Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

PresAIse, A Prescriptive AI Solution for Enterprises

Wei Sun, Scott McFaddin, Linh Ha Tran, Shivaram Subramanian, Kristjan Greenewald, Yeshi Tenzin, Zack Xue, Youssef Drissi, Markus Ettl

TL;DR

PresAIse tackles enterprise barriers to prescriptive AI by integrating scalable causal inference with interpretable policy learning and a natural language interface. It combines a sparsity-regularized structure-learning approach to identify confounders with a set-partitioning based prescriptive tree to yield readable, actionable policies, and couples this with an open-source, memory-enabled LLM agent (PresAIse) built on LangChain to bridge business users and data scientists. The work provides concrete architectural components, including back-end LLMs, tool APIs for optimization and evaluation, an agent for decisions, and a memory layer, plus a front-end for interactive querying. The PoC in airline pricing demonstrates how non-ML experts can interact with prescriptive models, potentially democratizing advanced analytics while preserving causal rigor and transparency.

Abstract

Prescriptive AI represents a transformative shift in decision-making, offering causal insights and actionable recommendations. Despite its huge potential, enterprise adoption often faces several challenges. The first challenge is caused by the limitations of observational data for accurate causal inference which is typically a prerequisite for good decision-making. The second pertains to the interpretability of recommendations, which is crucial for enterprise decision-making settings. The third challenge is the silos between data scientists and business users, hindering effective collaboration. This paper outlines an initiative from IBM Research, aiming to address some of these challenges by offering a suite of prescriptive AI solutions. Leveraging insights from various research papers, the solution suite includes scalable causal inference methods, interpretable decision-making approaches, and the integration of large language models (LLMs) to bridge communication gaps via a conversation agent. A proof-of-concept, PresAIse, demonstrates the solutions' potential by enabling non-ML experts to interact with prescriptive AI models via a natural language interface, democratizing advanced analytics for strategic decision-making.

PresAIse, A Prescriptive AI Solution for Enterprises

TL;DR

PresAIse tackles enterprise barriers to prescriptive AI by integrating scalable causal inference with interpretable policy learning and a natural language interface. It combines a sparsity-regularized structure-learning approach to identify confounders with a set-partitioning based prescriptive tree to yield readable, actionable policies, and couples this with an open-source, memory-enabled LLM agent (PresAIse) built on LangChain to bridge business users and data scientists. The work provides concrete architectural components, including back-end LLMs, tool APIs for optimization and evaluation, an agent for decisions, and a memory layer, plus a front-end for interactive querying. The PoC in airline pricing demonstrates how non-ML experts can interact with prescriptive models, potentially democratizing advanced analytics while preserving causal rigor and transparency.

Abstract

Prescriptive AI represents a transformative shift in decision-making, offering causal insights and actionable recommendations. Despite its huge potential, enterprise adoption often faces several challenges. The first challenge is caused by the limitations of observational data for accurate causal inference which is typically a prerequisite for good decision-making. The second pertains to the interpretability of recommendations, which is crucial for enterprise decision-making settings. The third challenge is the silos between data scientists and business users, hindering effective collaboration. This paper outlines an initiative from IBM Research, aiming to address some of these challenges by offering a suite of prescriptive AI solutions. Leveraging insights from various research papers, the solution suite includes scalable causal inference methods, interpretable decision-making approaches, and the integration of large language models (LLMs) to bridge communication gaps via a conversation agent. A proof-of-concept, PresAIse, demonstrates the solutions' potential by enabling non-ML experts to interact with prescriptive AI models via a natural language interface, democratizing advanced analytics for strategic decision-making.
Paper Structure (14 sections, 2 equations, 6 figures, 2 tables)

This paper contains 14 sections, 2 equations, 6 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related art
  3. Methodology
  4. Scalable causal inference via structure learning
  5. Interpretable policy via prescriptive trees
  6. PresAIse, a prescriptive AI agent driven by LLMs
  7. Overview
  8. Back-end
  9. Open-source LLMs
  10. Tools
  11. Agent
  12. Memory
  13. Front-end user experience
  14. Conclusion

Figures (6)

  • Figure 1: Causal graph. $\pi$ is a discrete treatment, taking up to $q$ values, and $Y$ is a scalar outcome. $X$ is an observed set of $p$ covariates.
  • Figure 2: A feature graph with two features and the action (price). Two policies or decision rules are highlighted.
  • Figure 3: A prescriptive student tree for the airline pricing example
  • Figure 4: The framework for PresAIse, the prescriptive AI agent
  • Figure 5: Query current pricing policy
  • ...and 1 more figures