governance.
  4. Scalability and Efficiency: Operating at just 2% of human labour costs, these agents make data governance scalable, freeing human resources for strategic tasks.
  5. Empowering a Data-Driven Culture: By enabling easy access to data, the virtual agents foster a data-driven culture throughout the company.

Participants of the peer exchange were invited to ask questions.

In response to a question regarding the reliability of these agents, Jakitowicz explained that Autopay’s virtual agent system uses a dual-agent architecture, where two agents cross-validate each other’s outputs, comparing results to deliver the most accurate answer. While acknowledging that AI tools are fallible, he noted that the system is faster and generally more accurate than human efforts, given the scale of data it processes.

In addition, when questioned on relying on question-answer databases – a common approach to validating AI outputs – referring to the insufficiency of validating billion-parameter models with limited datasets, Jakitowicz asked participants to consider which tasks in their own roles could be effectively managed by AI, emphasising that while not perfect, AI agents can be efficient tools in well-structured, repeatable processes.

An industry expert further clarified that validation practices from traditional software testing, like unit tests for intent recognition and entity extraction, are being adapted to large AI models. These unit tests are conducted before user testing with question-answer packs, ensuring robust validation in various scenarios.

Precision Marketing and Subscription Models: Monetising Data Directly

With high-quality data in place, Autopay drives revenue by leveraging this asset in targeted marketing and subscription models. Jakitowicz outlined how Autopay’s foundation models identify customers most likely to adopt subscription-based services, predicting price elasticity to maximise conversion. This precision-driven approach has allowed Autopay to offer tailored subscriptions far more effectively than generic advertising.

The results have been substantial. By using clean, AI-curated data, Autopay achieves a 92.9% targeting accuracy, yielding marketing performance that is eight times better than industry standards. This approach enables Autopay to offer differentiated services, from digital goods to premium loyalty cards. Jakitowicz shared that paid loyalty programmes consistently foster deeper engagement than free ones, illustrating the power of data-driven subscription models.

Insights into Autopay’s Proprietary Foundation Model

Autopay’s model is designed for time series and transactional data rather than general-purpose language processing, tailored specifically for subscription targeting and precision marketing.

Jakitowicz emphasised that creating such a model internally, rather than using third-party solutions, allows Autopay to refine and scale it efficiently. He distinguished this specific-purpose model from general-purpose language models, explaining that it provides insights across Autopay Partners without requiring additional training.

Final Thoughts: Data Governance as a Business Asset

For Chief Data Officers and data governance leads, Jakitowicz’s model at Autopay offers a new perspective on aligning data governance with revenue objectives. By treating data as a financial asset, deploying AI for scalable governance, and leveraging structured data models for targeted monetisation, Autopay illustrates how data governance can evolve from a compliance function to a driver of business value.

Jakitowicz’s advice for data leaders: “If you think about data governance, think about how to make money for the company on the data, and then you will get the funding to build tool you need.” Aligning data governance with revenue outcomes not only secures necessary resources but positions data teams as strategic contributors to the company’s success.

Ensure you get stories like this and many more interesting insights from data and analytics leaders like you – directly to your inbox – by signing up to our newsletter. Would like to become a member? Get in touch!

Membership

Make better informed decisions by assessing your data and analytics capability and use community intelligence to close the gap between strategy and execution.

Newsletter

Essential Data monthly newsletter is where you can discover limited availability articles, guides and frameworks otherwise only accessible to our members.