Empowering Market Insights and Decision-Making with Gen AI Agentic Workflows

February 25, 2025

By: Mridul Sinha, Associate Director - Product Engineering, Enterprise AI Advisory, LTIMindtree

The consumer-packaged goods (CPG) market is a sprawling, competitive landscape, offering a plethora of products that consumers use daily. Enter generative artificial intelligence (Gen AI) and large language models (LLMs), which are reshaping business operations, workflows, marketing, personalization, product innovation, and data privacy. These technologies are not just buzzwords; they are enhancing efficiency, personalizing offerings, optimizing supply chains, and enabling large-scale analysis of consumer data.

Initially, generative AI generated a lot of excitement, prompting brands to question its potential benefits and risks. Today, many brands are actively experimenting with Gen AI to improve brand-consumer interactions, earn trust, and lighten workloads. However, organizations face significant challenges with AI/ML technologies due to high costs, time consumption, and inefficient data management. The success of AI and ML systems hinges on high-quality, clean data, but data often comes from disparate sources with varying formats, complicating integration. Missing information, inconsistencies, and errors can significantly reduce efficiency and become resource-intensive.

By optimizing the data integration process, generative AI can streamline operations and reduce costs and time. According to marketresearch.biz, generative AI in the CPG market is projected to grow significantly, reaching USD 5452.4 million by 2032 with a compound annual growth rate (CAGR) of 9.1% from 2024 to 2033. CPG companies can leverage generative AI to improve supply chains, enhance loss prevention, and promote sustainability. Additionally, generative AI will accelerate R&D innovations, helping companies bring products to market faster than their competitors.

McKinsey’s research highlights that product managers (PMs) using gen AI can speed up time to market by 5% and enhance productivity by 40%.

Challenges faced by CPG businesses

Traditionally, the CPG industry has relied on after-the-fact measurement studies for precise market understanding, often partnering with third-party syndicated data providers. These partnerships grant CPG companies access to granular, zip code-level data, which can be merged with enterprise data to extract game-changing insights. However, the rapid pace of market change makes it challenging for CPG companies to analyze this data quickly and effectively due to:

Data format inconsistencies and quality issues: Different data providers often have unique formats and aggregation variations, leading to data quality challenges and inaccuracies. According to a report by Data Ladder in 2022, organizations face an average annual loss of $15 million due to poor data quality.
Figure 1: Challenges of data quality and inconsistencies
Data fragmentation: Critical information is often siloed across various internal and external sources, hindering swift and effective decision-making at the Stock Keeping Unit (SKU) level. The 2024 State of Data Governance Report by TDWI highlights that organizations with inadequate data governance frameworks experience a 25% reduction in operational efficiency due to data quality issues and inconsistencies.Figure 2: Challenges of fragmented insights
Inconsistent insights: Data fragmentation leads to the unavailability of holistic views, causing analysis to miss out on broader context and potentially leading to incomplete, skewed, or conflicting insights. According to a 2024 McKinsey report, 85% of CPG companies still struggle with data fragmentation and inconsistencies, hindering their ability to fully leverage advanced analytics and AI.

How Gen AI can overcome these challenges

Gen AI-based solutions can utilize a dynamic agent-based model tailored for the CPG industry, involving a sophisticated multi-agent network centered around a central agent.

This central agent, powered by a large language model, efficiently captures specific user requirements and breaks them down into smaller tasks assigned to specialized sub-agents.

Each sub-agent, also coupled with large language model capabilities, handles different data dimensions. When a query or task is received by the central agent, these sub-agents act on connected data dimensions to create relevant outputs.

The output generated by each queried sub-agent are then converged by the central agent to create meaningful insights.

Figure 3: Gen AI and LLM-enabled dynamic agent-based model approach

By integrating multiple data sources, this approach eliminates data silos and ensures cohesive and comprehensive analysis of market dynamics at a granular level. This allows for the delivery of real-time actionable insights tailored to meet the specific needs of different CPG personas, enhancing decision-making and giving businesses a competitive edge in a rapidly changing market.

Use cases in the CPG industry

Data harmonization using agentic workflows

Dynamic data integration: Generative AI agents autonomously handle ETL processes, adapting to different data formats and structures for seamless integration.
Enhanced data standardization: AI agents detect and resolve inconsistencies in data formats, ensuring uniformity across datasets.
Synthetic data generation: AI agents augment datasets by generating variations of existing data, enhancing model training.
Continuous learning and adaptation: AI agents continuously learn from new data and interactions, improving performance over time.
Automated workflows: Automation agents handle data validation, error detection, and correction
Output evaluation: AI agents evaluate data for completeness, accuracy, and consistency and validate harmonized data against benchmarks.

Converging insights from different generative AI-powered agentic workflow solutions

The central agent accepts user queries or actions through a conversation assistant UI/UX screen and breaks the user prompt into smaller chunks of queries and actions.

It explores the end-user intent to determine if the provided information is sufficient or requires more input.
The central agent can refer to available contexts in the form of vector databases to provide additional context while gathering more information.
It connects with specialized sub-agents powered by large language models, which are independently implemented solutions.

Sub-agents receive queries or actions from the central agent, process them, and send the output back to the central agent.

The central agent converges these outputs to create meaningful insights, which can be presented to the user in various formats such as reports, dashboards, text, voice, etc.

Advantages of multi-agent solutions

The introduction of a multi-agent solution approach brings many benefits to the CPG industry, including:

Productivity Boost: Agentic AI can improve productivity by up to 40% and enhance efficiency by 30%, significantly boosting strategic decision-making
Operational cost reduction: Agentic AI can reduce operational costs by 25% and increase customer satisfaction by 35% through more personalized and efficient services

LTIMindtree has worked extensively to create a Gen AI offering that follows the above solution approach to enable CPG organizations to be Gen AI ready and realize its benefits, making them more agile, data and insights driven.

Transforming decision-making with GenAI-powered agentic workflows

At LTIMindtree, we understand the ever-evolving technological needs of CPG businesses. Our tailored AI offering, leveraging Gen AI agentic workflows, transforms how businesses create actionable insights and accelerate decision-making through faster, more intelligent, and exceptionally reliable analysis. The integrated LLMs connected to multiple sub-agents, each handling a different data model, allow businesses to create comprehensive insights and navigate the complexities of the market with confidence, making informed decisions that drive growth and innovation.

Our GenAI-agentic workflow offering encompasses many sub-agent solutions that solve different business requirements in various data dimensions. While our eCommerce portfolio analyzer solution provides smarter and faster insights to decode market demand for a sustainable product portfolio, another solution solving 6P’s integrates data from various third-party syndicated data providers to create comprehensive consumer insights that go beyond surface-level analysis.

The application of generative AI and agentic workflows to create comprehensive, intelligent, and actionable holistic market insight is transforming and reshaping the CPG industry.

LTIMindtree, a pioneer in generative AI-related innovative solutions, along with its RCPG industry knowledge and ability, can help organizations maintain a competitive edge and drive sustained growth. Write to us at EAIBusiness.Advisory@LTIMindtree.com.

Blogs