Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Agentic Control Center for Data Product Optimization

Priyadarshini Tamilselvan, Gregory Bramble, Sola Shirai, Ken C. L. Wong, Faisal Chowdhury, Horst Samulowitz

TL;DR

A system that automates data product improvement through specialized AI agents operating in a continuous optimization loop that transforms data into observable and refinable assets that balance automation with trust and oversight is proposed.

Abstract

Data products enable end users to gain greater insights about their data by providing supporting assets, such as example question-SQL pairs which can be answered using the data or views over the database tables. However, producing useful data products is challenging, and typically requires domain experts to hand-craft supporting assets. We propose a system that automates data product improvement through specialized AI agents operating in a continuous optimization loop. By surfacing questions, monitoring multi-dimensional quality metrics, and supporting human-in-the-loop controls, it transforms data into observable and refinable assets that balance automation with trust and oversight.

Agentic Control Center for Data Product Optimization

TL;DR

A system that automates data product improvement through specialized AI agents operating in a continuous optimization loop that transforms data into observable and refinable assets that balance automation with trust and oversight is proposed.

Abstract

Data products enable end users to gain greater insights about their data by providing supporting assets, such as example question-SQL pairs which can be answered using the data or views over the database tables. However, producing useful data products is challenging, and typically requires domain experts to hand-craft supporting assets. We propose a system that automates data product improvement through specialized AI agents operating in a continuous optimization loop. By surfacing questions, monitoring multi-dimensional quality metrics, and supporting human-in-the-loop controls, it transforms data into observable and refinable assets that balance automation with trust and oversight.
Paper Structure (17 sections, 3 figures)

This paper contains 17 sections, 3 figures.

Table of Contents

  1. Introduction
  2. Related Work
  3. Architecture Overview
  4. State Management
  5. State Manager
  6. Data Product State
  7. Data Product Quality Metrics Management
  8. Data Product Quality Metrics and Dependencies
  9. Context Generation and Event-Driven Recalculation
  10. Tool Management and Integration
  11. AI Agents and Orchestration
  12. Planner Agent
  13. Input Planner Agent
  14. Specialized Agents
  15. Recalculate Metrics
  16. ...and 2 more sections

Figures (3)

  • Figure 1: The Agentic Control Center allows users to observe key statistics and trends as AI agents perform actions to improve the underlying data product.
  • Figure 2: The two-tier Agentic Control Center System Architecture where the frontend connects data sources and sets quality targets. The backend combines a State Manager (tracking state and metrics) with an Agentic Workflow that iteratively plans, executes, and optimizes to achieve those targets.
  • Figure 3: (a) Mapping of tools (e.g., question generation, view creation) to quality metrics such as coverage, complexity, and speed, with arrows showing positive (green) or optimizing (orange) effects. (b) The Planner Agent selects actions, the Input Planner Agent calibrates parameters, Specialized Agents execute updates, and metrics are recalculated for iterative optimization.